DE102019206844A9 - Analysis methods and devices therefor - Google Patents

Abstract

In order to provide a method for predicting process deviations in a process plant, for example in a paint shop, by means of which process deviations can be easily and reliably predicted, it is proposed according to the invention that the method comprises: automatic creation of a prediction model; prediction of process deviations during operation the process plant using the predictive model.

Description

The present invention relates to a method for fault analysis in a process engineering installation, for example in a painting installation.

The present invention is based on the object of providing a method for error analysis in a process engineering system, for example in a painting system, by means of which error situations can be analyzed simply and reliably.

This object is achieved by a method for error analysis in a process engineering system, for example in a painting system.

• - insbesondere automatisches Erkennen einer Fehlersituation in der verfahrenstechnischen Anlage;
• - Speichern eines Fehlersituationsdatensatzes zu der jeweiligen erkannten Fehlersituation in einer Fehlerdatenbank;
• - automatische Ermittlung einer Fehlerursache für die Fehlersituation und/oder automatische Ermittlung von für die Fehlersituation relevanten Prozesswerten auf Basis des Fehlerdatensatzes einer jeweiligen erkannten Fehlersituation.

The method for error analysis in a process engineering plant, for example in a painting plant, preferably includes the following:

• - in particular automatic detection of an error situation in the process engineering system;
• - Storage of an error situation data record for the respective recognized error situation in a defect database;
• - automatic determination of a cause of the error for the error situation and / or automatic determination of process values relevant to the error situation on the basis of the error data record of a respective recognized error situation.

In the context of this description and the appended claims, process values which are the cause of the error situation are understood to mean, in particular, process values which are the cause of the error situation and / or are related to them.

It can be beneficial if the error situation is recognized automatically by means of a reporting system.

• - einer Vorverknüpfung aus einem Meldesystem;
• - einer Zuordnung eines Prozesswerts zu demselben Anlagenteil der verfahrenstechnischen Anlage, in dem die Fehlersituation aufgetreten ist;
• - einer Verknüpfung eines Prozesswerts mit einer historischen Fehlersituation aufgrund einer aktiven Auswahl eines Benutzers;
• - einer aktiven Auswahl des Prozesswerts durch einen Benutzer.

In one embodiment of the method for error analysis in a process plant, one or more process values are automatically linked to the error situation based on one or more of the following linking criteria in order to automatically determine the cause of the error for the error situation and / or the process values relevant to the error situation :

• - a pre-link from a reporting system;
• - an assignment of a process value to the same part of the process plant in which the error situation occurred;
• a link between a process value and a historical error situation based on an active selection by a user;
• - an active selection of the process value by a user.

• - einer Prozessrelevanz der Prozesswerte;
• - einer Position eines Prozesswerts oder eines den Prozesswert ermittelnden Sensors innerhalb der verfahrenstechnischen Anlage;
• - einem Betrag einer Abweichung eines Prozesswerts von einem definierten Prozessfenster und/oder von einem Normalzustand;
• - einer Priorisierung von historischen Prozesswerten in historischen Fehlersituationen;
• - durch Übernehmen einer Priorisierung der Fehlerursache und/oder der Prozesswerte aus einem Meldesystem;
• - einer Priorisierung durch einen Benutzer.

In one embodiment of the method for error analysis in a process plant, it is provided that, for the automatic determination of the cause of the error for the error situation and / or the process values relevant to the error situation, an automatic prioritization of the process values linked to the error situation based on one or more of the following prioritization criteria is carried out:

• - a process relevance of the process values;
• a position of a process value or of a sensor determining the process value within the process engineering system;
• an amount of a deviation of a process value from a defined process window and / or from a normal state;
• - a prioritization of historical process values in historical error situations;
• - by adopting a prioritization of the cause of the error and / or the process values from a message system;
• - a prioritization by a user.

A prioritization carried out based on the process relevance of the process values is preferably carried out in such a way that process-critical process values are given higher priority.

In the context of this description and the appended claims, a process-critical process value is understood to mean, in particular, a process value that is stored in the reporting system as process-critical and / or has been defined as process-critical by a user.

A prioritization carried out based on a position of the process value or a sensor determining the process value within the process engineering system is preferably carried out in such a way that process values that are assigned to the same, a nearby and / or a comparable part of the system are given higher priority.

In the context of this description and the appended claims, comparable system parts are understood to mean, in particular, system parts with a similar or identical structure.

Comparable system parts are, for example, identical or similar industrial air supply systems, identical or similar Conditioning modules of an industrial air supply system or pumps or motors of the same or similar construction.

A position of the sensor determining the process value is preferably identified in the process engineering system using a system from a numbering system (so-called “plant numbering system”).

Process values are preferably clearly identified by means of the numbering system.

Process values are preferably prioritized in the numbering system depending on their designation.

For the unambiguous designation of sensors and / or process values, the numbering system preferably includes the designation of a functional unit, the designation of a functional group of the respective functional unit and / or the designation of a functional element of the respective functional group to which the respective sensor and / or process value is assigned.

It can also be favorable if the unambiguous designation of a process value by means of the numbering system includes a designation of a type of the measured variable, for example a temperature, a flow rate, a pressure.

For example, an air supply system of a painting system is a functional unit, a conditioning module of the air supply system being a functional group and a pump of the air supply system being a functional element.

A normal state of a process value is preferably determined by means of a method for anomaly and / or error detection.

A prioritization carried out based on a prioritization of historical process values in historical error situations is preferably carried out in such a way that process values are prioritized analogously to the historical error situation.

• - einer Prozessrelevanz der Prozesswerte;
• - einer Position eines Prozesswerts oder eines den Prozesswert ermittelnden Sensors innerhalb der verfahrenstechnischen Anlage;
• - einem Betrag einer Abweichung eines Prozesswerts von einem definierten Prozessfenster und/oder von einem Normalzustand;
• - einer Priorisierung von historischen Prozesswerten in historischen Fehlersituationen;
• - physikalischen Abhängigkeiten der Prozesswerte.

In one embodiment of the method for error analysis in a process engineering system, further error causes and / or process values are proposed to automatically determine the cause of the error for the error situation and / or the process values relevant to the error situation, the suggestion being based on one or more of the following Proposal criteria is carried out automatically:

• - a process relevance of the process values;
• a position of a process value or of a sensor determining the process value within the process engineering system;
• an amount of a deviation of a process value from a defined process window and / or from a normal state;
• - a prioritization of historical process values in historical error situations;
• - physical dependencies of the process values.

A process-critical process value is preferably suggested.

A proposal made based on a position of the process value or a sensor determining the process value within the process engineering system is preferably made such that process values that are assigned to the same, a nearby and / or a comparable part of the system are more likely to be proposed.

Process values are preferably proposed in the numbering system depending on their designation.

A proposal carried out on the basis of a prioritization of historical process values in historical error situations is preferably made in such a way that process values with a high priority are preferably proposed in the historical error situation.

To determine a proposal based on physical dependencies of the process values, the physical dependencies are preferably defined by a user as an expert rule.

A prioritization of the suggested causes of errors and / or process values can preferably be changed by a user.

• - einer Fehlerklassifikation der historischen Fehlersituation;
• - einer historischen Fehlersituation an demselben oder einem vergleichbaren Anlagenteil;
• - identischen oder ähnlichen Prozesswerten der historischen Fehlersituation zu Prozesswerten der erkannten Fehlersituation.

In one embodiment of the method for error analysis in a process plant, it is provided that historical error situations are determined from an error database on the basis of one or more of the following similarity criteria:

• - an error classification of the historical error situation;
• - a historical error situation on the same or a comparable part of the system;
• - identical or similar process values of the historical error situation to process values of the recognized error situation.

Historical error situations with an error classification that is identical to the detected error situation are preferably determined.

An identity or similarity of the process values of the historical error situation to the process values of the recognized error situation is preferably determined by a comparison algorithm.

In one embodiment of the method for error analysis in a process plant, it is provided that historical process values are determined from a process database, which are identical or similar to process values of the detected error situation.

A process database is preferably searched to determine the historical process values. It can be beneficial if an identity or similarity of the process values is determined by means of a comparison algorithm.

The historical process values are preferably determined automatically.

In one embodiment of the method for error analysis in a process engineering system, it is provided that the historical process values determined are identified as belonging to a historical error situation.

In one embodiment of the method for error analysis in a process engineering system, provision is made for an error situation data set to be stored in an error database for a recognized error situation.

• - eine Fehlerklassifikation der Fehlersituation;
• - mit der Fehlersituation verknüpfte Prozesswerte basierend auf einer Vorverknüpfung aus einem Meldesystem;
• - Informationen über einen Zeitpunkt des Auftretens einer jeweiligen Fehlersituation;
• - Informationen über eine Zeitdauer des Auftretens einer jeweiligen Fehlersituation;
• - Informationen über den Ort des Auftretens einer jeweiligen Fehlersituation;
• - Alarme;
• - Statusmeldungen.

In one embodiment of the method for error analysis in a process engineering system, it is provided that a respective error identification data record includes one or more of the following error situation data:

• - an error classification of the error situation;
• - Process values linked to the error situation based on a pre-link from a reporting system;
• - Information about a point in time when a respective error situation occurred;
• - Information about a duration of the occurrence of a respective error situation;
• - Information about the location of the occurrence of a respective error situation;
• - alarms;
• - Status messages.

In one embodiment of the method for error analysis in a process engineering system, it is provided that the error situation data record of a respective error situation includes error identification data for the unambiguous identification of the detected error situation.

The error identification data can preferably be used for the unambiguous designation of an error situation.

In one embodiment of the method for error analysis in a process engineering system, it is provided that documentation data and error correction data are stored in the error situation data record for a respective error situation.

Documentation data preferably include operating instructions, manuals, circuit diagrams, process diagrams and / or data sheets for the system parts affected by a particular error situation.

Error elimination data preferably include information for elimination of an error situation, in particular instructions for correcting an error situation.

In particular, documentation data and error correction data can also be added to the error situation data record by a user.

In one embodiment of the method for error analysis in a process engineering system, it is provided that process values are stored in the operation of the process engineering system in a time-synchronized manner with a recognized error situation.

In one embodiment of the method for error analysis in a process engineering system, it is provided that process values are provided with a time stamp, by means of which the process values can be clearly assigned to a point in time.

The present invention also relates to a fault analysis system for fault analysis in a process plant, for example in a paint shop, which is designed and set up to carry out the method according to the invention for fault analysis in a process plant, for example in a paint shop.

The present invention also relates to an industrial control system which comprises a fault analysis system according to the invention.

The present invention also relates to a method for predicting process deviations in a process plant, for example in a painting plant.

The present invention is based on the further object of providing a method for predicting process deviations in a process engineering installation, for example in a painting installation, by means of which process deviations can be easily and reliably predicted.

This object is achieved by a method for predicting process deviations in a process engineering system, for example in a painting system.

• - automatisches Erstellen eines Vorhersagemodells;
• - Vorhersage von Prozessabweichungen im Betrieb der verfahrenstechnischen Anlage unter Verwendung des Vorhersagemodells.

The method for predicting process deviations in a process plant, for example in a paint shop, preferably includes the following:

• - automatic creation of a predictive model;
• - Prediction of process deviations in the operation of the process plant using the prediction model.

A process deviation from production-critical process values can preferably be predicted by means of the prediction model.

In one embodiment of the method for predicting process deviations, it is provided that the method for predicting process deviations is carried out in an industrial air supply system, in a pretreatment station, in a station for cathodic dip painting and / or in a dryer station.

Industrial air supply systems, pretreatment stations and / or stations for cathodic dip painting are in particular very slow process engineering systems.

Production-critical process values of such process engineering systems therefore change only very slowly when they are in operation.

Due to the great inertia of such process engineering systems, a process deviation in the operation of the process engineering system can preferably be predicted at an early stage by means of the prediction model.

Time savings for repairing and / or maintaining such process engineering systems can thus preferably be achieved before a process deviation occurs.

An industrial air supply system preferably comprises several conditioning modules, for example a preheating module, a cooling module, a post-heating module and / or a humidifier module.

The prediction model created can preferably be transferred to similar process engineering systems.

For example, it is conceivable that a prediction model that was created for a pretreatment station can be used for a station for cathodic dip painting.

In one embodiment of the method for predicting process deviations, it is provided that the prediction model is used to predict process deviations from production-critical process values in the process plant, in particular on the basis of changing process values during operation of the process plant.

In the context of this description and the appended claims, production-critical process values are understood to mean, in particular, process values whose deviation from a predetermined process window leads to a quality deviation, in particular to quality defects.

Production-critical process values of an industrial air supply system are, for example, the temperature and the relative humidity of the air conditioned by means of the industrial air supply system, in particular at a discharge part of the industrial air supply system.

Air conditioned by means of an industrial air supply system is preferably supplied to a paint booth in a paint shop and thus preferably has a direct effect on the quality of treatment of the workpieces treated in the paint booth, in particular the vehicle bodies treated in the paint booth.

For example, it is possible to use the prediction model to predict process deviations from production-critical process values for a prediction horizon of at least approximately 10 minutes, for example of at least approximately 15 minutes, preferably of at least approximately 20 minutes.

In one embodiment of the method for predicting process deviations, provision is made for process values and / or status variables to be stored over a predetermined period of time during operation of the process engineering system in order to automatically create the prediction model.

• - Zielgrößen der industriellen Zuluftanlage, insbesondere Temperatur und relative Luftfeuchte der mittels der industriellen Zuluftanlage konditionierten Luft, insbesondere an einem Ausblasteil der industriellen Zuluftanlage;
• - Stellgrößen, insbesondere Ventilstellungen von Ventilen von Heiz- und/oder Kühlmodulen der industriellen Zuluftanlage, Drehfrequenzen von Pumpen, insbesondere der Befeuchterpumpe, und/oder Drehfrequenzen von Ventilatoren;
• - interne Größen, insbesondere Vor- und/oder Rücklauftemperaturen in der Heiz- und/oder Kühlmodulen der industriellen Zuluftanlage und/oder Luftkonditionen zwischen Konditionierungsmodulen;
• - gemessene Störgrößen, insbesondere Außentemperatur und/oder relative Außenluftfeuchte an einem Einblasteil der industriellen Zuluftanlage;
• - ungemessene Störgrößen; und/oder
• - Statusgrößen, insbesondere Befeuchterpumpe (an/aus), manueller Modus für Pumpen (an/aus), Versorgungsventile (offen/geschlossen), Ventilator (an/aus).

If the process plant is an industrial air supply system, the stored process values and / or status variables preferably include the following:

• - Target variables of the industrial air supply system, in particular the temperature and relative humidity of the air conditioned by means of the industrial air supply system, in particular at a discharge part of the industrial air supply system;
• - manipulated variables, in particular valve positions of valves of heating and / or cooling modules of the industrial air supply system, rotational frequencies of pumps, in particular of the humidifier pump, and / or rotational frequencies of fans;
• - internal variables, in particular flow and / or return temperatures in the heating and / or cooling modules of the industrial air supply system and / or air conditions between conditioning modules;
• - Measured disturbance variables, in particular outside temperature and / or relative outside air humidity at a blow-in part of the industrial air supply system;
• - unmeasured disturbances; and or
• - Status variables, in particular humidifier pump (on / off), manual mode for pumps (on / off), supply valves (open / closed), fan (on / off).

