DE102019116136A1 - Procedure for determining the causes of errors in automation components - Google Patents

Abstract

The invention relates to a method for determining the causes of errors in automation components which are used in systems of automation technology, comprising: Acquisition of event data (ED) of the automation components from a first database (DB1) or from the automation components directly, which event data (ED) at least information about an event of an automation component linked to identification information of the respective automation component; - acquisition of master data (MD), which are linked to the identification data of the respective automation components, from at least one second database (DB2, DB3, DB4); - correlated storage of the Event data (ED) of the automation components and the master data (MD) on a service platform (SP); and- statistical evaluation of the event data (ED) and master data (MD) to determine the cause of the events that have occurred, the statistical evaluation comprising a modeling of the reliability of the respective types of automation components and / or a dimension reduction, classification and clustering.

Description

The invention relates to a method for determining the causes of errors in automation components.

Automation components that are used in industrial systems of automation technology have already become known from the prior art. As an example of automation components, field devices are often used in process automation, just as in production automation. In principle, all devices that are used close to the process and that supply or process process-relevant information are referred to as field devices. Field devices are used to record and / or influence process variables. Sensor units are used to record process variables. These are used, for example, for pressure and temperature measurement, conductivity measurement, flow measurement, pH measurement, level measurement, etc. and record the corresponding process variables pressure, temperature, conductivity, pH value, level, flow, etc. Actuator systems are used to influence process variables. These are, for example, pumps or valves that can influence the flow of a liquid in a pipe or the level in a container. In addition to the aforementioned field devices, automation components also include remote I / Os, radio adapters or, in general, devices that are arranged on the field level.

The failure of an automation component means high costs and time expenditure. In critical processes in particular - until the automation component has been repaired or replaced - the relevant part of the system has to be shut down. As an alternative to this, redundancies are often built into the relevant system part in advance. As a rule, these are automation components that are identical to the automation components used and that replace an affected automation component in the event of a fault.

Sometimes it is difficult to determine the exact cause of a failure or a defect in an automation component. It has become known from the prior art, for example, to use reliability modeling to determine the reliability of automation components. Such a reliability modeling consists of one or more statistical distribution curves, the parameters of which are determined specifically for the type of automation component. The error curves include a behavior, in particular an occurrence, of specific event data of the automation component over the lifetime of the respective automation component. A deviation of an automation component from this characteristic indicates an error behavior of this automation component.

However, it is particularly desirable to determine whether a failure or malfunction of an automation component is an isolated case, or whether similar failures are to be expected for automation components from the same series. So far, however, the detection of series defects has been reactive, unsystematic or based on a subjectively perceived accumulation of complaints or service events in an automation component series.

This means that series errors in automation components can sometimes only be detected with a long delay, which leads to an increase in damage cases in the field before there is a possibility of remedial action. This increases the risk of damage as well as the costs of remedial measures and possible damage repair. In particular, it remains unclear whether and which usage properties of the automation components cause errors or whether an error can be traced back to the production of the automation component.

Proceeding from this problem, the invention is based on the object of presenting a method which, on the one hand, allows faulty behavior of automation components to be reliably detected and, on the other hand, to determine a cause for such a detected faulty behavior.

• - Erfassen von Eventdaten der Automatisierungskomponenten aus einer ersten Datenbank oder von den Automatisierungskomponenten direkt, welche Eventdaten zumindest eine Information über einen Event einer Automatisierungskomponente verknüpft mit Identifikationsinformationen der jeweiligen Automatisierungskomponente umfassen;
• - Erfassen von Masterdaten, welche mit den Identifikationsdaten der jeweiligen Automatisierungskomponenten verknüpft sind, aus zumindest einer zweiten Datenbank;
• - Korreliertes Speichern der Eventdaten der Automatisierungskomponenten und der Masterdaten auf einer Serviceptattform; und
• - Statistisches Auswerten der Eventdaten und Masterdaten zum Ermitteln der Ursache der aufgetretenen Events und/oder zur vorausschauenden Vorhersage, wobei das statistische Auswerten eine Modellierung der Zuverlässigkeit der jeweiligen Typen der Automatisierungskomponenten und/oder eine Dimensionsreduzierung, Klassifizierung und Clusterung umfasst.

