WO2017191253A1 - Alarm handling system and method in plant process automation - Google Patents

Alarm handling system and method in plant process automation Download PDF

Info

Publication number
WO2017191253A1
WO2017191253A1 PCT/EP2017/060651 EP2017060651W WO2017191253A1 WO 2017191253 A1 WO2017191253 A1 WO 2017191253A1 EP 2017060651 W EP2017060651 W EP 2017060651W WO 2017191253 A1 WO2017191253 A1 WO 2017191253A1
Authority
WO
WIPO (PCT)
Prior art keywords
alarm
setpoint
prediction
threshold
process variable
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/EP2017/060651
Other languages
French (fr)
Inventor
Martin Hollender
Benjamnin KLÖPPER
Moncef Chioua
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
ABB Schweiz AG
Original Assignee
ABB Schweiz AG
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by ABB Schweiz AG filed Critical ABB Schweiz AG
Priority to CN201780041170.6A priority Critical patent/CN109416534A/en
Priority to EP17720169.6A priority patent/EP3452878B1/en
Publication of WO2017191253A1 publication Critical patent/WO2017191253A1/en
Priority to US16/178,629 priority patent/US10824963B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring
    • G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
    • G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
    • G05B23/0224—Process history based detection method, e.g. whereby history implies the availability of large amounts of data
    • G05B23/0227—Qualitative history assessment, whereby the type of data acted upon, e.g. waveforms, images or patterns, is not relevant, e.g. rule based assessment; if-then decisions
    • G05B23/0232—Qualitative history assessment, whereby the type of data acted upon, e.g. waveforms, images or patterns, is not relevant, e.g. rule based assessment; if-then decisions based on qualitative trend analysis, e.g. system evolution
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
    • G05B19/41885—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by modeling, simulation of the manufacturing system
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00—Program-control systems
    • G05B2219/30—Nc systems
    • G05B2219/34—Director, elements to supervisory
    • G05B2219/34457—Emit alarm signal

Definitions

  • the invention relates to an alarm handling system and an alarm handling method in plant process automation and/or process automation technology for process plant facilities, like for example in food and beverage industry, oil and gas industry, chemical and pharmaceutical industry .
  • the object of the invention is to provide a more realistic and efficient alarm handling possibility in plant process automation.
  • the alarm handling system in plant process automation comprises a data processing device comprising
  • an alarm configuration device accessing and/or providing alarm configuration information comprising at least one setpoint for one or more determined process variables
  • a prediction device determining and processing the current rate of change of at least one process variable to predict how long it will take and/or the period until and/or predict at which date and/or time a provided setpoint and/or threshold, in particular a predefined setpoint and/or threshold and in particular a consequence threshold, is reached and/or crossed, and/or determines whether and /or when at least one of the monitored and/or determined process variable values will cross the respective setpoint, in particular the alarm setpoint, for example when indicating a return-to- normal scenario.
  • the alarm handling system and in particular the prediction device may provide and ensure a dynamic alarm feedback according to the alternating rate of at least one process variable and a countdown timer and/or trend determination for at least one correlated setpoint or threshold, which means correlated to said at least one process variable.
  • the prediction performed by the prediction device is executed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds.
  • the cycle length may be predefined and the cycle length or cycle period may be a few seconds to minutes.
  • the prediction performed by the prediction device is executed on an event triggered basis, for example when a change of the alternating rate and/or change rate occurs and/or is detected and/or determined.
  • the prediction performed by the prediction device is executed by request, in particular by request of an operator.
  • the prediction can be based on a simple linear interpolation with the current rate of change, but could also be based on more complex prediction models based on historical data, e.g. Hidden Markov Models.
  • the only additional required configuration parameter is the consequence threshold related to an alarm, which should be available anyhow as it is important for the calculation of the alarm setpoint. If the prediction and indication should be limited to the information if and/or when the alarm will return-to-normal even this additional required configuration parameter can be omitted.
  • the approach according to the invention generates the same alarms as the current state of the art approach and does not add risk or significant complexity but allows a more realistic and efficient handling as well as a more sophisticated assessment of alarms.
  • the prediction may be performed by applying a regression algorithm to the at least one respective process variable, which means that the time until the next threshold is crossed or the exceeded threshold is crossed again (return to normal) can be estimated by regression, e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na ' i ' ve Bayes, Random Forrests, Artificial Neural Networks/Deep Learning, or a combination of the two methods regression and classification.