In the context of this description and the appended claims, process values are understood to mean in particular time-dependent, continuous signals.

In the context of this description and the appended claims, status variables are understood to mean in particular time-dependent discrete events.

• - die verfahrenstechnische Anlage ist in dem vorgegebenen Zeitraum zu mindestens ungefähr 60 %, vorzugsweise zu mindestens ungefähr 80 %, in einem betriebsbereiten Zustand, insbesondere für einen Produktionsbetrieb;
• - die verfahrenstechnische Anlage ist in dem vorgegebenen Zeitraum zu mindestens ungefähr 60 %, vorzugsweise zu mindestens ungefähr 80 %, in einem produktionsfähigen Zustand;
• - die verfahrenstechnische Anlage wird in dem vorgegebenen Zeitraum insbesondere mit sämtlichen möglichen Betriebsstrategien betrieben;
• - einer vorgegebene Anzahl von Prozessabweichungen und/oder Störfällen in dem vorgegebenen Zeitraum.

In one embodiment of the method for predicting process deviations, it is provided that the specified period of time over which process values and / or status variables are stored during operation of the process engineering system is specified as a function of one or more of the following criteria:

• the process engineering system is in an operational state for at least approximately 60%, preferably at least approximately 80%, in the specified period of time, in particular for a production facility;
• the process-engineering plant is in a productive state for at least approximately 60%, preferably at least approximately 80%, in the specified period;
• - The process engineering system is operated in the specified period, in particular with all possible operating strategies;
• - a specified number of process deviations and / or incidents in the specified period.

• - ein Ventilator der industriellen Zuluftanlage in Betrieb ist (Statusgröße des Ventilators ist „an“);
• - Konditionierungsmodule der industriellen Zuluftanlage in einem Automatik-Modus betrieben werden;
• - mindestens ein Regelventil geöffnet ist; und/oder
• - eine Befeuchterpumpe in Betrieb ist (Statusgröße der Befeuchterpumpe ist „an“).

If the process plant is an industrial air supply system, it is preferably in an operational state for a production facility if:

• - a fan of the industrial supply air system is in operation (status variable of the fan is "on");
• - Conditioning modules of the industrial air supply system are operated in an automatic mode;
• - At least one control valve is open; and or
• - a humidifier pump is in operation (status variable of the humidifier pump is "on").

In the context of this description and the appended claims, a productive state of a process plant is understood in particular to mean that target variables of the process plant are within a predetermined process window.

If the process engineering system is an industrial air intake system, it is in a productive state if the target variables of the industrial air intake system, in particular a temperature and relative humidity of the air conditioned by means of the industrial air intake system, in particular at a discharge part of the industrial air intake system, are within a specified process window are located.

A pretreatment station or a station for cathodic dip painting can in particular only be operated with a single operating strategy.

An industrial air supply system can in particular be operated with several operating strategies, in particular as a function of ambient conditions.

An industrial air supply system can be operated, for example, with the following operating strategies: heating-humidifying, cooling-heating, cooling-humidifying, cooling, heating, humidifying.

If the process engineering system is an industrial air supply system, it is in particular conceivable that process values and / or status variables for automatically creating the forecast model over a period of, for example, at least approximately 6 months, in particular over for a period of at least approximately 9 months, preferably for a period of at least approximately 12 months.

If the process plant is a pretreatment station or a station for cathodic dip painting, it is particularly conceivable that process values and / or status variables for the automatic creation of the prediction model over a period of, for example, at least approximately 2 weeks, in particular over a period of at least approximately 4 weeks, preferably for a period of at least about 6 weeks.

For example, it is conceivable that at least approximately 30, preferably at least approximately 50, process deviations and / or malfunctions occur in the specified period.

In particular, it is also conceivable that the predefined period of time over which process values and / or status variables are stored in the operation of the process engineering system comprises several non-contiguous partial periods of time.

• - eine Mindestdauer des Teilzeitraums, beispielsweise mindestens ungefähr 30 Minuten;
• - ein Betrieb der verfahrenstechnischen Anlage in einem produktionsfähigen und/oder eingeschwungenen Zustand zu Beginn des Teilzeitraums;
• - ein Betrieb der verfahrenstechnischen Anlage in einem eingeschwungenen Zustand am Ende eines Teilzeitraums.

If the period of time over which process values and / or status variables are stored in the operation of the process engineering plant comprises several non-contiguous partial periods, the partial periods preferably each have one or more of the following criteria:

• - a minimum duration of the partial period, for example at least approximately 30 minutes;
• - an operation of the process engineering plant in a productive and / or steady state at the beginning of the partial period;
• - an operation of the process plant in a steady state at the end of a partial period.

In one embodiment of the method for predicting process deviations, provision is made for a machine learning method to be carried out to create the prediction model, the process values and / or status variables stored over the specified period of time being used to create the prediction model.

Machine learning methods carried out to automatically create the prediction model preferably include one or more of the following: Gradient Boosting, Random Forest, Support Vector Machine.

In one embodiment of the method for predicting process deviations, it is provided that the machine learning method is carried out on the basis of features which are extracted from the process values and / or status variables stored over the specified period.

• - statistische Kennzahlen;
• - Koeffizienten aus einer Hauptkomponentenanalyse;
• - lineare Regressionskoeffizienten;
• - dominante Frequenzen und/oder Amplituden aus dem Fourier-Spektrum.

In one embodiment of the method for predicting process deviations, it is provided that one or more of the following are used to extract the features:

• - statistical indicators;
• - coefficients from a principal component analysis;
• - linear regression coefficients;
• - dominant frequencies and / or amplitudes from the Fourier spectrum.

Statistical key figures include, for example, a minimum, a maximum, a median, a mean and / or a standard deviation.

In one embodiment of the method for predicting process deviations, it is provided that a selected number of prediction data sets with process deviations and a selected number of prediction data sets without process deviations are used for training the prediction model.

In particular, it is conceivable that the selected number of prediction data sets with process deviations corresponds at least approximately to the selected number of prediction data sets without process deviations.

In particular, it is conceivable that the selected number of prediction data sets with process deviations and the selected number of prediction data sets without process deviations are identical.

• - einem zeitlichen Mindestabstand zwischen zwei Prädiktionsdatensätzen mit Prozessabweichungen;
• - einer automatischen Auswahl anhand von definierten Regeln;
• - einer Auswahl durch einen Benutzer.

In one embodiment of the method for predicting process deviations, it is provided that the number of prediction data sets with process deviations is selected on the basis of one or more of the following criteria:

• - a minimum time interval between two prediction data sets with process deviations;
• - an automatic selection based on defined rules;
• - a selection by a user.

A minimum time interval between two prediction data sets with process deviations is, for example, at least approximately two hours.

In one embodiment of the method for predicting process deviations, it is provided that prediction data sets with process deviations are identified as such if a process deviation occurs within a predetermined time interval.

A predefined time interval preferably consists of a time span of the prediction data set and a selected prediction horizon.

For example, it is conceivable that the time span of the prediction data set is 30 minutes and that the selected prediction horizon is 15 minutes.

A prediction data set without process deviations is identified as such if there is no process deviation within the specified time interval.

In one embodiment of the method for predicting process deviations, it is provided that the process values and / or status variables stored over the specified period are summarized in prediction data sets by preprocessing.

• - Regularisierung der über den vorgegebenen Zeitraum gespeicherten Prozesswerte;
• - Zusammenfassung der Prozesswerte und/oder Statusgrößen in Prädiktionsdatensätze durch Einteilung der Prozesswerte und/oder Statusgrößen in Zeitfenster mit zeitlicher Verschiebung.

In one embodiment of the method for predicting process deviations, it is provided that the preprocessing comprises the following:

• - Regularization of the process values stored over the specified period;
• - Summary of the process values and / or status variables in prediction data sets by dividing the process values and / or status variables into time windows with a time shift.

A duration of a time window is preferably greater than the time shift.

The duration of a time window is 30 minutes, for example.

The time shift is, for example, 5 minutes.

Prediction data records that follow one another in time preferably each include process values and / or status variables with a time overlap, for example of 5 minutes.

The present invention also relates to a prediction system for predicting process deviations in a process engineering system, which is designed and set up to carry out the method according to the invention for predicting process deviations in a process engineering system, for example in a painting system.

The present invention also relates to an industrial control system comprising a prediction system according to the invention.

The method according to the invention for predicting process deviations preferably has one or more of the features and / or advantages described in connection with the method according to the invention for error analysis.

The method according to the invention for error analysis preferably also has one or more of the features and / or advantages described in connection with the method according to the invention for predicting process deviations.

The present invention also relates to a method for anomaly and / or error detection in a process engineering system, for example in a painting system.

The present invention is based on the further object of providing a method for anomaly and / or error detection in a process plant, for example in a painting plant, by means of which anomalies and / or error situations can be easily and reliably detected.

This object is achieved by a method for anomaly and / or error detection in a process plant, for example in a paint shop.

• - automatisches Erstellen eines Anomalie- und/oder Fehlermodells der verfahrenstechnischen Anlage, welches Informationen über die Auftretenswahrscheinlichkeit von Prozesswerten umfasst;
• - automatisches Einlesen von Prozesswerten der verfahrenstechnischen Anlage im Betrieb derselben;
• - automatische Erkennung einer Anomalie und/oder einer Fehlersituation durch Ermitteln einer Auftretenswahrscheinlichkeit mittels des Anomalie- und/oder Fehlermodells auf Basis der eingelesenen Prozesswerte der verfahrenstechnischen Anlage und durch Überprüfen der Auftretenswahrscheinlichkeit auf einen Grenzwert.

The method for anomaly and / or error detection in a process plant, for example in a paint shop, preferably includes the following:

• - Automatic creation of an anomaly and / or failure model of the process engineering system, which includes information about the probability of occurrence of process values;
• - automatic reading in of process values of the process engineering plant during operation of the same;
• - Automatic detection of an anomaly and / or an error situation by determining a probability of occurrence by means of the anomaly and / or error model on the basis of the read-in process values of the procedural plant and by checking the probability of occurrence for a limit value.

The method for anomaly and / or error detection can preferably be used to identify error situations, that is to say defects and / or failures in components, sensors and / or actuators.

With the method for anomaly and / or error detection in a process engineering system, a normal state of the process engineering system can preferably be determined in an automated manner.

In particular, static and / or dynamic relationships in the process engineering system can be described by means of the anomaly and / or error model.

In the context of this description and the appended claims, an anomaly is understood to mean, in particular, a deviation of a process value from a normal state.

The anomaly and / or error model preferably includes a structure graph.

The structure graph includes, in particular, several cliques, relationships between nodes of a respective clique preferably being described by a probability density function.

In each case, a clique of the structure graph is used to describe relationships for sensors and / or actuators of the process engineering system.

An anomaly is preferably recognized when a limit value for the probability of occurrence of a process value in a clique of a structure graph of the anomaly and / or error model is not reached.

It can be beneficial if a recognized anomaly is graphically displayed with anomalous process variables for a user.

• - das Anomalie- und/oder Fehlermodell Strukturdaten umfasst, welche Informationen über eine Prozessstruktur in der verfahrenstechnischen Anlage enthalten, und/oder dass
• - das Anomalie- und/oder Fehlermodell Parametrierungsdaten umfasst, welche Informationen über Zusammenhänge zwischen Prozesswerten der verfahrenstechnischen Anlage enthalten.

In one embodiment of the method for anomaly and / or error detection, it is provided that

• the anomaly and / or error model comprises structural data which contain information about a process structure in the process engineering plant, and / or that
• - The anomaly and / or error model comprises parameterization data which contain information about relationships between process values of the process engineering plant.

The structural data include, in particular, information about relationships between sensors and / or actuators in the process engineering system.

The parameterization data include, in particular, information about the probability of occurrence of process values.

In particular, structural data and / or parameterization data are used to create the anomaly and / or error model.

• - Strukturidentifikation zur Ermittlung einer Prozessstruktur der verfahrenstechnischen Anlage;
• - Ermittlung von Kausalitäten in der ermittelten Prozessstruktur der verfahrenstechnischen Anlage;
• - Strukturparametrierung der Zusammenhänge in der ermittelten Prozessstruktur der verfahrenstechnischen Anlage.

In one embodiment of the method for anomaly and / or error detection, it is provided that one or more of the following steps are carried out to create the anomaly and / or error model:

• - Structure identification to determine a process structure of the process plant;
• - Determination of causalities in the determined process structure of the procedural plant;
• - Structural parameterization of the interrelationships in the determined process structure of the process plant.

The anomaly and / or error model preferably includes structural information, causality information and / or structural parameterization information.

The structure parameterization can preferably be facilitated by means of the structure identification.

In particular, by means of the structure identification, a parameterization effort and thus in particular a computational effort for the structure parameterization can be reduced.

In one embodiment of the method for anomaly and / or error detection, it is provided that a structure graph is determined during the structure identification to determine a process structure of the process engineering system, which in particular depicts relationships in the process engineering system.

The structure graph preferably comprises several nodes and several edges connecting the nodes to one another in pairs.

The structure graph preferably comprises several cliques.

It can be favorable if relationships are determined in the determined structure graph by means of the structure identification.

• - eines maschinellen Lernverfahrens;
• - Expertenwissen;
• - bekannten Schaltplänen und/oder Verfahrensschemata;
• - Bezeichnungen in einem Nummerierungssystem der verfahrenstechnischen Anlage.

In one embodiment of the method for anomaly and / or error detection provides that the structure graph is determined using one or more of the following:

• - a machine learning process;
• - expert knowledge;
• - known circuit diagrams and / or process schemes;
• - Designations in a numbering system of the process plant.

The determination of the structure graph by means of a machine learning method is preferably carried out using correlation measures by means of which non-linear correlations can be reproduced, for example by means of transinformation (“mutual information”).

In the context of this description and the appended claims, expert knowledge is understood to mean, for example, knowledge about relationships between sensors in the process.

Edges between nodes of the structure graph can preferably be excluded by a pre-configuration of the structure graph using information from expert knowledge, known circuit diagrams and / or process schemes. In particular, a computational effort for determining the structure graph can be reduced.

Process values are preferably clearly identified by means of the numbering system (“plant numbering system”).

It can therefore be beneficial if the structure graph is determined using the unique designation of the process values.

In particular, it is conceivable that the structure graph determined by means of a machine learning process is checked for plausibility by means of expert knowledge, by means of known circuit diagrams and / or process schemes and / or by means of the designations in the numbering system of the process engineering plant.

In one embodiment of the method for anomaly and / or error detection, it is provided that the procedural system for structure identification, in particular for determining the structure graph, is stimulated with test signals.

Anomalies and / or error situations are preferably generated in a targeted manner during the excitation with test signals.

Test signals are generated in particular taking design data into account. In particular, limits for the test signals can be specified on the basis of the design data, for example a maximum amplitude of the manipulated variable jumps can be specified when specifying step functions.