The object is achieved by a method for determining the causes of errors in automation components that are used in systems of automation technology, comprising:

• Acquisition of event data of the automation components from a first database or from the automation components directly, which event data include at least one piece of information about an event of an automation component linked to identification information of the respective automation component;
• - Acquisition of master data, which are linked to the identification data of the respective automation components, from at least one second database;
• - Correlated storage of the event data of the automation components and the master data on a service platform; and
• - Statistical evaluation of the event data and master data to determine the cause of the events that have occurred and / or for predictive prediction, the statistical evaluation including a modeling of the reliability of the respective types of automation components and / or a dimension reduction, classification and clustering.

The essence of the invention is that event data that originate from the automation components themselves or were created for them are linked with so-called master data. The master data can represent all other automation component-related data that are not event data. The event data are necessarily time-linked, whereby certain master data can also be time-linked.

These linked data are subjected to statistical analysis. The statistical analysis can, for example, consist of the use of a modeling of the reliability, or reliability modeling, of the automation component described in the introductory part, or additionally or alternatively consist of clustering. Clustering refers to the detection of accumulations in the master and event data. Each type of data represents a dimension. This creates a multidimensional matrix in which the accumulations are searched for. Special statistical algorithms can be used for this, for example KI algorithms, which work in particular on the basis of neural networks. In order to reduce the computational effort, the dimension of the multidimensional matrix can be reduced. For example, a time period can be selected in which the statistical evaluation is to take place or for which device types this is to be carried out. Dimensions in which there is no or significantly little data in the relevant period or for the selected device types can be disregarded for the calculation. The service platform enables the mentioned algorithms to be carried out in an application application, as well as to carry out other application applications to carry out the method.

In this way, it is not only possible to find the causes of errors that have occurred. Predictions can also be made, for example whether errors are imminent for automation components.

• Der Hersteller der Automatisierungskomponenten kann diese Ergebnisse für dessen Forschung und Entwicklung nutzen. Es können mithilfe der Ergebnisse gezielt Schwachstellen geschlossen werden. Die Ergebnisse resultieren auch in einer verbesserten Kenntnis, welche Komponenten und Anforderungen die jeweiligen Automatisierungskomponenten in entsprechenden Zielindustrien tatsächlich benötigen, was beispielsweise auch zeit- und ressourcenintensives over-engineering verhindert.

Both the causes of the errors and the predictions can be useful for different areas:

• The manufacturer of the automation components can use these results for their research and development. The results can be used to specifically close weak points. The results also result in improved knowledge of which components and requirements the respective automation components actually need in the corresponding target industries, which for example also prevents time-consuming and resource-intensive over-engineering.

In the service area, for example, the results can be used advantageously to offer the customer of the manufacturer of the automation components an optimized maintenance plan for his automation components. Warranties such as the guarantee of the automation components can also be extended since the guarantee costs can be estimated. A so-called field performance index can be introduced, which contains obsolescence data of those parts of the automation components that can only be exchanged or replaced with considerable effort.

For the customer, all of this offers the advantage that the downtime of his system can be reduced.

• - Die Kategorie eines aufgetretenen Events und insbesondere den Status des Events;
• - Das Datum, bzw. den Zeitpunkt eines aufgetretenen Events;
• - Daten bezüglich des Kunden, bei welchem der Event aufgetreten ist;
• - Parameterdaten der jeweiligen Automatisierungskomponente, welcher dem Event zugeordnet ist;
• - Ein Fehlerprotokoll;
• - Informationen über den verfahrenstechnischen Prozess, in welchem die Automatisierungskomponente eingesetzt wird.

According to a preferred embodiment of the method according to the invention, it is provided that the event data contain one or more of the following data:

• - The category of an event that has occurred and in particular the status of the event;
• - The date or time of an event that occurred;
• - data relating to the customer at whom the event occurred;
• - Parameter data of the respective automation component which is assigned to the event;
• - An error log;
• - Information about the procedural process in which the automation component is used.

• - Die Bestellnummer einer jeweiligen Automatisierungskomponente;
• - Applikationsspezifische Daten einer jeweiligen Automatisierungskomponente, insbesondere der Industriesektor, in welchem die jeweilige Automatisierungskomponente eingesetzt wird;
• - Produktionsdaten der jeweiligen Automatisierungskomponente, wobei die Produktionsdaten enthalten:
  1. a) Produktionsparameter;
  2. b) Daten über den Ort der Produktion;
  3. c) Zeitstempel über erfolgte Produktionsschritte;
  4. d) Messwerte der Produktionsanlage während des Produktionsprozesses;
  5. e) Eventuelle Reparaturmaßnahmen der Automatisierungskomponente während der Produktion der Automatisierungskomponente; und/oder
  6. f) Losdaten eines Produktionslieferantens.