  • regression e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na ' i ' ve Bayes, Random Forrests, Artificial Ne
  • the prediction may be performed by applying a linear regression algorithm or function to the at least one respective process variable.
  • a configuration learner device wherein alarm configuration information, like for example suitable and/or appropriate process variables, corresponding alarm set point and consequence thresholds are derived and/or determined from historical data, in particular stored on at least one historical database, including for example alarm logs and process measurements.
  • a prediction model learner is provided, wherein a prediction model applied and executed that takes the current process signal values as input values which are processed to predict future values and/or a trend for one or more process variable values.
  • the processing predictor device is executing and/or performing the prediction taking into account the most recent readings from the signal and predicting future values and checking if and when a cross consequence threshold or alarm set point and threshold respectively is reached and/or crossed.
  • an alarm display device which is presenting alarm messages in list form and which is updated periodically or event triggered.
  • the predicted information like duration and/or date and/or time is disclosed and/or shown together with the correlated alarm and/or alarm message.
  • the prediction device processes at least one of the following configuration parameters, in particular provided by the alarm configuration device: the alarm limit, the related trip limit, wherein often these are "Hi” and "HiHi” alarms, and the related process signal, which can be shown as an option in the 800xA alarm list already today
  • the alarm handling system means are provided, by which stepwise
  • the alarm configuration information which comprise signal, alarm set point and consequence threshold, can be derived from historical data, in particular alarm logs and process measurements, performed by the configuration learner device,
  • the prediction model is created and/or learned that takes the current signal as input and predicts future values
  • the prediction is performed in the simplest case by linear extrapolation taken into account the most recent readings from the signal and predicting future values and checking if and when the cross consequence threshold or alarm set point,
  • the fourth step may include the filtering (not displaying) of alarms based on the information provided by the third step (e.g. time until Return-to- Normal or additional information like probability or confidence provided by more complex prediction methods) and/or a fifth optional step might include a periodic recalculation of prediction to update the screen, in particular including displaying alarms previously filtered in the fourth step.
  • the object of the invention is also solved by an alarm handling method according to the features of claim 7.
  • the alarm handling method in plant process automation according to the invention stepwise comprises
  • the prediction performed is executed and /or performed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds.
  • the cycle length may be predefined and the cycle length or period may be a few seconds up to one or more minutes.
  • the prediction performed is executed on an event triggered basis, for example in case a change of the alternating rate and/or change rate occurs and/or is detected and/or is determined.
  • the prediction is executed by request, in particular by request of an operator.
  • the prediction can be based on a simple linear interpolation with the current rate of change, but could also be based on more complex prediction models based on historical data, e.g. Hidden Markov Models.
  • the only additional required configuration parameter is the consequence threshold related to an alarm, which should be available anyhow as it is important for the calculation of the alarm setpoint. If the prediction and indication should be limited to the information if and/or when the alarm will return-to-normal even this additional required configuration parameter can be omitted.
  • the approach according to the invention generates the same alarms as the current state of the art approach and does not add risk or significant complexity but allows a more realistic and efficient handling as well as a more sophisticated assessment of alarms.
  • the prediction may be performed by applying a regression algorithm to the at least one respective process variable, which means that the time until the next threshold is crossed or the exceeded threshold is crossed again (return to normal) can be estimated by regression, e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na ' i ' ve Bayes, Random Forrests, Artificial Neural Networks/Deep Learning, or a combination of the two methods regression and classification.
  • regression e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na ' i ' ve Bayes, Random Forrests, Artificial Ne
  • the prediction may be performed by applying a linear regression algorithm or function to the at least one respective process variable.
  • alarm configuration information like for example suitable and/or appropriate process variables, corresponding alarm set point and
  • consequence thresholds are derived and/or determined from historical data, in particular stored on at least one historical database, including for example alarm logs and process measurements.
  • a prediction model is provided or created and/or applied and executed that takes the current process signal values as input values which are processed to predict future values and/ or a trend for one or more process variables.
  • the prediction takes into account the most recent readings from the signal and predicting future values and checking if and when a cross consequence threshold or alarm set point and threshold respectively is reached and/or crossed.
  • alarms and/or alarm messages are displayed in list form and are updated periodically or event triggered.