• - Sensortyp (Temperatursensor, Durchflusssensor, Ventilstellungssensor, Drucksensor etc.) und/oder Aktortyp (Ventil, Ventilator, Klappe, E-Motor);
• - zulässige Wertebereiche von Sensoren und/oder Aktoren;
• - Signaltyp von Sensor und/oder Aktor (float, integer).

In the context of this description and the appended claims, design data is understood to mean in particular one or more of the following information:

• - Sensor type (temperature sensor, flow sensor, valve position sensor, pressure sensor, etc.) and / or actuator type (valve, fan, flap, electric motor);
• - Permissible value ranges of sensors and / or actuators;
• - Signal type from sensor and / or actuator (float, integer).

In particular, the process engineering system is dynamically excited by means of the test signals.

The test signals are, in particular, signals by means of which manipulated variables can be changed in the process engineering system. For example, manipulated variables of valves and / or pumps of the process engineering system are changed by means of the test signals.

• - bei der Anregung der verfahrenstechnischen Anlage mit Testsignalen generierte Systemeingangssignale und Systemausgangssignale;
• - Expertenwissen;
• - bekannte Schaltpläne und/oder Verfahrensschemata;
• - Bezeichnungen in einem Nummerierungssystem der verfahrenstechnischen Anlage.

In one embodiment of the method for anomaly and / or error detection, it is provided that the determination of causalities in the determined process structure of the procedural system takes place using one or more of the following:

• - System input signals and system output signals generated when the process plant is excited with test signals;
• - expert knowledge;
• - known circuit diagrams and / or process schemes;
• - Designations in a numbering system of the process plant.

Causalities in the ascertained process structure are derived, for example, from system input signals and system output signals of the process engineering system ascertained when the process engineering system is excited with test signals, for example based on the respective temporal course of the system input signals and the system output signals.

As an alternative or in addition to this, it is conceivable that causalities from system input signals and System output signals are derived by means of methods for causal inference.

In the context of this description and the appended claims, causalities are understood to mean in particular directions of causality, that is to say “arrow directions”, in the structure graph determined.

Preferably, by means of the determined causalities in the determined process structure or in the determined structure graph, the process values causing a recognized anomaly can be found.

• - Verfahren zur Ermittlung von Wahrscheinlichkeitsdichtefunktionen, insbesondere Gaußsche Mischmodelle;
• - bekannte physikalische Zusammenhänge zwischen Prozesswerten;
• - physikalische Kennfelder von Funktionselementen der verfahrenstechnischen Anlage, beispielsweise Kennfelder von Ventilen.

In one embodiment of the method for anomaly and / or error detection, it is provided that one or more of the following are used to parameterize the structure of the relationships in the determined process structure of the procedural plant:

• - Method for determining probability density functions, in particular Gaussian mixed models;
• - known physical relationships between process values;
• - physical maps of functional elements of the process engineering system, for example maps of valves.

The structure parameterization is preferably carried out using methods for determining probability density functions, in particular using Gaussian mixed models.

It can be beneficial if relationships between two variables of the functional element can be described by means of physical characteristics maps of functional elements of the process engineering system.

A relationship between a valve position and a volume flow can, for example, be described by means of a known valve characteristics map of a valve.

In one embodiment of the method for anomaly and / or error detection, it is provided that, for structural parameterization using methods for determining probability density functions, in particular using Gaussian mixed models, data from the regular operation of the process engineering system and / or through an excitation of the process engineering system data obtained by means of test signals are used.

For example, manipulated, measured and / or controlled variables stored in a database are used for structure parameterization using methods for determining probability density functions.

For structural parameterization using methods for determining probability density functions, data from the ongoing operation of the process engineering system are preferably used, which are stored over a period of at least 2 weeks, preferably at least 4 weeks, for example at least 8 weeks.

In one embodiment of the method for anomaly and / or error detection, it is provided that the data used for structural parameterization using methods for determining probability density functions, in particular using Gaussian mixed models, are preprocessed prior to structural parameterization.

During preprocessing, data from the regular operation of the plant are preferably excluded using alarms and status bits that describe the state of the process plant, which are not assigned to operational or production-ready operating states of the process plant (e.g. shut down plant, maintenance phases, etc.).

It can also be favorable if data from the regular operation of the system are preprocessed by filtering, for example by means of low-pass filters and / or by means of Butterworth filters.

Data from regular operation are preferably interpolated to a uniform time step size.

In one embodiment of the method for anomaly and / or error detection, it is provided that when creating the anomaly and / or error model, a limit value for the probability of occurrence of a process value is established, an anomaly being detected if the limit value is not reached.

The limit value is preferably established by means of a non-linear optimization method, for example by means of the Nelder-Mead method.

As an alternative or in addition to this, it is conceivable that the limit value is determined by means of quantiles.

Limit values for the probability of occurrence of the process values can preferably be optimized, for example, by specifying a "false positive rate".

The limit values are preferably adapted after the anomaly and / or error model has been created for the first time, in particular in the case of an excessive number of false alarms.

In one embodiment of the method for anomaly and / or error detection, it is provided that the method for anomaly and / or error detection is used to identify a cause of a detected anomaly and / or a detected error situation.

In particular, the cause of the error can be identified by means of the structure graph of the anomaly and / or error model.

The structure graph is preferably visualized for a user to identify the anomaly and / or the error situation and / or to identify the cause of the error.

In particular, a “root cause analysis” can be implemented using the structure graph. In particular, abnormal process values can be identified within a process structure of the process engineering plant.

A recognized anomaly can preferably be flagged by a user as an error situation or false alarm.

Error situations are stored, in particular, in an error database.

• - Vorbehandlungsstation;
• - Station zur kathodischen Tauchlackierung;
• - Trocknerstationen;
• - industrielle Zuluftanlage;
• - Lackierroboter.

In one embodiment of the method for anomaly and / or error detection, it is provided that the procedural system comprises one or more of the following treatment stations of a painting system or is formed by them:

• - pre-treatment station;
• - Station for cathodic dip painting;
• - dryer stations;
• - industrial air supply system;
• - painting robots.

The present invention also relates to an anomaly and / or error detection system for anomaly and / or error detection, which is designed and set up to carry out the method according to the invention for anomaly and / or error detection in a process plant, for example in a paint shop.

The anomaly and / or error detection system forms, in particular, a reporting system by means of which an error situation in the process engineering system can be automatically identified.

The present invention also relates to an industrial control system which comprises an anomaly and / or error detection system according to the invention.

The method according to the invention for anomaly and / or error detection preferably has one or more of the features and / or advantages described in connection with the method according to the invention for error analysis and / or the method according to the invention for predicting process deviations.

The method according to the invention for error analysis and / or the method according to the invention for predicting process deviations preferably have one or more of the features and / or advantages described in connection with the method according to the invention for anomaly and / or error detection.

Further features and / or advantages of the invention are the subject matter of the following description and the graphic representation of exemplary embodiments.

• 1 eine schematische Darstellung einer verfahrenstechnischen Anlage und eines industriellen Steuerungssystems;
• 2 eine schematische Darstellung einer verfahrenstechnischen Anlage, insbesondere einer Lackieranlage;
• 3 eine schematische Darstellung einer industriellen Zuluftanlage;
• 4 die schematische Darstellung der industriellen Zuluftanlage aus 3 beim Auftreten einer Fehlersituation;
• 5 die schematische Darstellung der industriellen Zuluftanlage aus 3 beim Auftreten einer weiteren Fehlersituation;
• 6 die schematische Darstellung der industriellen Zuluftanlage aus 3 beim Auftreten einer weiteren Fehlersituation;
• 7 eine weitere schematische Darstellung einer industriellen Zuluftanlage;
• 8 die schematische Darstellung der industriellen Zuluftanlage aus 7 in einem Betriebszustand ohne Prozessabweichung;
• 9 die schematische Darstellung der industriellen Zuluftanlage aus 7 in einem Betriebszustand mit einer Prozessabweichung aufgrund von sich ändernden Umgebungsbedingungen;
• 10 die schematische Darstellung der industriellen Zuluftanlage aus 7 in einem Betriebszustand mit einer Prozessabweichung aufgrund des Einschaltens eines Wärmerückgewinnungssystems;
• 11 die schematische Darstellung der industriellen Zuluftanlage aus 7 in einem Betriebszustand mit einer Prozessabweichung aufgrund des Ausfalls eines Ventils;
• 12 eine schematische Darstellung von in Prädiktionsdatensätze zusammengefassten Prozesswerten;
• 13 eine schematische Darstellung der Prädiktionsdatensätze aus 12, welche als Prädiktionsdatensätze mit Prozessabweichungen und als Prädiktionsdatensätze ohne Prozessabweichung gekennzeichnet sind;
• 14 eine schematische Darstellung einer Vorbehandlungsstation;
• 15 eine schematische Darstellung von Verfahrensschritten zur Erstellung eines Anomalie- und/oder Fehlermodells der Vorbehandl ungsstation;
• 16 eine schematische Darstellung eines Graphen mit einer aus der Vorbehandlungsstation aus 14 abgeleiteten Prozessstruktur;
• 17 eine Clique eines Faktorgraphen;
• 18 ein Modell eines funktionalen Zusammenhangs in der Clique aus 17; und
• 19 eine der Clique aus 17 entsprechende Clique, welche durch Zuweisung einer Fehlerursache um einen Knoten erweitert ist.

In the drawings show:

• 1 a schematic representation of a process plant and an industrial control system;
• 2 a schematic representation of a process engineering system, in particular a painting system;
• 3rd a schematic representation of an industrial air supply system;
• 4th the schematic representation of the industrial air supply system 3rd when an error situation occurs;
• 5 the schematic representation of the industrial air supply system 3rd if another error situation occurs;
• 6th the schematic representation of the industrial air supply system 3rd if another error situation occurs;
• 7th another schematic representation of an industrial air supply system;
• 8th the schematic representation of the industrial air supply system 7th in an operating state without process deviation;
• 9 the schematic representation of the industrial air supply system 7th in one Operating status with a process deviation due to changing environmental conditions;
• 10 the schematic representation of the industrial air supply system 7th in an operating state with a process deviation due to the activation of a heat recovery system;
• 11 the schematic representation of the industrial air supply system 7th in an operating state with a process deviation due to the failure of a valve;
• 12th a schematic representation of process values summarized in prediction data sets;
• 13th a schematic representation of the prediction data sets 12th , which are identified as prediction data sets with process deviations and as prediction data sets without process deviations;
• 14th a schematic representation of a pretreatment station;
• 15th a schematic representation of method steps for creating an anomaly and / or failure model of the pretreatment station;
• 16 a schematic representation of a graph with one from the pretreatment station 14th derived process structure;
• 17th a clique of a factor graph;
• 18th a model of a functional relationship in the clique 17th ; and
• 19th one of the clique out 17th corresponding clique, which is expanded by a node by assigning a cause of the error.

Identical or functionally equivalent elements are provided with the same reference symbols in all figures.

1 shows an industrial control system, designated as a whole by 100, for a process engineering plant 101 .

The process plant 101 is for example a paint shop 104 , which in particular in 2 is shown.

Based on 1 to 6th In particular, a method for error analysis in the process engineering system is used 101 , especially in the paint shop 102 explained.

Based on 1 , 2 and 7th to 13th In particular, a method for predicting process deviations in the process engineering system is being developed 101 , especially in the paint shop 102 explained.

Based on 1 , 2 and 14th to 19th In particular, a method for anomaly and / or error detection in the process plant is used 101 , especially in the paint shop 102 explained.

In the 2 illustrated process plant 101 , especially the paint shop 102 , preferably comprises several treatment stations 104 for the treatment of workpieces 106 , in particular for treating vehicle bodies 108 .

The treatment stations 104 are at the in 2 illustrated embodiment of the painting system 102 in particular chained together and therefore form a painting line 110 .

For the treatment of workpieces 106 , especially for painting vehicle bodies 108 , run through the workpieces 106 the treatment stations 104 preferably one after the other.

For example, it is conceivable that a workpiece 106 successive treatment stations 104 runs in the specified order.

One workpiece 106 is in a pre-treatment station 112 pre-treated and from the pre-treatment station 112 to a station for cathodic dip painting 114 promoted.

From the station to cathodic dip painting 114 becomes the workpiece 106 after applying a coating thereon to a drying station 116 after the station for cathodic dip painting 114 promoted.

After the drying of the cathodic dip painting in the station 114 on the workpiece 106 applied coating in the drying station 116 becomes the workpiece 106 preferably in a base coat cabin 118 promoted, in which turn a coating on the workpiece 106 is applied.

After applying the coating in the base coat booth 118 becomes the workpiece 106 preferably in a base coat drying station 120 promoted.

After drying in the base coat booth 118 on the workpiece 106 applied coating in the base coat drying station 120 becomes the workpiece 106 preferably in a clear coat booth 122 promoted, in which one further coating on the workpiece 106 is applied.

After applying the coating in the clear coat booth 122 becomes the workpiece 106 preferably a clear coat drying station 124 fed.

After drying in the clear coat booth 122 on the workpiece 106 applied coating in the Clear-Coat drying station 124 becomes the workpiece 106 preferably a control station 126 fed at the end of the production process.

In the control station 126 a quality control is preferably carried out by a quality inspector, for example by means of a visual inspection.

The process plant 101 , especially the paint shop 102 , preferably further comprises an industrial air supply system 128 for conditioning the air, for example in the base coat cabin 118 and / or the Clear Coat booth 122 is fed.

By means of the industrial air supply system 128 is preferably a temperature and / or a relative humidity of the air in the base coat cabin 118 and / or the Clear Coat booth 122 is supplied, adjustable.

Using the industrial control system 100 is preferably a production process, in particular a painting process, in treatment stations 104 the process plant 101 , especially the paint shop 102 , controllable.

Preferably the industrial control system comprises 100 plus a process control system 130 .

This in 1 illustrated industrial control system 100 preferably further comprises a database 132 .

Database 132 of the industrial control system 100 preferably comprises a process database 134 and a bug tracking system 136 .

It can also be beneficial if the industrial control system 100 a reporting system 138 and an analysis system 140 includes.

The industrial control system 100 preferably further comprises a visualization system 142 , by means of which information can be visualized for a user.

The visualization system preferably comprises 142 one or more screens on which information can be displayed.

The analysis system 140 preferably comprises a fault analysis system 144 or is formed by this.

It can also be beneficial if the reporting system 138 a prediction system 146 to predict process deviations in the process plant 101 includes or is formed by this.

Alternatively or in addition to this, it is conceivable that the reporting system 138 an anomaly and / or error detection system 148 includes.

The failure analysis system 144 is set up and designed in particular, based on the 1 to 6th explained procedures for error analysis in the process plant 101 perform.

The prediction system 146 is set up and designed in particular, based on the 1 , 2 and 7th to 13th explained procedures for predicting process deviations in the process plant 101 perform.

The anomaly and / or error detection system 148 is designed in particular that based on the 1 , 2 and 14th to 19th explained procedures for anomaly and / or error detection in the process plant 101 perform.

The ones in the 3rd to 6th illustrated industrial air supply system 128 preferably comprises several conditioning modules 150 , for example a preheating module 154 , a cooling module 156 , a post-heating module 158 and / or a humidifier module 160 .