According to a preferred embodiment of the method according to the invention, it is provided that the master data contain one or more of the following data about the automation components:

• - The order number of a respective automation component;
• - Application-specific data of a respective automation component, in particular the industrial sector in which the respective automation component is used;
• - Production data of the respective automation component, with the production data containing:
  1. a) production parameters;
  2. b) data on the place of production;
  3. c) time stamp of completed production steps;
  4. d) measured values of the production plant during the production process;
  5. e) Any repair measures for the automation component during the production of the automation component; and or
  6. f) Lot data of a production supplier.

According to an advantageous embodiment of the method according to the invention, it is provided that the service platform is embedded in a cloud architecture and can be contacted via the Internet. In this case, cloud computing is understood to mean the storage of information and the access to the stored information via the Internet.

According to an advantageous embodiment of the method according to the invention, it is provided that the first database and the at least second database are embedded in a cloud architecture and can be contacted via the Internet.

According to an advantageous embodiment of the method according to the invention, it is provided that the first database and the at least second database are local databases. Local databases are databases which are located in particular in the vicinity of a system of the customer who uses the method and in which system the automation components are used, and which are integrated into the customer's company and / or system network.

• 1: einen schematischen Überblick über das erfindungsgemäße Verfahren.

The invention is explained in more detail with reference to the following figure. Show it

• 1 : a schematic overview of the method according to the invention.

In 1 the inventive method is shown schematically. The core component, which carries out the method according to the invention, is a service platform SP educated. The service platform SP is designed as a cloud-capable service platform and can be contacted by an operator via the Internet. On the service platform SP In addition to a data memory, one or more application applications can be implemented and executed, which are designed as automated processes to execute the following method steps. The operator is in particular the manufacturer of the automation components or a customer of the manufacturer who uses the automation components in a process engineering system.

In a first step, so-called event data are created ED collected from automation components of the customer. The event data contain the type of event, for example "maintenance", "service case", "repair", a time stamp with the time the event occurred and identification information of the automation component to which the event is assigned. The event dates ED can from the automation components in the sense of IloT ("Industrial Internet of Things") directly to the service platform SP be transmitted. Alternatively, the service data SD are sent from a first database DB to the service platform SP transmitted, preferably over the Internet. Alternatively, the operator can also independently enter the service data SD into the service platform SP enter.

In a second process step, so-called master data are created MD of the automation components. The master data MD are in particular all data relating to the automation components that are not service data. In particular, it is provided that a second database DB2 or multiple databases DB2 , DB3 , DB4 are available from which the master data MD can be read out.

For example, the database contains DB2 Master data MD in the form of application-specific data from the automation components. These contain information regarding the application in which the respective automation components are used, for example the plant number, the plant sector, the respective industrial type, environmental information, etc.

For example, the database contains DB3 Information regarding the respective system components themselves, for example the device type of the system component, module / component information, etc.

• - Produktionsparameter;
• - Daten über den Ort der Produktion;
• - Zeitstempel über erfolgte Produktionsschritte;
• - Messwerte der Produktionsanlage während des Produktionsprozesses;
• - Eventuelle Reparaturmaßnahmen der Automatisierungskomponente während der Produktion der Automatisierungskomponente; und/oder
• - Losdaten eines Produktionslieferantens

For example, the database contains DB4 Production data of the respective automation components, for example:

• - production parameters;
• - data on the place of production;
• - Time stamp of completed production steps;
• - Measured values of the production plant during the production process;
• - Any repair measures for the automation component during the production of the automation component; and or
• - Lot data of a production supplier

In a third process step, the event data ED with the master data MD enriched. It is provided here that the event and master data ED , MD are constantly updated by adding the latest data at regular intervals ED , MD from the databases DB1 , DB2 , DB3 , DB4 can be accessed so that the wealth of experience on the service platform SP , formed by the enrichment of the event data ED with the master data MD , is always up to date.

In a fourth step, statistical calculations are applied to the wealth of experience. The statistical analysis can, for example, consist of the use of a modeling of the reliability, or reliability modeling, of the automation component, as disclosed in the introductory part of the description, for example a Weibull modeling, or additionally or alternatively consist of a clustering. An alternative approach is to create an empirical reliability model based on data collected for each cluster. A third approach consists in creating an AI (“Artificial Intelligence”) model, both for clustering and for classification. Clustering refers to the detection of accumulations in the master and event data. Each type of data represents a dimension. This creates a multidimensional matrix in which a search is made for clusters or other patterns. Special statistical algorithms can be used for this, for example KI algorithms, which work in particular on the basis of neural networks. To reduce the computational effort, the dimension of the multidimensional matrix can be reduced, for example using KI methods such as principal component analysis. Furthermore, a time period can be selected in which the statistical evaluation should take place or for which device types this should be carried out. Dimensions in which there is no or significantly little data in the relevant period or for the selected device types can be disregarded for the calculation. The service platform enables the mentioned algorithms to be carried out in an application application, as well as to carry out other application applications to carry out the method. The algorithm is executed event-controlled, time-controlled or through interaction with the user.