  • the predicted information like duration and/or date and/or time is disclosed and/or shown together with the correlated alarm and/or alarm message.
  • the operator can use said information to better assess and prioritize which alarm is most critical for the plant and/or which alarm is to act upon first to avoid any damages or disturbances for the respective process.
  • the alarm configuration information which comprise signal, alarm set point and consequence threshold, can be derived from historical data, in particular alarm logs and process measurements, • in an optional second step the prediction model is created and/or learned that takes the current signal as input and predicts future values,
  • the prediction is performed in the simplest case by linear extrapolation taken into account the most recent readings from the signal and predicting future values and checking if and when the cross consequence threshold or alarm set point,
  • FIG. 2 schematic presentation of an exemplary alarm handling system according to the invention
  • Figure 5 exemplary alarm list including predicted trends
  • FIG 1 a state of the art scenario is presented, wherein in plant process automation alarm setpoints are calculated assuming a worst case scenario with regard to the maximum alternation rate for a process variable, like for example if in a tank reaching a level L(trip) causes an automatic shutdown action ,e.g. stop all inflow, the alarm setpoint L(alarm) is calculated in such a way that even with maximum speed of increase (maximum alternation rate) of the level and/or the respective process variable, so that the operator still has enough time to avoid the trip.
  • This situation is taken from IEC 62682. Accordingly in figure 1 all the possible elements related to an implementation of an alarm handling system are disclosed.
  • Figure 1 shows a diagram wherein the y-axis refers to a process variable and the x- axis to the time.
  • an alarm setpoint and a consequence threshold are defined for the exemplary process variable.
  • process variable value reaches and/or crosses the alarm setpoint am alarm is generated.
  • periods and instances of time are determined defining the allowable response time, process dead time, time when the process variable crosses the consequence threshold, process response delay delay time and deadband delay, when the process variable returns to normal values and normal scenario.
  • the measured alternating rate differs from the defined maximum alternating rate thus the calculated and/or determined instances of time are overly pessimistic and/or not correct.
  • an alarm handling system for a plant process automation facility comprising a data processing device with at least one interface, accessing and/or processing one or more process signals and determining corresponding process variable values.
  • an alarm configuration device 20 is provided accessing and/or providing alarm configuration information comprising at least one setpoint 22 for one or more determined process variables 24.
  • a prediction device 30 is provided determining and processing the current rate of change of at least one process variable to predict how long it will take and/or the period until and/or predict at which date and/or time a provided setpoint and/or threshold, in particular a predefined setpoint and/or threshold and in particular a consequence threshold 28, is reached and/or crossed, and/or determines whether and /or when at least one of the monitored and/or determined process variable values 24 will cross the respective setpoint 26, in particular the alarm setpoint 26, for example indicating a return-to- normal scenario.
  • the prediction process performed by the prediction device may be executed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds.
  • the cycle length may be predefined and the cycle length or period may be from a few seconds to minutes.
  • the prediction may be executed on an event triggered basis, for example an alarm, a change of the alternating rate and/or change rate of the respective process variable, in particular of temperature, pressure, or flow, occurs and/or is detected and/or determined, and/or by request.
  • an event triggered basis for example an alarm, a change of the alternating rate and/or change rate of the respective process variable, in particular of temperature, pressure, or flow, occurs and/or is detected and/or determined, and/or by request.
  • the prediction can be based on a simple linear interpolation with the current rate of change, but could also be based on more complex prediction models based on historical data, e.g. Hidden Markov Models.
  • the approach according to the invention generates the same alarms as the current state of the art approach and does not add risk or significant complexity but allows a more realistic and efficient handling as well as a more sophisticated assessment of alarms.
  • the prediction may be performed by applying a regression algorithm to the at least one respective process variable, which means that the time until the next threshold is crossed or the exceeded threshold is crossed again (return to normal) can be estimated by regression, e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na ' i ' ve Bayes, Random Forrests, Artificial Neural Networks/Deep Learning, or a combination of the two methods regression and classification.
  • regression e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na ' i ' ve Bayes, Random Forrests, Artificial Ne
  • a configuration learner device 42 wherein alarm configuration information, like for example suitable and/or appropriate process variables, corresponding alarm setpoints and consequence thresholds are derived and/or determined from historical data, in particular stored on at least one historical database 44,46, including for example alarm logs and process measurements.