An example is the industrial air intake system 128 the paint shop 102 a functional unit, wherein a conditioning module 150 the supply air system 128 is a functional group and where a circulation pump 152 the supply air system is a functional element (cf. 3rd to 6th ).

In addition to the circulation pumps 152 the preheating module, the cooling module 156 and the post-heating module 158 includes the industrial air supply system 128 preferably also a humidifier pump 153 of the humidifier module 160 .

It can also be beneficial if the supply air system 128 a fan 162 includes.

The supply air system 128 preferably further comprises a heat recovery system 164 for heat recovery.

The supply air system 128 is preferably an air stream from an environment thereof 165 feedable.

One by means of the supply air system 128 conditioned airflow 167 is preferably the base coat booth 118 and / or the Clear Coat booth 122 feedable.

The air supply system preferably comprises sensors, not shown in the drawings, by means of which process values can be recorded.

• - Außentemperatur 166;
• - Außenfeuchte 168;
• - Temperatur der mittels der industriellen Zuluftanlage konditionierten Luft 170;
• - Feuchte der mittels der industriellen Zuluftanlage konditionierten Luft 172;
• - Volumenströme 174, 176, 178 in den Konditionierungsmodulen 150;
• - Ventilstellungen 180, 182, 184 von Ventilen 181, 183, 185 in den Konditionierungsmodulen 150.

For example, the following process values can be recorded by means of the sensors, which are in the 3rd to 6th are preferably each identified by a reference symbol:

• - outside temperature 166 ;
• - outside humidity 168 ;
• - Temperature of the air conditioned by means of the industrial supply air system 170 ;
• - Humidity of the air conditioned by means of the industrial supply air system 172 ;
• - volume flows 174 , 176 , 178 in the conditioning modules 150 ;
• - valve positions 180 , 182 , 184 of valves 181 , 183 , 185 in the conditioning modules 150 .

It can also be beneficial if a rotational frequency 193 the humidifier pump 153 and a rotation frequency 195 of the fan 194 are recorded.

Preferably the process values 166 to 184 in the process database 134 saved.

• - Pumpenstatus 186, 188, 190 der Umlaufpumpen 152 der Konditionierungsmodule 150 und Pumpenstatus 192 der Befeuchterpumpe 153 (an/aus);
• - Ventilatorstatus 194 (an/aus);
• - Ventilstatus 196, 198, 200 der Konditionierungsmodule 150 (geschlossen/geöffnet) ;
• - Status 202 des Wärmerückgewinnungssystems 164 (an/aus).

Furthermore, it can be provided that the following status variables are recorded, which are in the 3rd to 6th are preferably also identified by a reference symbol:

• - pump status 186 , 188 , 190 the circulation pumps 152 the conditioning modules 150 and pump status 192 the humidifier pump 153 (On off);
• - fan status 194 (On off);
• - valve status 196 , 198 , 200 the conditioning modules 150 (closed open) ;
• - status 202 the heat recovery system 164 (On off).

The status variables are also preferred 186 to 202 in the process database 134 saved.

The procedure for error analysis in the process plant 101 is preferably now made with reference to the 3rd to 6th explained.

The industrial air supply system 128 forms the process engineering system in particular 101 .

In the following, various example situations are described from which the functionality of the fault analysis system 144 can be seen.

• (Ventilleckage)

Example situation 1 (cf. 4th ):

• (Valve leakage)

A valve leak on the preheating module 154 occurs. A volume flow 174> 0 is measured. The pump status 186 is "off" and the valve status 196 of the control valve is "closed".

The error situation is in the message system as logic (pump status 186 = "off"; volume flow 174> 0 and valve status 196 "Closed"). The message system has therefore preferably saved the process and status values as a pre-link.

• 1) Die Meldung wird einem Benutzer mittels des Visualisierungssystems 142, beispielsweise mittels eines Bildschirms des Visualisierungssystems 142, angezeigt.
• 2) Ein Benutzer möchte die Situation analysieren und öffnet ein Diagnosefenster.
• 3) Im Diagnosefenster werden dem Benutzer der mit der Meldung vorverknüpfte Prozesswert 174 sowie die Statusgrößen 186 und 196 angezeigt. Das Fehleranalysesystem 144 bekommt diese Information vorzugsweise direkt aus dem Meldesystem 138.
• 4) Eine Priorisierung der mit der Fehlersituation verknüpften Prozesswerte wird vorzugsweise nicht vorgenommen, da der Fehlersituation nur der Prozesswert 174 zugeordnet ist.
• 5) Der Benutzer speichert die Fehlersituation mit der Verknüpfung in der Fehlerdatenbank 136 ab.
• 6) Bei erneutem Auftreten der Fehlersituation „Ventilleckage“ an dem gleichen oder vergleichbaren Ventil wird dem Benutzer vorzugsweise die vergleichbare Fehlersituation angezeigt.

The error analysis steps when a message occurs due to valve leakage are preferably as follows:

• 1) The message is sent to a user by means of the visualization system 142 , for example by means of a screen of the visualization system 142 displayed.
• 2) A user wants to analyze the situation and opens a diagnostic window.
• 3) The process value linked to the message is shown to the user in the diagnostics window 174 as well as the status sizes 186 and 196 displayed. The failure analysis system 144 preferably receives this information directly from the reporting system 138 .
• 4) The process values linked to the error situation are preferably not prioritized, since only the process value is the error situation 174 assigned.
• 5) The user saves the error situation with the link in the error database 136 from.
• 6) If the error situation “valve leakage” occurs again on the same or comparable valve, the user is preferably shown the comparable error situation.

• (Überschreiten der Temperatur 170 der mittels der industriellen Zuluftanlage 128 konditionierten Luft aufgrund einer zu hohen Außentemperatur 166)

Example situation 2 (cf. 5 ):

• (Exceeding the temperature 170 by means of the industrial air supply system 128 conditioned air due to too high an outside temperature 166 )

The temperature 170 by means of the industrial air supply system 128 conditioned air is exceeded because the outside temperature 166 outside a design window of the industrial air supply system 128 is located.

The message is displayed when leaving a specified process window for the temperature 170 by means of the industrial air supply system 128 conditioned air is generated and is used by the reporting system 138 to the visualization system 142 reported.

The message is with the value of the temperature 170 by means of the industrial air supply system 128 conditioned air linked. However, the message is not related to the outside temperature 166 connected.

• 1) Die Meldung wird einem Benutzer mittels des Visualisierungssystems 142, beispielsweise mittels eines Bildschirms des Visualisierungssystems 142, angezeigt.
• 2) Ein Benutzer möchte die Situation analysieren und öffnet ein Diagnosefenster.
• 3) Im Diagnosefenster wird dem Benutzer der mit der Meldung verknüpfte Prozesswert 170 angezeigt.
• 4) Eine Priorisierung der mit der Fehlersituation verknüpften Prozesswerte wird vorzugsweise nicht vorgenommen, da der Fehlersituation nur der Prozesswert 166 zugeordnet ist.
• 5) Zusätzlich werden dem Benutzer vorzugsweise folgende Prozesswerte vorgeschlagen:
  • - Feuchte der mittels der industriellen Zuluftanlage konditionierten Luft 172 (prozesskritische Größe, zeigt Anomalie im Verhalten);
  • - Außentemperatur 166 (zeigt Anomalie im Verhalten);
  • - Außenfeuchte 168 (zeigt Anomalie im Verhalten).
• 6) Der Benutzer wählt die vorgeschlagenen Prozesswerte aus und fügt sie der Fehlersituation hinzu.
• 7) Der Benutzer kann die Ursache der Temperaturabweichung aus dem Analysesystem 140, insbesondere aus dem Fehleranalysesystem 144, direkt erfassen, da ihm die relevanten Prozesswerte vorgeschlagen werden.
• 8) Der Benutzer fügt eine Dokumentation zu der Fehlersituation mit einem Fehlerbehebungsvorschlag hinzu.
• 9) Der Benutzer speichert die Fehlersituation mit der Verknüpfung und den Dokumenten in der Fehlerdatenbank ab.
• 10) Das Fehleranalysesystem 144 erfasst vorzugsweise neben einer Fehler-ID, einer Fehlerklassifikation (Temperaturerhöhung), einem Fehlerort (Ausblasteil der industriellen Zuluftanlage 128) auch Referenzen (IDs) der Prozessgrößen 170, 172, 166, 168 in der priorisierten Reihenfolge und eine Menge von Merkmalen (z.B. Mittelwerte, Minimum, Maximum, Streuung der Prozessgrößen während des Auftretens der Fehlersituation) sowie die Zeitdauer vom Auftreten der Fehlersituation bis zum Speicherzeitpunkt beziehungsweise Ende der Fehlersituation.

The error analysis steps when the message occurs due to the temperature change are preferably as follows:

• 1) The message is sent to a user by means of the visualization system 142 , for example by means of a screen of the visualization system 142 displayed.
• 2) A user wants to analyze the situation and opens a diagnostic window.
• 3) The process value linked to the message is shown to the user in the diagnostics window 170 displayed.
• 4) The process values linked to the error situation are preferably not prioritized, since only the process value is the error situation 166 assigned.
• 5) In addition, the following process values are suggested to the user:
  • - Humidity of the air conditioned by means of the industrial supply air system 172 (process-critical variable, shows anomaly in behavior);
  • - outside temperature 166 (shows anomaly in behavior);
  • - outside humidity 168 (shows anomaly in behavior).
• 6) The user selects the proposed process values and adds them to the error situation.
• 7) The user can determine the cause of the temperature deviation from the analysis system 140 , especially from the error analysis system 144 , record directly, as the relevant process values are suggested to him.
• 8) The user adds documentation on the error situation with a suggested remedy.
• 9) The user saves the error situation with the link and the documents in the error database.
• 10) The failure analysis system 144 preferably detects an error ID, an error classification (temperature increase), an error location (exhaust part of the industrial air supply system) 128 ) also references (IDs) of the process variables 170 , 172 , 166 , 168 in the prioritized order and a set of features (e.g. mean values, minimum, maximum, scatter of the process variables during the occurrence of the error situation) as well as the time from the occurrence of the error situation to the time of storage or the end of the error situation.

• (Überschreiten der Temperatur 170 der mittels der industriellen Zuluftanlage konditionierten Luft aufgrund einer zu hohen Außentemperatur)

Example situation 3 (cf. 5 ):

• (Exceeding the temperature 170 the air conditioned by means of the industrial supply air system due to an excessively high outside temperature)

The temperature 170 the air conditioned by the industrial supply air system is exceeded again due to the outside temperature. The error pattern is similar to example situation 2.

• 1) Die Meldung wird einem Benutzer mittels des Visualisierungssystems 142, beispielsweise mittels eines Bildschirms des Visualisierungssystems 142, angezeigt.
• 2) Der Benutzer möchte die Fehlersituation analysieren und öffnet ein Diagnosefenster.
• 3) Im Diagnosefenster werden dem Benutzer die mit der Meldung verknüpften Prozesswerte: 170, 172, 166, 168 in beschriebener Reihenfolge angezeigt.
• 4) Eine Prozessliste und deren Priorisierung ergeben sich aus einer ähnlichen Fehlersituation. Die Ähnlichkeit zur Fehlersituation aus Beispielsituation 2 wird vom Fehleranalysesystem 144 über einen metrischen Abgleich der Prozesswerte bestimmt.
• 5) Die ähnliche Fehlersituation wird dem Benutzer angezeigt
• 6) Der Benutzer kann die Verläufe der Prozesswerte in der gezeigten priorisierten Reihenfolge sowie die Dokumente der ihm gezeigten ähnlichen Fehlersituation zur Lösungsfindung nutzen.
• 7) Der Benutzer speichert die Fehlersituation ab.

The analysis steps when the message due to the temperature change occurs are preferably as follows:

• 1) The message is sent to a user by means of the visualization system 142 , for example by means of a screen of the visualization system 142 displayed.
• 2) The user wants to analyze the error situation and opens a diagnosis window.
• 3) In the diagnostics window, the user is shown the process values linked to the message: 170, 172, 166, 168 in the order described.
• 4) A process list and its prioritization result from a similar error situation. The similarity to the error situation from example situation 2 is made by the error analysis system 144 determined via a metric comparison of the process values.
• 5) The similar error situation is displayed to the user
• 6) The user can view the courses of the process values in the prioritized order shown as well as the documents of the similar error situation shown to him to find a solution.
• 7) The user saves the error situation.

• (Störung in einem Versorgungssystem)

Example situation 4 (cf. 6th ):

• (Disruption in a supply system)

A burner in the preheating module 154 too little fuel gas is being supplied. The volume flow 174 sinks.

To get more fuel gas, the valve will 181 of the preheating module 154 further open, the valve position 180 changes.

The valve too 185 of the post-heating module 158 opens to the malfunction of the preheating module 154 to compensate. The valve position 184 changes.

The malfunction may be due to the low outside temperature 166 not be compensated and the temperature 170 the air conditioned by the industrial supply air system breaks in.

• 1) Die Meldung wird einem Benutzer mittels des Visualisierungssystems 142, beispielsweise mittels eines Bildschirms des Visualisierungssystems 142, angezeigt.
• 2) Der Benutzer möchte die Fehlersituation analysieren und öffnet ein Diagnosefenster.
• 3) Im Diagnosefenster wird dem Benutzer aufgrund der Vorpriorisierung im Meldesystem 138 der mit der Meldung verknüpfte Prozesswert 170 angezeigt.
• 4) Eine Priorisierung findet nicht statt.
• 5) Es werden folgende Prozesswerte vorgeschlagen:
  • - Feuchte der mittels der industriellen Zuluftanlage konditionierten Luft 172 (prozesskritische Größe, zeigt Anomalie im Verhalten);
  • - Volumenstrom 174, Ventilstellung 180, Ventilstellung 186 (abhängig von der Abweichung vom Normalzustand).
• 6) Die Fehlersituationen aus den Beispielsituationen 2 und 3 werden nicht als ähnlich eingestuft (unterschiedliches Signalverhalten, da hoher metrischer Abstand der Prozesswerte).
• 7) Die vorgeschlagenen Prozesswerte können der Fehlersituation hinzugefügt und in der Fehlerdatenbank abgespeichert werden.

The analysis steps when the message occurs due to the malfunction in the supply system are preferably as follows:

• 1) The message is sent to a user by means of the visualization system 142 , for example by means of a screen of the visualization system 142 displayed.
• 2) The user wants to analyze the error situation and opens a diagnosis window.
• 3) In the diagnosis window, the user is informed based on the pre-prioritization in the message system 138 the process value linked to the message 170 displayed.
• 4) There is no prioritization.
• 5) The following process values are suggested:
  • - Humidity of the air conditioned by means of the industrial supply air system 172 (process-critical variable, shows anomaly in behavior);
  • - volume flow 174 , Valve position 180 , Valve position 186 (depending on the deviation from the normal state).
• 6) The error situations from example situations 2 and 3 are not classified as similar (different signal behavior, as there is a high metric distance between the process values).
• 7) The proposed process values can be added to the error situation and saved in the error database.