• Der erste Anwendungsbereich 1 ist der Service-, bzw. Kundenbereich. Im Servicebereich können die Ergebnisse beispielsweise vorteilhaft verwendet werden, um dem Kunden des Herstellers der Automatisierungskomponenten einen optimierten Wartungsplan für seine Automatisierungskomponenten anzubieten. Ebenso können Gewährleistungen wie die Garantie der Automatisierungskomponenten verlängert werden, da eine Abschätzung der Garantiekosten getroffen werden kann. Es kann ein sogenannter Feld-Performance-Index eingeführt werden, welcher Obsoleszenzdaten jener Teile der Automatisierungskomponenten enthält, welche nur mit nicht unerheblichem Aufwand getauscht, bzw. ersetzt werden können.

The results of the statistical evaluation can be used for at least four different areas of application 1 , 2 , 3 , 4th be used:

• The first area of application 1 is the service or customer area. In the service area, for example, the results can be used advantageously to offer the customer of the manufacturer of the automation components an optimized maintenance plan for his automation components. Warranties such as the guarantee of the automation components can also be extended since the guarantee costs can be estimated. A so-called field performance index can be introduced which contains obsolescence data of those parts of the automation components that can only be exchanged or replaced with considerable effort.

In the second application 2 the manufacturer of the automation components can use these results for research and development. The results can be used to specifically close weak points. The results also result in improved knowledge of which components and requirements the respective automation components actually need in the corresponding target industries, which for example also prevents time-consuming and resource-intensive over-engineering.

In the third application 3 the results can be used for production. In particular, series errors in the automation components can be detected and thus eliminated.

A fourth use case 4th relates to sales support. Specifically, the results can be used to suggest the most suitable device for a given application - in terms of reliability.

List of reference symbols

1, 2, 3, 4

Application examples

AL

algorithm

CL

Cloud architecture

DB1, DB2, DB3, DB4

Databases

ED

Event dates

MD

Master data

SP

Service platform

Claims (6)

Method for determining the causes of errors in automation components which are used in automation technology systems, including: - Acquisition of event data (ED) of the automation components from a first database (DB1) or from the automation components directly, which event data (ED) comprise at least one piece of information about an event of an automation component linked to identification information of the respective automation component; - Acquisition of master data (MD), which are linked to the identification data of the respective automation components, from at least one second database (DB2, DB3, DB4); - Correlated storage of the event data (ED) of the automation components and the master data (MD) on a service platform (SP); and - Statistical evaluation of the event data (ED) and master data (MD) to determine the cause of the events that have occurred and / or for predictive prediction, the statistical evaluation including modeling the reliability of the respective types of automation components and / or dimension reduction, classification and clustering . Procedure according to Claim 1 , the event data (ED) containing one or more of the following data: The category of an event that has occurred and in particular the status of the event; - The date or time of an event that occurred; - data relating to the customer at whom the event occurred; - Parameter data of the respective automation component which is assigned to the event; - An error log; - Information about the procedural process in which the automation component is used. Procedure according to Claim 1 or Claim 2 , the master data (MD) containing one or more of the following data about the automation components: the order number of a respective automation component; - Application-specific data of a respective automation component, in particular the industrial sector in which the respective automation component is used; - Production data of the respective automation component, the production data containing: a) production parameters; b) data on the place of production; c) time stamp of completed production steps; d) measured values of the production plant during the production process; e) Any repair measures for the automation component during the production of the automation component; and / or f) lot data of a production supplier. Method according to at least one of the preceding claims, wherein the service platform (SP) is embedded in a cloud architecture and can be contacted via the Internet. Method according to at least one of the preceding claims, wherein the first database (DB1) and the at least second database (DB2, DB3, DB4) are embedded in a cloud architecture (CL) and can be contacted via the Internet. Method according to at least one of the preceding claims, wherein the first database (DB1) and the at least second database (DB2, DB3, DB4) are local databases.

Images