  • a prediction model learner 48 may be provided, wherein a prediction model is applied and executed that takes the current process signal values as input values which are processed to predicts future values and/ or a trend for one or more process variable values.
  • measurements and alarm and event logs can be leveraged by the configuration learner device 42 and /or a prediction model learner device 46 to re-engineer the alarm set points 26 in an automated and data driven fashion.
  • the predictor device 30 is executing and/or performing the prediction taking into account the most recent readings from the signal and predicting future values and checking if and when a cross consequence threshold or alarm set point and threshold respectively is reached and/or crossed.
  • An alarm display device 40 is provided, which is presenting alarm messages in list form comprising trend indicators for specific process variables and alarms and/or setpoints, wherein the presentation and the respective alarm information is updated, wherein the predicted information like duration and/or date and/or time is disclosed and/or shown together with the correlated alarm and/or alarm message, as disclosed in figure 5, periodically and/or event triggered and/or by request. Also a prioritization matrix, as disclosed in figure 4, may be provided and presented.
  • the operator can use said information to better assess and prioritize which alarm is most critical for the plant and/or which alarm is to act upon first to avoid any damages or disturbances for the respective process.
  • FIG 3 a schematic presentation of an exemplary embodiment of the alarm handling method according to the invention is disclosed alarm handling system means are provided, by which stepwise
  • the alarm configuration information which comprise signal, alarm set point and consequence threshold, can be derived from histor- ical data, in particular alarm logs and process measurements, in particular performed by the configuration learner device 42,
  • the prediction model is created and/or learned that takes the current signal as input and predicts future values
  • the prediction is performed in the simplest case by linear extrapolation taken into account the most recent readings from the signal 24 and predicting future values and checking if and when the respective consequence threshold 28 or alarm set point 26 is reached and/or crossed, wherein the prediction firstly may be triggered by alarm activation
  • the alarm display or list is updated accordingly.
  • the prediction process performed by the prediction device may be executed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds.
  • the cycle length may be predefined and the cycle length or period may be from a few seconds to minutes.
  • the prediction may be executed on an event triggered basis, for example an alarm, a change of the alternating rate and/or change rate of the respective process variable, in particular of temperature, pressure, or flow, occurs and/or is detected and/or determined, and/or by request.
  • an event triggered basis for example an alarm, a change of the alternating rate and/or change rate of the respective process variable, in particular of temperature, pressure, or flow, occurs and/or is detected and/or determined, and/or by request.
  • the filling height is calculated in such a way for the triggering of the alarm that at a maximum filling speed the operator disposes of a certain minimum response time, e.g. three minutes.
  • the required response time is usually covered by the alarm priority, e.g. priority "red” (high) because a quick reaction is necessary and the potential damage quite high.
  • the assigned priority is based on the highest assumed filling speed. The actual filling speed however may be much lower in that specific case, insofar the indicated priority is considered "wrong" It is important to know that from the alarm it cannot be deduced what may be the possible time period until the overflow happens, which however is very important for the operator.
  • FIG 4 an exemplary prioritization matrix for alarm assessment and handling is presented. It shows the current state of the art how static off-line configuration of alarm priority is currently done.
  • the table of consequences on the left shows an example how four different levels of potential damage can be defined.
  • the first row shows the most severe potential consequences in case the operator does not react to the alarm. In this example this means either that at least one person will die, or that more than fifty barrel of crude oil will be released into the ocean or that a financial damage bigger than five million euro will occur.
  • the matrix on the right introduces three time ranges in which the operator has to react (immediate, prompt, soon). The matrix determines the priority for the operator: for example an alarm requiring immediate response and most severe potential consequences gets the priority "emergency". If such an alarm occurs, the operators know that they should focus on this alarm first.
  • an exemplary list of alarm messages comprising predicted trends and time schedules for the respective setpoint and/or threshold for several alarms and/or alarm messages referring to different measured process temperature and pressure variables, in particular of sensors of different technical equipment installed in the plant.
  • the difficult question for the operator is which alarm to handle first. In this case there are three alarms with priority "critical” which usually means that these alarms need immediate response and that a severe damage can occur if no action is taken.