The method for predicting process deviations in the process plant 101 is preferably now made with reference to the 7th to 13th explained.

• - Zielgrößen 204 der industriellen Zuluftanlage 128, insbesondere Temperatur 170 und relative Luftfeuchte 172 der mittels der industriellen Zuluftanlage 128 konditionierten Luft, insbesondere an einem Ausblasteil der industriellen Zuluftanlage 128;
• - Stellgrößen 206, insbesondere Ventilstellungen 180, 182, 184 von Ventilen von Heiz- und/oder Kühlmodulen 154, 156, 158 der industriellen Zuluftanlage 128, Drehfrequenzen 193 von Pumpen 152, insbesondere der Befeuchterpumpe 153, und/oder Drehfrequenzen 195 von Ventilatoren 162;
• - interne Größen 208, insbesondere Vor- und/oder Rücklauftemperaturen 210 in den Heiz- und/oder Kühlmodulen 154, 156, 158 der industriellen Zuluftanlage 128 und/oder Luftkonditionen zwischen Konditionierungsmodulen 150;
• - gemessene Störgrößen 210, insbesondere Außentemperatur 166 und/oder relative Außenluftfeuchte 168 an einem Einblasteil der industriellen Zuluftanlage 128;
• - ungemessene Störgrößen 212; und/oder
• - Statusgrößen 214, insbesondere Befeuchterpumpe 153 (an/aus); manueller Modus für Pumpen 152, 153 (an/aus); Versorgungsventile 181, 183, 185, Ventilator 162 (an/aus).

If the process plant 101 an industrial air supply system 128 stored process values and / or status variables preferably include the following (cf. 7th ):

• - target values 204 the industrial air supply system 128 , especially temperature 170 and relative humidity 172 by means of the industrial air supply system 128 conditioned air, in particular at a blow-out part of the industrial air supply system 128 ;
• - manipulated variables 206 , especially valve positions 180 , 182 , 184 of valves of heating and / or cooling modules 154 , 156 , 158 the industrial air supply system 128 , Rotational frequencies 193 of pumps 152 , especially the humidifier pump 153 , and / or rotational frequencies 195 of fans 162 ;
• - internal sizes 208 , especially flow and / or return temperatures 210 in the heating and / or cooling modules 154 , 156 , 158 the industrial air supply system 128 and / or air conditions between conditioning modules 150 ;
• - measured disturbance variables 210 , especially outside temperature 166 and / or relative outside air humidity 168 on a blow-in part of the industrial air supply system 128 ;
• - unmeasured disturbances 212 ; and or
• - Status sizes 214 , especially humidifier pump 153 (On off); manual mode for pumps 152 , 153 (On off); Supply valves 181 , 183 , 185 , Fan 162 (On off).

In the following, various exemplary operating states are described from which the functionality of the prediction system 146 can be seen.

Exemplary operating state 1:

(without deviation)

8th shows a first exemplary operating state of the industrial air supply system 128 without process deviation. The first exemplary operating state thus preferably represents a positive case.

The outside temperature 166 and outside humidity 168 are not constant. The preheating module 154 and the humidifier module 160 are active.

A control system for the industrial air supply system 128 keeps the temperature 170 and the relative humidity 172 by means of the industrial air supply system 128 conditioned air at a constant value.

According to the status sizes 214 (Fan 162 = "On" and valve and pump mode = "automatic") is the industrial air supply system 128 operational.

Because of the constant temperature 170 and relative humidity 172 by means of the industrial air supply system 128 conditioned air is the industrial supply air system 128 preferably also ready for production.

Exemplary operating state 2:

(Increase in outside temperature 166 with a decrease in the relative outdoor humidity 168 above the design parameters; Influence of a measured disturbance variable 210 on target size 204 )

9 shows a second exemplary operating state of the industrial air supply system 128 when the outside temperature increases 166 with a decrease in the relative outdoor humidity 168 by a change in weather.

1. a) Die erhöhte Außentemperatur 166 nach dem Wetterumschwung ist außerhalb der Auslegungsparameter.
2. b) Die Leistung des bisher aktiven Vorheizers-Moduls 154 wird von der Regelung reduziert.
3. c) Da die Reduktion der Heizleistung nicht ausreicht wird von der Regelung das Ventil 183 des Kühlmoduls 156 geöffnet und damit die Kühlleistung erhöht.
4. d) Die Drehfrequenz 193 der Befeuchterpumpe 153 wird von der Regelung erhöht, um die abnehmende Außenfeuchte 168 auszugleichen.
5. e) Da die Kühlleistung aufgrund der zu geringen Auslegung des Kühlmoduls 156 nicht ausreicht, kommt es zu einer Abweichung der Temperatur 170 der mittels der industriellen Zuluftanlage 128 konditionierten Luft.

The process values show the following behavior:

1. a) The increased outside temperature 166 after the weather change is outside the design parameters.
2. b) The output of the previously active preheater module 154 is reduced by the regulation.
3. c) Since the reduction in the heating output is not sufficient, the control system closes the valve 183 of the cooling module 156 opened and thus the cooling capacity increased.
4. d) The rotational frequency 193 the humidifier pump 153 is increased by the regulation to the decreasing outdoor humidity 168 balance.
5. e) Since the cooling capacity due to the insufficient design of the cooling module 156 is not sufficient, there will be a deviation in temperature 170 by means of the industrial air supply system 128 conditioned air.

Leaving the specified process window for the temperature 170 is due to the inertia of the industrial air intake system 128 and the compensation of the regulation delayed.

Exemplary operating state 3:

(Changeover to winter operation with heat recovery; influence of an unmeasured disturbance variable 212 on target size 204 )

10 shows a third exemplary operating state of the industrial air supply system 128 when the heat recovery system is switched on 164 which the airflow 165 heated from waste heat in cold climatic conditions.

The activation of the heat recovery system 164 takes place through a manual valve, which is why the influence of the heat recovery through that of the heat recovery system 164 cannot be measured (unmeasured disturbance 212 ).

1. a) Die Heizleistung des Wärmerückgewinnungssystems 164 wird erhöht, der Wert ist nicht messbar.
2. b) Das Ventil 181 des Vorheiz-Moduls 154 schließt aufgrund der Erhöhung der Heizleistung.
3. c) Im Kühlmodul 156 öffnet das Ventil 183, um durch zusätzliche Kühlleistung die Temperatur zu halten.
4. d) Die Drehfrequenz 193 der Befeuchterpumpe 153 wird von der Regelung so angepasst, dass die Luftfeuchte 172 der mittels der industriellen Zuluftanlage 128 konditionierten Luft gehalten wird.
5. e) Es kommt zu einer Abweichung der Temperatur 170 der mittels der industriellen Zuluftanlage 128 konditionierten Luft, da das Kühlmodul 156 die Wärmezufuhr nicht schnell genug kompensieren kann. Die Abweichung erfolgt aufgrund der Trägheit der industriellen Zuluftanlage 128 und der Kompensation der Regelung verzögert.

The process values show the following behavior:

1. a) The heating capacity of the heat recovery system 164 is increased, the value cannot be measured.
2. b) The valve 181 of the preheating module 154 closes due to the increase in heating output.
3. c) In the cooling module 156 opens the valve 183 to keep the temperature through additional cooling.
4. d) The rotational frequency 193 the humidifier pump 153 is adjusted by the control so that the air humidity 172 by means of the industrial air supply system 128 conditioned air is maintained.
5. e) The temperature deviates 170 by means of the industrial air supply system 128 conditioned air as the cooling module 156 cannot compensate for the heat supply quickly enough. The deviation is due to the inertia of the industrial air supply system 128 and the compensation of the regulation delayed.

Exemplary operating state 4:

(Failure of the valve 181 of the preheater module 154 )

11 shows a fourth exemplary operating state of the industrial air supply system 128 with a failure of the valve 181 of the preheating module 154 .

1. a) Das Ventil 181 des Vorheiz-Moduls 154 schließt aufgrund eines Gerätefehlers, wodurch sich die Heizleistung verringert.
2. b) Das Ventil 185 des Nachheiz-Moduls 158 öffnet, um die fehlende Heizleistung zu kompensieren.
3. c) Da die Heizleistung des Nachheiz-Moduls 158 nicht ausreicht, kommt es zu einer Abweichung der Temperatur 170 der mittels der industriellen Zuluftanlage 128 konditionierten Luft.

The process values show the following behavior:

1. a) The valve 181 of the preheating module 154 closes due to a device fault, which reduces the heating output.
2. b) The valve 185 of the post-heating module 158 opens to compensate for the lack of heating power.
3. c) Since the heating output of the post-heating module 158 is not sufficient, there will be a deviation in temperature 170 by means of the industrial air supply system 128 conditioned air.

The method for predicting process deviations in the process plant 101 , in particular in the industrial air supply system, is explained below with regard to the operating states 1 to 4 described above.

The operating states 2 to 4 are preferably with process deviations in the operation of the industrial air supply system 128 by means of the method for predicting process deviations with a prediction horizon 216 for example about 15 minutes predictable.

The data basis for training a forecast model is considered to be a period of time in which the industrial air supply system 128 normally (> 80%) in ready-to-use operating state (see exemplary operating state 1).

The recorded data contain the exemplary operating states 2 to 4, preferably several times in each case. These can occur during operation or, alternatively, have been brought about intentionally, for example by closing a valve 181 , 183 , 185 .

The data are then preferably preprocessed and regularized, which for example 12th can be found. The valve position is an example there 180 of the valve 181 of the preheating module 154 as well as the temperature 170 and relative humidity 172 by means of the industrial air supply system 128 conditioned air applied.

The regularized data are for example in a 30 minute time window 218 divided, each with a time shift of, for example, 5 minutes.

The in time window 218 regularized data form, in particular, prediction data sets, in particular prediction data sets without process deviations 220 and prediction data sets with process deviation 222 (cf. 7th ).

• - Falls nein: entsprechende Prädiktionsdatensätze mit Prozessabweichung 222 werden verworfen und nicht zum Training des Vorhersagemodells verwendet.
• - Falls ja: entsprechende Prädiktionsdatensätze mit Prozessabweichung 222 kommen für das Training des Vorhersagemodells in Frage.

For the prediction data sets with process deviation 222 is based on the status sizes 214 checked whether the process plant 101 was ready for operation (e.g. fan 162 "On", conditioning modules 150 in automatic mode).

• - If no: corresponding prediction data sets with process deviation 222 are discarded and not used to train the predictive model.
• - If yes: corresponding prediction data sets with process deviation 222 are suitable for training the predictive model.

Due to a minimum time interval of, for example, one hour, in 7th only one prediction data set with process deviation 222 selected.

A selection of the prediction data sets without process deviations 220 preferably takes place analogously to the selection of the prediction data sets with process deviation 222 .

In the case of a minimum time interval of, for example, one hour, only one prediction data record is used without process deviations 220 selected for training the predictive model.

From the selected prediction data sets without process deviations 220 and from the selected prediction data sets with process deviations 222 features are then preferably extracted.

For example, statistical key figures, for example minimum, maximum, median, mean and / or the standard deviation, are used to extract the features. It can also be beneficial if linear regression coefficients are used to extract the features.

The prediction model is preferably trained on the basis of the extracted features of the selected prediction data sets without process deviations 220 and on the basis of the selected prediction data sets with process deviations 222 , in particular by means of a machine learning method, for example by means of gradient boosting.

By means of the trained forecast model, process deviations from production-critical process values in the industrial supply air systems are preferably identified 128 predicted on the basis of changing process values during operation of the industrial air supply system 128 .

The prediction model is explained in particular on the basis of the exemplary operating states 2 to 4:

Exemplary operating state 2:

The prediction model predicts a process deviation after the temperature increase has occurred. The basis for this are the measured disturbance variables 210 , especially outside temperature 166 and the outside humidity 168 , the response of the conditioning modules 150 as well as the course of the temperature 170 by means of the industrial air supply system 128 conditioned air at the discharge part.

Exemplary operating state 3:

The prediction model predicts a temperature increase after the heat recovery system has been switched on 170 by means of the industrial air supply system 128 conditioned air. The basis is the reaction of the conditioning modules 150 as well as the course of the temperature 170 by means of the industrial air supply system 128 conditioned air at the discharge part.

Exemplary operating state 4:

The predictive model says the temperature increase in temperature 170 by means of the industrial air supply system 128 conditioned air based on weather conditions and valve position 180 of the valve 181 of the preheating module 156 ahead.

The procedure for anomaly and / or error detection in the process plant 101 is preferably now made with reference to the 14th to 19th explained.

The method for anomaly and / or error detection can preferably be used to identify error situations, in particular defects and / or failures in components, sensors and / or actuators.

The pre-treatment station 112 forms here, for example, the process engineering system 101 .

The pre-treatment station 112 preferably comprises a pretreatment basin 224 in which work piece 106 , preferably vehicle bodies 108 , are pretreatable.

The pre-treatment station 112 preferably further comprises a first pump 226 , a second pump 228 , a heat exchanger 230 and a valve 232 .

The process values V62Point, S86, T95, T85, T15 and T05 receive their designation based on a unique designation in a numbering system of the process plant 101 .

The process values T95 , T85 , T15 and T05 represent in particular temperatures within the process plant 101 , especially within the pre-treatment station 112 , dar.

The process value S86 is a valve position of the valve 232 .

The process value V62Point is a volume flow.

An anomaly and / or error model is preferably used to carry out the method for anomaly and / or error detection 233 the process plant 101 , especially the pre-treatment station 112 , which contains information about the probability of occurrence of the above-mentioned process values (cf. 15th ).

• Zunächst werden Testsignale insbesondere unter Berücksichtigung von Auslegungsdaten 234 im Rahmen einer Testsignalgenerierung 236 erzeugt.

The creation of the anomaly and / or failure model 233 preferably takes place as follows:

• First of all, test signals are given, in particular, taking into account design data 234 as part of a test signal generation 236 generated.

In particular, due to the design data 234 Limits specified for the test signals, for example a maximum amplitude of manipulated variable jumps when specifying step functions.

• - Sensortyp (Temperatursensor, Durchflusssensor, Ventilstellung, Drucksensor etc.) und/oder Aktortyp (Ventil, Ventilator, Klappe, E-Motor);
• - zulässige Wertebereiche von Sensoren und/oder Aktoren;
• - Signaltyp von Sensor und/oder Aktor (float, integer).

The design data 234 include, for example, one or more of the following information:

• - Sensor type (temperature sensor, flow sensor, valve position, pressure sensor, etc.) and / or actuator type (valve, fan, flap, electric motor);
• - Permissible value ranges of sensors and / or actuators;
• - Signal type from sensor and / or actuator (float, integer).

The process plant 101 , especially the pre-treatment station 112 , is preferably excited dynamically by means of the test signals. This is in 15th with the reference number 238 marked. It is conceivable that anomalies and / or error situations are specifically generated during the excitation with test signals.

When stimulating the process plant 101 , especially the pre-treatment station 112 , with test signals are preferably system input signals 240 and system output signals 242 generated.

The system input signals 240 and the system output signals 242 are preferably in a test signal database 244 saved.