  • the additional information from this invention shows that "p123” is rapidly rising and that a trip is predicted in five minutes, whereas for the other two critical alarms more time is predicted (30 minutes and more than 60 minutes) It is rather obvious that the operator should focus on "p123" first. In a conventional system there would be no easy way to know on which of the three critical alarms to focus first.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Manufacturing & Machinery (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

Alarm handling method and system in plant process automation Alarm handling method and system in plant process automation including a data processing device comprising • · at least one interface (10), accessing or processing one or more process signals and determining corresponding process variables (24), • · an alarm configuration device (20), accessing or providing alarm configuration information comprising at least one setpoint for one or more determined process variables, • · a prediction device (30) determining and processing the current rate of change of at least one process variable to predict how long it will take or predict at which time a provided setpoint or threshold, in particular a predefined setpoint or threshold and in particular a consequence threshold (28), is reached or crossed, or determines whether and when at least one of the monitored or determined process variable values will cross the respective setpoint, in particular the alarm setpoint (26), for example when indicating a return-to-normal scenario.

Description

Alarm handling system and method in plant process automation
Description
The invention relates to an alarm handling system and an alarm handling method in plant process automation and/or process automation technology for process plant facilities, like for example in food and beverage industry, oil and gas industry, chemical and pharmaceutical industry .
Today in plant process automation alarm setpoints are calculated or determined assuming a worst case scenario with regard to the alternation rate for a process variable, like for example if in a tank reaching a level L(trip) causes an automatic shutdown action ,e.g. stop all inflow, the alarm setpoint L(alarm) is calculated in such a way that even with maximum speed of increase (maximum alternation rate) of the level and/or the respective process variable, so that the operator still has enough time to avoid the trip. This situation is disclosed in figure 1 which is taken from I EC 62682 disclosing the common alarm handling situation in plant process automation.
Disadvantageously, even in case if the rate of change is lower than the maximum alternation rate of the respective process variable and accordingly the rate of change of a respective level is lower than the worst case scenario, wherein the alarm setpoint could be much closer to the consequence threshold still giving the operator sufficient time to react, the present or real rate of change is not taken into account and the once defined alarm setpoint may not be adapted.
Thus, today there is no dynamic feedback adaption of alarm setpoints in plant process automation, and accordingly a more realistic and more efficient assessment and handling of alarms in process plant automation and/or accordingly a more precise assessment of alarms is not available and/or not possible.
Thus the object of the invention is to provide a more realistic and efficient alarm handling possibility in plant process automation.
This object is solved by an alarm handling system in plant process automation according to the features of claim 1 . Further embodiments as well as an alarm handling method are disclosed in further claims and the following description.
The alarm handling system in plant process automation according to the invention comprises a data processing device comprising
• at least one interface, accessing and/or processing one or more process
signals and determining corresponding process variable values,
• an alarm configuration device, accessing and/or providing alarm configuration information comprising at least one setpoint for one or more determined process variables,
a prediction device determining and processing the current rate of change of at least one process variable to predict how long it will take and/or the period until and/or predict at which date and/or time a provided setpoint and/or threshold, in particular a predefined setpoint and/or threshold and in particular a consequence threshold, is reached and/or crossed, and/or determines whether and /or when at least one of the monitored and/or determined process variable values will cross the respective setpoint, in particular the alarm setpoint, for example when indicating a return-to- normal scenario.
Accordingly the alarm handling system and in particular the prediction device may provide and ensure a dynamic alarm feedback according to the alternating rate of at least one process variable and a countdown timer and/or trend determination for at least one correlated setpoint or threshold, which means correlated to said at least one process variable. In a further embodiment the prediction performed by the prediction device is executed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds. The cycle length may be predefined and the cycle length or cycle period may be a few seconds to minutes.
In a further embodiment the prediction performed by the prediction device is executed on an event triggered basis, for example when a change of the alternating rate and/or change rate occurs and/or is detected and/or determined.
In a further embodiment the prediction performed by the prediction device is executed by request, in particular by request of an operator.
The prediction can be based on a simple linear interpolation with the current rate of change, but could also be based on more complex prediction models based on historical data, e.g. Hidden Markov Models.
The only additional required configuration parameter is the consequence threshold related to an alarm, which should be available anyhow as it is important for the calculation of the alarm setpoint. If the prediction and indication should be limited to the information if and/or when the alarm will return-to-normal even this additional required configuration parameter can be omitted.