A structure identification is then preferably used 246 the process plant 101 , especially the pre-treatment station 112 , carried out. A structure graph is preferably used 247 the process plant 101 , especially the pre-treatment station 112 , determined (cf. 16 ).

The structure identification 246 , in particular the determination of the structure graph is preferably carried out using a machine learning method, preferably using correlation measures by means of which non-linear correlations can be reproduced, for example by means of transinformation (“mutual information”).

It can also be beneficial if for structure identification 246 Expert knowledge 248 is used, i.e. in particular knowledge about relationships in the process.

In this case, for example, edges between nodes of the structure graph to be determined are provided by a pre-configuration of the structure graph using information from expert knowledge, known circuit diagrams and / or process schemes 250 excludable. In particular, a computational effort for determining the structure graph can be reduced.

It can also be favorable if the structure graph is made using the unique designation of the process values based on a numbering system of the process engineering plant 101 , especially the pre-treatment station 112 , is determined, that is, on the basis of semantics 252 the designation of the process values.

In particular, it is conceivable that the structure graph ascertained by means of the machine learning method is implemented by means of expert knowledge 248 , using known circuit diagrams and / or process diagrams 250 and / or by means of the designations in the numbering system of the process plant 101 (Semantics 252 ) is checked for plausibility.

Causalities are then preferably used 254 determined in the determined process structure, in particular “arrow directions” in the determined structure graph.

Causalities 254 in the determined process structure, for example, when the process engineering system is excited 101 System input signals determined with test signals 240 and system output signals 242 the process plant 101 derived, for example based on the respective temporal course of the system input signals 240 and the system output signals 242 .

Alternatively or in addition to this, it is conceivable that causalities 254 off when the process plant is excited 101 System input signals determined with test signals 240 and system output signals 242 can be derived by means of causal inference methods.

To determine the causalities 254 are preferably also expert knowledge 248 , Procedural schemes 250 and / or the designations in the numbering system of the process plant 101 (Semantics 252 ) are used.

Preferably, by means of the determined causalities 254 the process values which are the cause of a recognized anomaly can be found in the ascertained process structure or in the ascertained structure graph.

After the structure identification 246 and / or the determination of the causalities 254 is preferably a structure parameterization 256 carried out.

By means of the structure identification 246 is preferably the structure parameterization 256 easable. A computational effort for the structure parameterization 256 is by means of the structure identification 246 preferably reducible.

The structure parameterization 256 preferably takes place using a method for determining probability density functions, in particular using Gaussian mixed models.

The structure parameterization 256 For example, for the common probability density function f1, the in 17th depicted clique 258 carried out using Gaussian mixed models (cf. 18th ).

Preferably for the structure parameterization 256 also expert knowledge 248 used.

For structure parameterization 256 For example, known physical relationships between process values and / or physical characteristics maps of functional elements of the process engineering system 101 , especially the pretreatment station. For example, a map of the valve 232 used.

It can also be beneficial if expert knowledge 248 about error situations for structure parameterization 256 is used.

By means of a known valve map of the valve 232 is, for example, a relationship between the valve position S86 and the volume flow rate V62 point can be described.

For structure parameterization 256 using methods for determining probability density functions, in particular using Gaussian mixed models, are preferably stored in an operating database 260 Stored data from the regular operation of the process plant 101 , especially the pre-treatment station 112 , and / or data from the test signal database 244 used.

For example, for structure parameterization 256 using methods for determining probability density functions, in particular in a database 244 , 260 stored manipulated, measured and / or controlled variables are used.

Preferably used for structure parameterization 256 using methods for the determination of probability density functions data from the ongoing operation of the process engineering plant 101 , in particular the pretreatment station, which are stored over a period of at least 2 weeks, preferably at least 4 weeks, for example at least 8 weeks.

The data are stored before the structure parameterization 256 preferably preprocessed.

During the preprocessing, data from the process engineering system are preferably used 101 which operating states of the process engineering plant that are not operational or ready for production 101 are assigned (for example shutdown system, maintenance phases, etc.), excluded, in particular on the basis of alarms and status bits, which indicate the state of the process engineering system 101 , especially the pre-treatment station 112 describe.

It can also be beneficial if data from the process engineering system 101 be preprocessed by filtering, for example by means of low-pass filters and / or by means of Butterworth filters.

The data are preferably also interpolated to a uniform time step size.

When creating the anomaly and / or failure model 233 is preferably used as part of a limit value optimization 264 a limit value for the probability of occurrence of a process value is specified.

The limit value for the probability of occurrence is preferably established in such a way that an anomaly is recognized if the limit value is not reached.

The limit value is preferably established by means of a non-linear optimization method, for example by means of the Nelder-Mead method.

As an alternative or in addition to this, it is conceivable that the limit value is determined by means of quantiles.

Limit values for the probability of occurrence of the process values can preferably be optimized, for example by specifying a “false positive rate”.

It is also conceivable that the limit values after the anomaly and / or error model has been created for the first time 233 adjusted, especially in the case of too high a number of false positives.

• Es kommt beispielsweise zu einem Ventilausfall des Ventils 232 und damit zu einer Abweichung der Sensorwerte vom abgebildeten Normalzustand in den einzelnen Cliquen.

An anomaly and / or error detection based on the anomaly and / or error model 233 preferably takes place as follows:

• For example, a valve failure of the valve occurs 232 and thus to a deviation of the sensor values from the depicted normal state in the individual cliques.

The probabilities of occurrence of the sensor values in the cliques are determined during the operation of the process plant 101 , especially in the operation of the pre-treatment station 112 , evaluated and if the calculated limit values are not reached, there are detected anomalies in the various cliques.

The valve failure of the valve 232 initially leads to an anomaly in the clique 258 the valve position S86 , with a message by the anomaly and / or error detection system 148 is issued.

In the further course of time, further anomalies arise due to the propagation of errors, which later also change the process-relevant variable, for example the pool temperature T35 of the pelvis 224 , influence.

• - Detektionszeitpunkt der Anomalie;
• - Clique(n) des Auftretens der Anomalie mit beteiligten Sensoren.

Preferably contains the message by the anomaly and / or error detection system 148 one or more of the following information:

• - Time of detection of the anomaly;
• - Clique (s) of the occurrence of the anomaly with involved sensors.

Due to the early detection of the anomaly and the notification to the user, if intervention is timed, a deviation of the process-relevant variable, that is to say the tank temperature, can preferably be detected T35 of the pelvis 224 , be prevented.

The user can then define a cause of failure, i.e. valve failure, for the occurrence of the anomaly.

By assigning the cause of the error, the clique 258 around a knot 266 expanded and the probability density function of the anomalous data integrated into the functional context (cf. 19th ).

After integrating the cause of the error, the method for anomaly and / or error detection is carried out as before. If an anomaly occurs, the probabilities of the defined error causes are also output.

• - Detektionszeitpunkt der Anomalie;
• - Clique(n) des Auftretens der Anomalie mit beteiligten Sensoren;
• - Wahrscheinlichkeiten der definierten Fehlerursachen.

A user now receives the notification from the anomaly and / or error detection system 148 one or more of the following information:

• - time of detection of the anomaly;
• - Clique (s) of the occurrence of the anomaly with involved sensors;
• - Probabilities of the defined causes of errors.

• Ausführungsform 1:
  • Verfahren zur Fehleranalyse in einer verfahrenstechnischen Anlage (101), beispielsweise in einer Lackieranlage (102), wobei das Verfahren Folgendes umfasst:
  • - insbesondere automatisches Erkennen einer Fehlersituation in der verfahrenstechnischen Anlage (101);
  • - Speichern eines Fehlersituationsdatensatzes zu der jeweiligen erkannten Fehlersituation in einer Fehlerdatenbank (136);
  • - automatische Ermittlung einer Fehlerursache für die Fehlersituation und/oder automatische Ermittlung von für die Fehlersituation relevanten Prozesswerten auf Basis des Fehlerdatensatzes einer jeweiligen erkannten Fehlersituation.
• Ausführungsform 2:
  • Verfahren nach Ausführungsform 1, dadurch gekennzeichnet, dass zur automatischen Ermittlung der Fehlerursache für die Fehlersituation und/oder der für die Fehlersituation relevanten Prozesswerte einer oder mehrere Prozesswerte automatisch basierend auf einem oder mehreren der folgenden Verknüpfungskriterien mit der Fehlersituation verknüpft werden:
    • - einer Vorverknüpfung aus einem Meldesystem;
    • - einer Zuordnung eines Prozesswerts zu demselben Anlagenteil der verfahrenstechnischen Anlage (101), in dem die Fehlersituation aufgetreten ist;
    • - einer Verknüpfung eines Prozesswerts mit einer historischen Fehlersituation aufgrund einer aktiven Auswahl eines Benutzers;
    • - einer aktiven Auswahl des Prozesswerts durch einen Benutzer.
• Ausführungsform 3:
  • Verfahren nach Ausführungsform 2, dadurch gekennzeichnet, dass zur automatischen Ermittlung der Fehlerursache für die Fehlersituation und/oder der für die Fehlersituation relevanten Prozesswerte eine automatische Priorisierung der mit der Fehlersituation verknüpften Prozesswerte basierend auf einem oder mehreren der folgenden Priorisierungskriterien automatisch durchgeführt wird:
    • - einer Prozessrelevanz der Prozesswerte;
    • - einer Position eines Prozesswerts oder eines den Prozesswert ermittelnden Sensors innerhalb der verfahrenstechnischen Anlage (101) ;
    • - einem Betrag einer Abweichung eines Prozesswerts von einem definierten Prozessfenster und/oder von einem Normalzustand;
    • - einer Priorisierung von historischen Prozesswerten in historischen Fehlersituationen;
    • - durch Übernehmen einer Priorisierung der Fehlerursache und/oder der Prozesswerte aus einem Meldesystem (138);
    • - einer Priorisierung durch einen Benutzer.
• Ausführungsform 4:
  • Verfahren nach einer der Ausführungsformen 1 bis 3, dadurch gekennzeichnet, dass zur automatischen Ermittlung der Fehlerursache für die Fehlersituation und/oder der für die Fehlersituation relevanten Prozesswerte weitere Fehlerursachen und/oder Prozesswerte vorgeschlagen werden, wobei das Vorschlagen anhand eines oder mehrerer der folgenden Vorschlagskriterien automatisch durchgeführt wird:
    • - einer Prozessrelevanz der Prozesswerte;
    • - einer Position eines Prozesswerts oder eines den Prozesswert ermittelnden Sensors innerhalb der verfahrenstechnischen Anlage (101) ;
    • - einem Betrag einer Abweichung eines Prozesswerts von einem definierten Prozessfenster und/oder von einem Normalzustand;
    • - einer Priorisierung von historischen Prozesswerten in historischen Fehlersituationen;
    • - physikalischen Abhängigkeiten der Prozesswerte.
• Ausführungsform 5:
  • Verfahren nach einer der Ausführungsformen 1 bis 4, dadurch gekennzeichnet, dass historische Fehlersituationen aus einer Fehlerdatenbank (136) anhand eines oder mehrerer der folgenden Ähnlichkeitskriterien ermittelt werden:
    • - einer Fehlerklassifikation der historischen Fehlersituation;
    • - einer historischen Fehlersituation an demselben oder einem vergleichbaren Anlagenteil;
    • - identischen oder ähnlichen Prozesswerten der historischen Fehlersituation zu Prozesswerten der erkannten Fehlersituation.
• Ausführungsform 6:
  • Verfahren nach einer der Ausführungsformen 1 bis 5, dadurch gekennzeichnet, dass historische Prozesswerte aus einer Prozessdatenbank (134) ermittelt werden, welche zu Prozesswerten der erkannten Fehlersituation identisch oder ähnlich sind.
• Ausführungsform 7:
  • Verfahren nach Ausführungsform 6, dadurch gekennzeichnet, dass die ermittelten historischen Prozesswerte als zu einer historischen Fehlersituation gehörig gekennzeichnet werden.
• Ausführungsform 8:
  • Verfahren nach einer der Ausführungsformen 1 bis 7, dadurch gekennzeichnet, dass zu einer erkannten Fehlersituation ein Fehlersituationsdatensatz in einer Fehlerdatenbank (136) gespeichert wird.
• Ausführungsform 9:
  • Verfahren nach Ausführungsform 8, dadurch gekennzeichnet, dass ein jeweiliger Fehleridentifikationsdatensatz eine oder mehrere der folgenden Fehlersituationsdaten umfasst:
    • - eine Fehlerklassifikation der Fehlersituation;
    • - mit der Fehlersituation verknüpfte Prozesswerte basierend auf einer Vorverknüpfung aus einem Meldesystem;
    • - Informationen über einen Zeitpunkt des Auftretens einer jeweiligen Fehlersituation;
    • - Informationen über eine Zeitdauer des Auftretens einer jeweiligen Fehlersituation;
    • - Informationen über den Ort des Auftretens einer jeweiligen Fehlersituation;
    • - Alarme;
    • - Statusmeldungen.
• Ausführungsform 10:
  • Verfahren nach Ausführungsform 8 oder 9, dadurch gekennzeichnet, dass der Fehlersituationsdatensatz einer jeweiligen Fehlersituation Fehleridentifikationsdaten zur eindeutigen Identifikation der erkannten Fehlersituation umfasst.
• Ausführungsform 11:
  • Verfahren nach einer der Ausführungsformen 8 bis 10, dadurch gekennzeichnet, dass in dem Fehlersituationsdatensatz einer jeweiligen Fehlersituation Dokumentationsdaten und Fehlerbehebungsdaten gespeichert werden.
• Ausführungsform 12:
  • Verfahren nach einer der Ausführungsformen 8 bis 11, dadurch gekennzeichnet, dass Prozesswerte im Betrieb der verfahrenstechnischen Anlage (101) zeitlich synchronisiert mit einer erkannten Fehlersituation gespeichert werden.
• Ausführungsform 13:
  • Verfahren nach einer der Ausführungsformen 8 bis 12, dadurch gekennzeichnet, dass Prozesswerte mit einem Zeitstempel versehen werden, mittels welchem die Prozesswerte einem Zeitpunkt eindeutig zuordenbar sind.
• Ausführungsform 14:
  • Fehleranalysesystem (144) zur Fehleranalyse in einer verfahrenstechnischen Anlage (101), beispielsweise in einer Lackieranlage (102), welches zur Durchführung des Verfahrens zur Fehleranalyse in einer verfahrenstechnischen Anlage (101), beispielsweise in einer Lackieranlage (102), nach einer der Ausführungsformen 1 bis 13 ausgebildet und eingerichtet ist.
• Ausführungsform 15:
  • Industrielles Steuerungssystem (100), welches ein Fehleranalysesystem (144) nach Ausführungsform 14 umfasst.
• Ausführungsform 16:
  • Verfahren zur Vorhersage von Prozessabweichungen in einer verfahrenstechnischen Anlage (101), beispielsweise in einer Lackieranlage (102), wobei das Verfahren Folgendes umfasst:
    • - automatisches Erstellen eines Vorhersagemodells;
    • - Vorhersage von Prozessabweichungen im Betrieb der verfahrenstechnischen Anlage (101) unter Verwendung des Vorhersagemodells.
• Ausführungsform 17:
  • Verfahren nach Ausführungsform 16, dadurch gekennzeichnet, dass das Verfahren zur Vorhersage von Prozessabweichungen in einer industriellen Zuluftanlage (128), in einer Vorbehandlungsstation (112), in einer Station zur kathodischen Tauchlackierung (114) und/oder in einer Trocknerstation (116, 120, 124) durchgeführt wird.
• Ausführungsform 18:
  • Verfahren nach Ausführungsform 16 oder 17, dadurch gekennzeichnet, dass mittels des Vorhersagemodells Prozessabweichungen von produktionskritischen Prozesswerten in der verfahrenstechnischen Anlage (101) vorhergesagt werden auf Basis von sich ändernden Prozesswerten während einem Betrieb der verfahrenstechnischen Anlage (101).
• Ausführungsform 19:
  • Verfahren nach einer der Ausführungsformen 16 bis 18, dadurch gekennzeichnet, dass zum automatischen Erstellen des Vorhersagemodells Prozesswerte und/oder Statusgrößen im Betrieb der verfahrenstechnischen Anlage (101) über einen vorgegebenen Zeitraum gespeichert werden.
• Ausführungsform 20:
  • Verfahren nach Ausführungsform 19, dadurch gekennzeichnet, dass der vorgegebene Zeitraum, über welchen Prozesswerte und/oder Statusgrößen im Betrieb der verfahrenstechnischen Anlage (101) gespeichert werden, abhängig von einem oder mehreren der folgenden Kriterien vorgegeben wird:
    • - die verfahrenstechnische Anlage (101) ist in dem vorgegebenen Zeitraum zu mindestens ungefähr 60 %, vorzugsweise zu mindestens ungefähr 80 %, in einem betriebsbereiten Zustand, insbesondere für einen Produktionsbetrieb;
    • - die verfahrenstechnische Anlage (101) ist in dem vorgegebenen Zeitraum zu mindestens ungefähr 60 %, vorzugsweise zu mindestens ungefähr 80 %, in einem produktionsfähigen Zustand;
    • - die verfahrenstechnische Anlage (101) wird in dem vorgegebenen Zeitraum insbesondere mit sämtlichen möglichen Betriebsstrategien betrieben;
    • - einer vorgegebene Anzahl von Prozessabweichungen und/oder Störfällen in dem vorgegebenen Zeitraum.
• Ausführungsform 21:
  • Verfahren nach Ausführungsform 19 oder 20, dadurch gekennzeichnet, dass zum Erstellen des Vorhersagemodells ein maschinelles Lernverfahren durchgeführt wird, wobei die über den vorgegebenen Zeitraum gespeicherten Prozesswerte und/oder Statusgrößen zur Erstellung des Vorhersagemodells verwendet werden.
• Ausführungsform 22:
  • Verfahren nach Ausführungsform 21, dadurch gekennzeichnet, dass das maschinelle Lernverfahren auf Basis von Features durchgeführt wird, welche aus den über den vorgegebenen Zeitraum gespeicherten Prozesswerten und/oder Statusgrößen extrahiert werden.
• Ausführungsform 23:
  • Verfahren nach Ausführungsform 22, dadurch gekennzeichnet, dass zur Extraktion der Features eines oder mehrere der folgenden verwendet werden:
    • - statistische Kennzahlen;
    • - Koeffizienten aus einer Hauptkomponentenanalyse;
    • - lineare Regressionskoeffizienten;
    • - dominante Frequenzen und/oder Amplituden aus dem Fourier-Spektrum.
• Ausführungsform 24:
  • Verfahren nach einer der Ausführungsform 16 bis 23, dadurch gekennzeichnet, dass zu einem Training des Vorhersagemodells eine ausgewählte Anzahl von Prädiktionsdatensätzen mit Prozessabweichungen (222) und eine ausgewählte Anzahl von Prädiktionsdatensätzen ohne Prozessabweichungen (220) verwendet werden.
• Ausführungsform 25:
  • Verfahren nach Ausführungsform 24, dadurch gekennzeichnet, dass die Auswahl der Anzahl von Prädiktionsdatensätzen mit Prozessabweichung aufgrund eines oder mehrerer der folgenden Kriterien erfolgt:
    • - einem zeitlichen Mindestabstand zwischen zwei Prädiktionsdatensätzen mit Prozessabweichungen;
    • - einer automatischen Auswahl anhand von definierten Regeln;
    • - einer Auswahl durch einen Benutzer.
• Ausführungsform 26:
  • Verfahren nach Ausführungsform 24 oder 25, dadurch gekennzeichnet, dass Prädiktionsdatensätze mit Prozessabweichungen als solche gekennzeichnet werden, wenn innerhalb eines vorgegebenen Zeitintervalls eine Prozessabweichung auftritt.
• Ausführungsform 27:
  • Verfahren nach Ausführungsform 26, dadurch gekennzeichnet, dass die über den vorgegebenen Zeitraum gespeicherten Prozesswerte und/oder Statusgrößen durch eine Vorverarbeitung in Prädiktionsdatensätze zusammengefasst werden.
• Ausführungsform 28:
  • Verfahren nach Ausführungsform 27, dadurch gekennzeichnet, dass die Vorverarbeitung Folgendes umfasst:
    • - Regularisierung der über den vorgegebenen Zeitraum gespeicherten Prozesswerte;
    • - Zusammenfassung der Prozesswerte und/oder Statusgrößen in Prädiktionsdatensätze durch Einteilung der Prozesswerte und/oder Statusgrößen in Zeitfenster mit zeitlicher Verschiebung.
• Ausführungsform 29:
  • Vorhersagesystem (146) zur Vorhersage von Prozessabweichungen in einer verfahrenstechnischen Anlage, welches zur Durchführung des Verfahrens zur Vorhersage von Prozessabweichungen in einer verfahrenstechnischen Anlage (101), beispielsweise in einer Lackieranlage (102), nach einer der Ausführungsformen 16 bis 29 ausgebildet und eingerichtet ist.
• Ausführungsform 30:
  • Industrielles Steuerungssystem (100), welches ein Vorhersagesystem (146) nach Ausführungsform 29 umfasst.
• Ausführungsform 31:
  • Verfahren zur Anomalie- und/oder Fehlererkennung in einer verfahrenstechnischen Anlage (101), beispielsweise in einer Lackieranlage (102), wobei das Verfahren Folgendes umfasst:
    • - automatisches Erstellen eines Anomalie- und/oder Fehlermodells (233) der verfahrenstechnischen Anlage (101), welches Informationen über die Auftretenswahrscheinlichkeit von Prozesswerten umfasst;
    • - automatisches Einlesen von Prozesswerten der verfahrenstechnischen Anlage (101) im Betrieb derselben;
    • - automatische Erkennung einer Anomalie und/oder einer Fehlersituation durch Ermitteln einer Auftretenswahrscheinlichkeit mittels des Anomalie- und/oder Fehlermodells (233) auf Basis der eingelesenen Prozesswerte der verfahrenstechnischen Anlage (101) und durch Überprüfen der Auftretenswahrscheinlichkeit auf einen Grenzwert.
• Ausführungsform 32:
  • Verfahren nach Ausführungsform 31, dadurch gekennzeichnet, dass
    • - das Anomalie- und/oder Fehlermodell (233) Strukturdaten umfasst, welche Informationen über eine Prozessstruktur in der verfahrenstechnischen Anlage (101) enthalten, und/oder dass
    • - das Anomalie- und/oder Fehlermodell (233) Parametrierungsdaten umfasst, welche Informationen über Zusammenhänge zwischen Prozesswerten der verfahrenstechnischen Anlage (101) enthalten.
• Ausführungsform 33:
  • Verfahren nach Ausführungsform 31 oder 32, dadurch gekennzeichnet, dass zum Erstellen des Anomalie- und/oder Fehlermodells (233) einer oder mehrere der folgenden Schritte durchgeführt werden:
    • - Strukturidentifikation (246) zur Ermittlung einer Prozessstruktur der verfahrenstechnischen Anlage (101);
    • - Ermittlung von Kausalitäten (254) in der ermittelten Prozessstruktur der verfahrenstechnischen Anlage (101);
    • - Strukturparametrierung (256) der Zusammenhänge in der ermittelten Prozessstruktur der verfahrenstechnischen Anlage (101).
• Ausführungsform 34:
  • Verfahren nach Ausführungsform 33, dadurch gekennzeichnet, dass bei der Strukturidentifikation (246) zur Ermittlung einer Prozessstruktur der verfahrenstechnischen Anlage (101) ein Strukturgraph ermittelt wird, welcher insbesondere Zusammenhänge in der verfahrenstechnischen Anlage (101) abbildet.
• Ausführungsform 35:
  • Verfahren nach Ausführungsform 34, dadurch gekennzeichnet, dass die Ermittlung des Strukturgraphen unter Verwendung eines oder mehrerer der folgenden erfolgt:
    • - eines maschinellen Lernverfahrens;
    • - Expertenwissen (248);
    • - bekannte Schaltplänen und/oder Verfahrensschemata (250);
    • - Bezeichnungen in einem Nummerierungssystem der verfahrenstechnischen Anlage (101).
• Ausführungsform 36:
  • Verfahren nach Ausführungsform 33 bis 35, dadurch gekennzeichnet, dass die verfahrenstechnische Anlage (101) zur Strukturidentifikation, insbesondere zur Ermittlung des Strukturgraphen, mit Testsignalen angeregt wird.
• Ausführungsform 37:
  • Verfahren nach einer der Ausführungsformen 33 bis 36, dadurch gekennzeichnet, dass die Ermittlung von Kausalitäten (254) in der ermittelten Prozessstruktur der verfahrenstechnischen Anlage (101) unter Verwendung eines oder mehrerer der folgenden erfolgt:
    • - bei der Anregung der verfahrenstechnischen Anlage (101) mit Testsignalen generierte Systemeingangssignale (240) und Systemausgangssignale (242);
    • - Expertenwissen (248);
    • - bekannten Schaltplänen und/oder Verfahrensschemata (252);
    • - Bezeichnungen in einem Nummerierungssystem der verfahrenstechnischen Anlage (101).
• Ausführungsform 38:
  • Verfahren nach einer der Ausführungsformen 33 bis 37, dadurch gekennzeichnet, dass zur Strukturparametrierung (246) der Zusammenhänge in der ermittelten Prozessstruktur der verfahrenstechnischen Anlage (101) eines oder mehrere der folgenden verwendet werden:
    • - Verfahren zur Ermittlung von Wahrscheinlichkeitsdichtefunktionen, insbesondere Gaußsche Mischmodelle;
    • - bekannte physikalische Zusammenhänge zwischen Prozesswerten;
    • - physikalische Kennfelder von Funktionselementen der verfahrenstechnischen Anlage (101), beispielsweise Kennfelder von Ventilen (232).
• Ausführungsform 39:
  • Verfahren nach Ausführungsform 38, dadurch gekennzeichnet, dass zur Strukturparametrierung (246) unter Verwendung von Verfahren zur Ermittlung von Wahrscheinlichkeitsdichtefunktionen, insbesondere unter Verwendung Gaußscher Mischmodelle, Daten aus dem regulären Betrieb der verfahrenstechnischen Anlage (101) und/oder durch eine Anregung der verfahrenstechnischen Anlage (101) mittels Testsignalen gewonnene Daten verwendet werden.
• Ausführungsform 40:
  • Verfahren nach Ausführungsform 39, dadurch gekennzeichnet, dass die zur Strukturparametrierung (246) unter Verwendung von Verfahren zur Ermittlung von Wahrscheinlichkeitsdichtefunktionen, insbesondere unter Verwendung Gaußscher Mischmodelle, verwendeten Daten vor der Strukturparametrierung (246) vorverarbeitet werden.
• Ausführungsform 41:
  • Verfahren nach einer der Ausführungsformen 31 bis 40, dadurch gekennzeichnet, dass bei der Erstellung des Anomalie- und/oder Fehlermodells (233) ein Grenzwert für die Auftretenswahrscheinlichkeit eines Prozesswertes festgelegt wird, wobei eine Anomalie bei einer Unterschreitung des Grenzwertes erkannt wird.
• Ausführungsform 42:
  • Verfahren nach einer der Ausführungsformen 31 bis 41, dadurch gekennzeichnet, dass mittels des Verfahrens zur Anomalie- und/oder Fehlererkennung eine Fehlerursache einer erkannten Anomalie und/oder einer erkannten Fehlersituation identifiziert wird.
• Ausführungsform 43:
  • Verfahren nach einer der Ausführungsformen 31 bis 42, dadurch gekennzeichnet, dass die verfahrenstechnische Anlage (101) eine oder mehrere der folgenden Behandlungsstationen (104) einer Lackieranlage umfasst oder durch diese gebildet wird:
    • - Vorbehandlungsstation (112);
    • - Station zur kathodischen Tauchlackierung (114);
    • - Trocknerstationen (116, 120, 124);
    • - eine industrielle Zuluftanlage (128);
    • - Lackierroboter.
• Ausführungsform 44:
  • Anomalie- und/oder Fehlererkennungssystem (148) zur Anomalie- und/oder Fehlererkennung, welches zur Durchführung des Verfahrens zur Anomalie- und/oder Fehlererkennung in einer verfahrenstechnischen Anlage (101), beispielsweise in einer Lackieranlage (102), nach einer der Ausführungsformen 31 bis 43 ausgebildet und eingerichtet ist.
• Ausführungsform 45:
  • Industrielles Steuerungssystem (100), welches ein Anomalie- und/oder Fehlererkennungssystem (148) nach Ausführungsform 44 umfasst.
• Ausführungsform 46:
  • Industrielles Steuerungssystem, welches ein Fehleranalysesystem nach Ausführungsform 14, ein Vorhersagesystem zur Vorhersage von Prozessabweichungen in einer verfahrenstechnischen Anlage nach Ausführungsform 29 und/oder ein Anomalie- und/oder Fehlererkennungssystem nach Ausführungsform 44 umfasst.