Advantageously the approach according to the invention generates the same alarms as the current state of the art approach and does not add risk or significant complexity but allows a more realistic and efficient handling as well as a more sophisticated assessment of alarms.
Furthermore, the prediction may be performed by applying a regression algorithm to the at least one respective process variable, which means that the time until the next threshold is crossed or the exceeded threshold is crossed again (return to normal) can be estimated by regression, e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na'i've Bayes, Random Forrests, Artificial Neural Networks/Deep Learning, or a combination of the two methods regression and classification.
In a further embodiment the prediction may be performed by applying a linear regression algorithm or function to the at least one respective process variable.
In a further embodiment a configuration learner device is provided, wherein alarm configuration information, like for example suitable and/or appropriate process variables, corresponding alarm set point and consequence thresholds are derived and/or determined from historical data, in particular stored on at least one historical database, including for example alarm logs and process measurements.
In a further embodiment a prediction model learner is provided, wherein a prediction model applied and executed that takes the current process signal values as input values which are processed to predict future values and/or a trend for one or more process variable values.
In a further embodiment the processing predictor device is executing and/or performing the prediction taking into account the most recent readings from the signal and predicting future values and checking if and when a cross consequence threshold or alarm set point and threshold respectively is reached and/or crossed.
In a further embodiment an alarm display device is provided, which is presenting alarm messages in list form and which is updated periodically or event triggered.
In a further embodiment the predicted information like duration and/or date and/or time is disclosed and/or shown together with the correlated alarm and/or alarm message.
The operator can use said information to better assess and prioritize which alarm is most critical for the plant and/or which alarm is to act upon first to avoid any damages or disturbances for the respective process. In a further embodiment the prediction device processes at least one of the following configuration parameters, in particular provided by the alarm configuration device: the alarm limit, the related trip limit, wherein often these are "Hi" and "HiHi" alarms, and the related process signal, which can be shown as an option in the 800xA alarm list already today
In a further embodiment in case when even the alarm set points and the related process signal are not readily available, e.g. hard-coded in legacy controllers, historical process measurements and alarm and event logs could be leveraged by a configuration learner device and /or a prediction model learner device to re-engineer the alarm set points in an automated and data driven fashion.
According to another embodiment the alarm handling system means are provided, by which stepwise
• in a first optional step, the alarm configuration information, which comprise signal, alarm set point and consequence threshold, can be derived from historical data, in particular alarm logs and process measurements, performed by the configuration learner device,
• in an optional second step the prediction model is created and/or learned that takes the current signal as input and predicts future values,
• in a third and first mandatory step, the prediction is performed in the simplest case by linear extrapolation taken into account the most recent readings from the signal and predicting future values and checking if and when the cross consequence threshold or alarm set point,
• In the fourth and second mandatory step, the alarm display or list is updated accordingly.
In a further embodiment the fourth step may include the filtering (not displaying) of alarms based on the information provided by the third step (e.g. time until Return-to- Normal or additional information like probability or confidence provided by more complex prediction methods) and/or a fifth optional step might include a periodic recalculation of prediction to update the screen, in particular including displaying alarms previously filtered in the fourth step. Moreover, the object of the invention is also solved by an alarm handling method according to the features of claim 7.
The alarm handling method in plant process automation according to the invention stepwise comprises
• accessing and/or processing of one or more process signals and determining corresponding process variables and/or process variable values,
• accessing and/or providing alarm configuration information comprising at least one setpoint and/or threshold for one or more determined process variables,
• determining and processing the current rate of change of at least one process variable and predicting how long it will take and/or predicting the period until and/or predicting at which date and/or time a provided setpoint and/or threshold, in particular a predefined setpoint or threshold and in particular a consequence threshold, is reached and/or crossed, and/or determines whether and /or when at least one of the monitored and/or determined process variable values will cross the respective setpoint, in particular the alarm setpoint indicating a return-to-normal scenario.
In a further embodiment the prediction performed is executed and /or performed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds. The cycle length may be predefined and the cycle length or period may be a few seconds up to one or more minutes.
In a further embodiment the prediction performed is executed on an event triggered basis, for example in case a change of the alternating rate and/or change rate occurs and/or is detected and/or is determined.
In a further embodiment the prediction is executed by request, in particular by request of an operator. The prediction can be based on a simple linear interpolation with the current rate of change, but could also be based on more complex prediction models based on historical data, e.g. Hidden Markov Models.