Special embodiments are the following:

• Embodiment 1:
  • Procedure for error analysis in a process plant ( 101 ), for example in a paint shop ( 102 ), the method comprising:
  • - in particular automatic detection of an error situation in the process engineering system ( 101 );
  • - Saving an error situation data record for the respective recognized error situation in an error database ( 136 );
  • - automatic determination of a cause of the error for the error situation and / or automatic determination of process values relevant to the error situation on the basis of the error data record of a respective recognized error situation.
• Embodiment 2:
  • Method according to embodiment 1, characterized in that, in order to automatically determine the cause of the error for the error situation and / or the process values relevant to the error situation, one or more process values are automatically linked to the error situation based on one or more of the following linking criteria:
    • - a pre-link from a reporting system;
    • - an assignment of a process value to the same part of the process plant ( 101 ) in which the error situation occurred;
    • a link between a process value and a historical error situation based on an active selection by a user;
    • - an active selection of the process value by a user.
• Embodiment 3:
  • Method according to embodiment 2, characterized in that, in order to automatically determine the cause of the error for the error situation and / or the process values relevant to the error situation, an automatic prioritization of the process values linked to the error situation based on one or more of the following prioritization criteria is carried out automatically:
    • - a process relevance of the process values;
    • - a position of a process value or a sensor that determines the process value within the process plant ( 101 );
    • an amount of a deviation of a process value from a defined process window and / or from a normal state;
    • - a prioritization of historical process values in historical error situations;
    • - by adopting a prioritization of the cause of the error and / or the process values from a message system ( 138 );
    • - a prioritization by a user.
• Embodiment 4:
  • Method according to one of the embodiments 1 to 3, characterized in that further error causes and / or process values are proposed for the automatic determination of the cause of the error for the error situation and / or the process values relevant to the error situation, the suggestion being automatic based on one or more of the following criteria is carried out:
    • - a process relevance of the process values;
    • - a position of a process value or a sensor that determines the process value within the process plant ( 101 );
    • an amount of a deviation of a process value from a defined process window and / or from a normal state;
    • - a prioritization of historical process values in historical error situations;
    • - physical dependencies of the process values.
• Embodiment 5:
  • Method according to one of the embodiments 1 to 4, characterized in that historical error situations from an error database ( 136 ) can be determined using one or more of the following similarity criteria:
    • - an error classification of the historical error situation;
    • - a historical error situation on the same or a comparable part of the system;
    • - identical or similar process values of the historical error situation to process values of the recognized error situation.
• Embodiment 6:
  • Method according to one of the embodiments 1 to 5, characterized in that historical process values from a process database ( 134 ) which are identical or similar to the process values of the detected error situation.
• Embodiment 7:
  • Method according to embodiment 6, characterized in that the historical process values determined are identified as belonging to a historical error situation.
• Embodiment 8:
  • Method according to one of the embodiments 1 to 7, characterized in that a fault situation data record is stored in a fault database ( 136 ) is saved.
• Embodiment 9:
  • Method according to embodiment 8, characterized in that a respective error identification data record comprises one or more of the following error situation data:
    • - an error classification of the error situation;
    • - Process values linked to the error situation based on a pre-link from a reporting system;
    • - Information about a point in time when a respective error situation occurred;
    • - Information about a duration of the occurrence of a respective error situation;
    • - Information about the location of the occurrence of a respective error situation;
    • - alarms;
    • - Status messages.
• Embodiment 10:
  • Method according to embodiment 8 or 9, characterized in that the error situation data set of a respective error situation comprises error identification data for the unambiguous identification of the detected error situation.
• Embodiment 11:
  • Method according to one of the embodiments 8 to 10, characterized in that documentation data and error correction data are stored in the error situation data record of a respective error situation.
• Embodiment 12:
  • Method according to one of the embodiments 8 to 11, characterized in that process values in the operation of the process engineering plant ( 101 ) are saved in time synchronized with a recognized error situation.
• Embodiment 13:
  • Method according to one of the embodiments 8 to 12, characterized in that process values are provided with a time stamp, by means of which the process values can be clearly assigned to a point in time.
• Embodiment 14:
  • Failure analysis system ( 144 ) for error analysis in a process plant ( 101 ), for example in a paint shop ( 102 ), which is used to carry out the procedure for error analysis in a process plant ( 101 ), for example in a paint shop ( 102 ), is designed and set up according to one of the embodiments 1 to 13.
• Embodiment 15:
  • Industrial control system ( 100 ), which is an error analysis system ( 144 ) according to embodiment 14 comprises.
• Embodiment 16:
  • Procedure for predicting process deviations in a process plant ( 101 ), for example in a paint shop ( 102 ), the method comprising:
    • - automatic creation of a predictive model;
    • - Prediction of process deviations in the operation of the process plant ( 101 ) using the predictive model.
• Embodiment 17:
  • Method according to embodiment 16, characterized in that the method for predicting process deviations in an industrial air supply system ( 128 ), in a pre-treatment station ( 112 ), in a station for cathodic dip painting ( 114 ) and / or in a drying station ( 116 , 120 , 124 ) is carried out.
• Embodiment 18:
  • Method according to embodiment 16 or 17, characterized in that process deviations from production-critical process values in the process engineering system ( 101 ) are predicted on the basis of changing process values during operation of the process plant ( 101 ).
• Embodiment 19:
  • Method according to one of the embodiments 16 to 18, characterized in that for the automatic creation of the prediction model process values and / or status variables in the operation of the process engineering plant ( 101 ) can be saved for a specified period.
• Embodiment 20:
  • Method according to embodiment 19, characterized in that the predetermined period of time over which process values and / or status variables in the operation of the process engineering plant ( 101 ), depending on one or more of the following criteria:
    • - the process plant ( 101 ) is at least approximately 60%, preferably at least approximately 80%, in an operational state in the specified period, in particular for a production facility;
    • - the process plant ( 101 ) is at least approximately 60%, preferably at least approximately 80%, in a productive state in the specified period of time;
    • - the process plant ( 101 ) is operated in particular with all possible operating strategies in the specified period;
    • - a specified number of process deviations and / or incidents in the specified period.
• Embodiment 21:
  • Method according to embodiment 19 or 20, characterized in that a machine learning method is carried out to create the prediction model, the process values and / or status variables stored over the predetermined period of time being used to create the prediction model.
• Embodiment 22:
  • Method according to embodiment 21, characterized in that the machine learning method is carried out on the basis of features which are extracted from the process values and / or status variables stored over the predetermined period of time.
• Embodiment 23:
  • Method according to embodiment 22, characterized in that one or more of the following are used to extract the features:
    • - statistical indicators;
    • - coefficients from a principal component analysis;
    • - linear regression coefficients;
    • - dominant frequencies and / or amplitudes from the Fourier spectrum.
• Embodiment 24:
  • Method according to one of the embodiment 16 to 23, characterized in that a selected number of prediction data sets with process deviations ( 222 ) and a selected number of prediction data sets without process deviations ( 220 ) be used.
• Embodiment 25:
  • Method according to embodiment 24, characterized in that the selection of the number of prediction data sets with process deviations is based on one or more of the following criteria:
    • - a minimum time interval between two prediction data sets with process deviations;
    • - an automatic selection based on defined rules;
    • - a selection by a user.
• Embodiment 26:
  • Method according to embodiment 24 or 25, characterized in that prediction data sets with process deviations are identified as such if a process deviation occurs within a predetermined time interval.
• Embodiment 27:
  • Method according to embodiment 26, characterized in that the process values and / or status variables stored over the predetermined period of time are combined into prediction data sets by preprocessing.
• Embodiment 28:
  • Method according to embodiment 27, characterized in that the preprocessing comprises:
    • - Regularization of the process values stored over the specified period;
    • - Summary of the process values and / or status variables in prediction data sets by dividing the process values and / or status variables into time windows with a time shift.
• Embodiment 29:
  • Forecast system ( 146 ) to predict process deviations in a process plant, which is used to carry out the method for predicting process deviations in a process plant ( 101 ), for example in a paint shop ( 102 ), is designed and set up according to one of the embodiments 16 to 29.
• Embodiment 30:
  • Industrial control system ( 100 ), which is a prediction system ( 146 ) according to embodiment 29 comprises.
• Embodiment 31:
  • Procedure for anomaly and / or error detection in a process plant ( 101 ), for example in a paint shop ( 102 ), the method comprising:
    • - automatic creation of an anomaly and / or failure model ( 233 ) of the process plant ( 101 ), which includes information about the probability of occurrence of process values;
    • - automatic reading in of process values from the process plant ( 101 ) in the operation of the same;
    • - Automatic detection of an anomaly and / or an error situation by determining a probability of occurrence using the anomaly and / or error model ( 233 ) on the basis of the process values read in from the process plant ( 101 ) and by checking the probability of occurrence for a limit value.
• Embodiment 32:
  • Method according to embodiment 31, characterized in that
    • - the anomaly and / or failure model ( 233 ) Includes structural data, which contains information about a process structure in the process plant ( 101 ) included, and / or that
    • - the anomaly and / or failure model ( 233 ) Includes parameterization data, which contains information about relationships between process values of the process plant ( 101 ) contain.
• Embodiment 33:
  • Method according to embodiment 31 or 32, characterized in that to create the anomaly and / or error model ( 233 ) one or more of the following steps must be carried out:
    • - structure identification ( 246 ) to determine a process structure of the process plant ( 101 );
    • - Determination of causalities ( 254 ) in the determined process structure of the process plant ( 101 );
    • - Structure parameterization ( 256 ) the relationships in the determined process structure of the process plant ( 101 ).
• Embodiment 34:
  • Method according to embodiment 33, characterized in that in the structure identification ( 246 ) to determine a process structure of the process plant ( 101 ) a structure graph is determined, which in particular shows relationships in the process engineering system ( 101 ) maps.
• Embodiment 35:
  • Method according to embodiment 34, characterized in that the structure graph is determined using one or more of the following:
    • - a machine learning process;
    • - Expert knowledge ( 248 );
    • - known circuit diagrams and / or process diagrams ( 250 );
    • - Designations in a numbering system of the process plant ( 101 ).
• Embodiment 36:
  • Method according to embodiment 33 to 35, characterized in that the procedural plant ( 101 ) for structure identification, in particular for determining the structure graph, is stimulated with test signals.
• Embodiment 37:
  • Method according to one of the embodiments 33 to 36, characterized in that the determination of causalities ( 254 ) in the determined process structure of the process plant ( 101 ) is made using one or more of the following:
    • - when stimulating the process plant ( 101 ) system input signals generated with test signals ( 240 ) and system output signals ( 242 );
    • - Expert knowledge ( 248 );
    • - known circuit diagrams and / or process diagrams ( 252 );
    • - Designations in a numbering system of the process plant ( 101 ).
• Embodiment 38:
  • Method according to one of the embodiments 33 to 37, characterized in that for structural parameterization ( 246 ) the relationships in the determined process structure of the process plant ( 101 ) one or more of the following can be used:
    • - Method for determining probability density functions, in particular Gaussian mixed models;
    • - known physical relationships between process values;
    • - physical maps of functional elements of the process plant ( 101 ), for example valve maps ( 232 ).
• Embodiment 39:
  • Method according to embodiment 38, characterized in that for structural parameterization ( 246 ) using methods for determining probability density functions, in particular using Gaussian mixed models, data from the regular operation of the process engineering plant ( 101 ) and / or by stimulating the process engineering system ( 101 ) data obtained by means of test signals can be used.
• Embodiment 40:
  • Method according to embodiment 39, characterized in that the structure parameterization ( 246 ) using methods for determining probability density functions, in particular using Gaussian mixed models, data used before the structure parameterization ( 246 ) are preprocessed.
• Embodiment 41:
  • Method according to one of the embodiments 31 to 40, characterized in that when creating the anomaly and / or error model ( 233 ) a limit value for the probability of occurrence of a process value is established, an anomaly being detected if the limit value is not reached.
• Embodiment 42:
  • Method according to one of the embodiments 31 to 41, characterized in that the method for anomaly and / or error detection is used to identify an error cause of a recognized anomaly and / or a recognized error situation.
• Embodiment 43:
  • Method according to one of the embodiments 31 to 42, characterized in that the procedural plant ( 101 ) one or more of the following treatment stations ( 104 ) includes or is formed by a paint shop:
    • - pre-treatment station ( 112 );
    • - Station for cathodic dip painting ( 114 );
    • - dryer stations ( 116 , 120 , 124 );
    • - an industrial air supply system ( 128 );
    • - painting robots.
• Embodiment 44:
  • Anomaly and / or error detection system ( 148 ) for anomaly and / or error detection, which is used to carry out the method for anomaly and / or error detection in a process engineering plant ( 101 ), for example in a paint shop ( 102 ), is designed and set up according to one of the embodiments 31 to 43.
• Embodiment 45:
  • Industrial control system ( 100 ), which has an anomaly and / or error detection system ( 148 ) according to embodiment 44 comprises.
• Embodiment 46:
  • Industrial control system which comprises a fault analysis system according to embodiment 14, a prediction system for predicting process deviations in a process engineering plant according to embodiment 29 and / or an anomaly and / or error detection system according to embodiment 44.

Claims (15)

A method for predicting process deviations in a process plant (101), for example in a painting plant (102), the method comprising the following: - automatic creation of a predictive model; - Prediction of process deviations in the operation of the process plant (101) using the prediction model. Procedure according to Claim 1 , characterized in that the method for predicting process deviations is carried out in an industrial air supply system (128), in a pretreatment station (112), in a station for cathodic dip painting (114) and / or in a dryer station (116, 120, 124) . Procedure according to Claim 1 or 2 , characterized in that process deviations from production-critical process values in the process engineering system (101) are predicted by means of the forecast model on the basis of changing process values during operation of the process engineering system (101). Method according to one of the Claims 1 to 3rd , characterized in that, for the automatic creation of the prediction model, process values and / or status variables are stored during the operation of the process engineering system (101) over a predetermined period of time. Procedure according to Claim 4 , characterized in that the predefined period of time over which process values and / or status variables are stored in the operation of the procedural plant (101) is predefined as a function of one or more of the following criteria: the procedural plant (101) is in the predefined period of time at least approximately 60%, preferably at least approximately 80%, in an operational state, in particular for a production plant; - The process-engineering plant (101) is in a productive state for at least approximately 60%, preferably at least approximately 80%, in the predetermined period of time; - The process engineering system (101) is operated in the specified period, in particular with all possible operating strategies; - a specified number of process deviations and / or incidents in the specified period. Procedure according to Claim 4 or 5 , characterized in that a machine learning method is carried out to create the prediction model, the process values and / or status variables stored over the predetermined period of time being used to create the prediction model. Procedure according to Claim 6 , characterized in that the machine learning method is carried out on the basis of features which are extracted from the process values and / or status variables stored over the predetermined period of time. Procedure according to Claim 7 , characterized in that one or more of the following are used to extract the features: - statistical key figures; - coefficients from a principal component analysis; - linear regression coefficients; - dominant frequencies and / or amplitudes from the Fourier spectrum. Method according to one of the Claims 1 to 8th , characterized in that a selected number of prediction data sets with process deviations (222) and a selected number of prediction data sets without process deviations (220) are used for training the prediction model. Procedure according to Claim 9 , characterized in that the number of prediction data sets with process deviations is selected on the basis of one or more of the following criteria: a minimum time interval between two prediction data sets with process deviations; - an automatic selection based on defined rules; - a selection by a user. Procedure according to Claim 9 or 10 , characterized in that prediction data sets marked as such with process deviations if a process deviation occurs within a specified time interval. Procedure according to Claim 11 , characterized in that the process values and / or status variables stored over the specified period are summarized in prediction data sets by preprocessing. Procedure according to Claim 12 , characterized in that the preprocessing comprises the following: regularization of the process values stored over the predetermined period of time; - Summary of the process values and / or status variables in prediction data sets by dividing the process values and / or status variables into time windows with a time shift. Prediction system (146) for predicting process deviations in a process engineering plant, which is used to carry out the method for predicting process deviations in a process engineering plant (101), for example in a painting plant (102), according to one of the Claims 1 to 13th is trained and established. Industrial control system (100) comprising a prediction system (146) according to Claim 14 includes.

Images