The only additional required configuration parameter is the consequence threshold related to an alarm, which should be available anyhow as it is important for the calculation of the alarm setpoint. If the prediction and indication should be limited to the information if and/or when the alarm will return-to-normal even this additional required configuration parameter can be omitted.
Advantageously the approach according to the invention generates the same alarms as the current state of the art approach and does not add risk or significant complexity but allows a more realistic and efficient handling as well as a more sophisticated assessment of alarms.
Furthermore, the prediction may be performed by applying a regression algorithm to the at least one respective process variable, which means that the time until the next threshold is crossed or the exceeded threshold is crossed again (return to normal) can be estimated by regression, e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na'i've Bayes, Random Forrests, Artificial Neural Networks/Deep Learning, or a combination of the two methods regression and classification.
In a further embodiment the prediction may be performed by applying a linear regression algorithm or function to the at least one respective process variable.
In a further embodiment alarm configuration information, like for example suitable and/or appropriate process variables, corresponding alarm set point and
consequence thresholds are derived and/or determined from historical data, in particular stored on at least one historical database, including for example alarm logs and process measurements. In a further embodiment a prediction model is provided or created and/or applied and executed that takes the current process signal values as input values which are processed to predict future values and/ or a trend for one or more process variables.
In a further embodiment the prediction takes into account the most recent readings from the signal and predicting future values and checking if and when a cross consequence threshold or alarm set point and threshold respectively is reached and/or crossed.
In a further embodiment alarms and/or alarm messages are displayed in list form and are updated periodically or event triggered.
In a further embodiment the predicted information like duration and/or date and/or time is disclosed and/or shown together with the correlated alarm and/or alarm message.
The operator can use said information to better assess and prioritize which alarm is most critical for the plant and/or which alarm is to act upon first to avoid any damages or disturbances for the respective process.
In a further embodiment at least one of the following configuration parameters, the alarm limit, the related trip limit, wherein often these are "Hi" and "HiHi" alarms, and the related process signal, which can be shown as an option in the 800xA alarm list already today, is processed.
In a further embodiment in case when even the alarm set points and the related process signal are not readily available, e.g. hard-coded in legacy controllers, historical process measurements and alarm and event logs could be leveraged to re- engineer the alarm set points in an automated and data driven fashion.
According to another embodiment the alarm handling method:
• in a first optional step, the alarm configuration information, which comprise signal, alarm set point and consequence threshold, can be derived from historical data, in particular alarm logs and process measurements, • in an optional second step the prediction model is created and/or learned that takes the current signal as input and predicts future values,
• in a third and first mandatory step, the prediction is performed in the simplest case by linear extrapolation taken into account the most recent readings from the signal and predicting future values and checking if and when the cross consequence threshold or alarm set point,
• In the fourth and second mandatory step, the alarm display or list is updated accordingly.
The claimed invention and advantageous embodiments are disclosed and explained in more detail according to several figures and execution examples.
The Figures disclose:
Figure 1 alarm handling according to the state of the art
Figure 2 schematic presentation of an exemplary alarm handling system according to the invention
Figure 3 schematic presentation of an exemplary alarm handling method according to the invention
Figure 4 prioritization matrix for alarms
Figure 5 exemplary alarm list including predicted trends
In figure 1 a state of the art scenario is presented, wherein in plant process automation alarm setpoints are calculated assuming a worst case scenario with regard to the maximum alternation rate for a process variable, like for example if in a tank reaching a level L(trip) causes an automatic shutdown action ,e.g. stop all inflow, the alarm setpoint L(alarm) is calculated in such a way that even with maximum speed of increase (maximum alternation rate) of the level and/or the respective process variable, so that the operator still has enough time to avoid the trip. This situation is taken from IEC 62682. Accordingly in figure 1 all the possible elements related to an implementation of an alarm handling system are disclosed. Figure 1 shows a diagram wherein the y-axis refers to a process variable and the x- axis to the time. For the exemplary process variable an alarm setpoint and a consequence threshold are defined. When the measures process variable value reaches and/or crosses the alarm setpoint am alarm is generated. According to known systems and methods on the basis of a maximum alternation rate of said process values periods and instances of time are determined defining the allowable response time, process dead time, time when the process variable crosses the consequence threshold, process response delay delay time and deadband delay, when the process variable returns to normal values and normal scenario.
Disadvantageously in most cases the measured alternating rate differs from the defined maximum alternating rate thus the calculated and/or determined instances of time are overly pessimistic and/or not correct.
Assuming that not solely one specific process variable and alarm has to be handled but at least from ten up to several hundred alarms have to be handled this may lead to an incorrect and faulty ranking of said alarms, which may lead to misdiagnoses and finally dysfunction of the whole plant or facility.
In figure 2 an alarm handling system for a plant process automation facility according to the invention is presented comprising a data processing device with at least one interface, accessing and/or processing one or more process signals and determining corresponding process variable values.
Furthermore, an alarm configuration device 20 is provided accessing and/or providing alarm configuration information comprising at least one setpoint 22 for one or more determined process variables 24. A prediction device 30 is provided determining and processing the current rate of change of at least one process variable to predict how long it will take and/or the period until and/or predict at which date and/or time a provided setpoint and/or threshold, in particular a predefined setpoint and/or threshold and in particular a consequence threshold 28, is reached and/or crossed, and/or determines whether and /or when at least one of the monitored and/or determined process variable values 24 will cross the respective setpoint 26, in particular the alarm setpoint 26, for example indicating a return-to- normal scenario. The prediction process performed by the prediction device may be executed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds. The cycle length may be predefined and the cycle length or period may be from a few seconds to minutes.
Furthermore, alternatively or in combination the prediction may be executed on an event triggered basis, for example an alarm, a change of the alternating rate and/or change rate of the respective process variable, in particular of temperature, pressure, or flow, occurs and/or is detected and/or determined, and/or by request.
The prediction can be based on a simple linear interpolation with the current rate of change, but could also be based on more complex prediction models based on historical data, e.g. Hidden Markov Models.
Advantageously, the approach according to the invention generates the same alarms as the current state of the art approach and does not add risk or significant complexity but allows a more realistic and efficient handling as well as a more sophisticated assessment of alarms.
Furthermore, the prediction may be performed by applying a regression algorithm to the at least one respective process variable, which means that the time until the next threshold is crossed or the exceeded threshold is crossed again (return to normal) can be estimated by regression, e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na'i've Bayes, Random Forrests, Artificial Neural Networks/Deep Learning, or a combination of the two methods regression and classification.
Moreover, a configuration learner device 42 is provided, wherein alarm configuration information, like for example suitable and/or appropriate process variables, corresponding alarm setpoints and consequence thresholds are derived and/or determined from historical data, in particular stored on at least one historical database 44,46, including for example alarm logs and process measurements. Additionally, a prediction model learner 48 may be provided, wherein a prediction model is applied and executed that takes the current process signal values as input values which are processed to predicts future values and/ or a trend for one or more process variable values.
Thus, in case when even the alarm set points and the related process signal are not readily available, e.g. hard-coded in legacy controllers, historical process
measurements and alarm and event logs can be leveraged by the configuration learner device 42 and /or a prediction model learner device 46 to re-engineer the alarm set points 26 in an automated and data driven fashion.
The predictor device 30 is executing and/or performing the prediction taking into account the most recent readings from the signal and predicting future values and checking if and when a cross consequence threshold or alarm set point and threshold respectively is reached and/or crossed.
An alarm display device 40 is provided, which is presenting alarm messages in list form comprising trend indicators for specific process variables and alarms and/or setpoints, wherein the presentation and the respective alarm information is updated, wherein the predicted information like duration and/or date and/or time is disclosed and/or shown together with the correlated alarm and/or alarm message, as disclosed in figure 5, periodically and/or event triggered and/or by request. Also a prioritization matrix, as disclosed in figure 4, may be provided and presented.
The operator can use said information to better assess and prioritize which alarm is most critical for the plant and/or which alarm is to act upon first to avoid any damages or disturbances for the respective process.
In figure 3 a schematic presentation of an exemplary embodiment of the alarm handling method according to the invention is disclosed alarm handling system means are provided, by which stepwise
• in a first optional step 50, the alarm configuration information, which comprise signal, alarm set point and consequence threshold, can be derived from histor- ical data, in particular alarm logs and process measurements, in particular performed by the configuration learner device 42,
• in an optional second step 52 the prediction model is created and/or learned that takes the current signal as input and predicts future values,
• in a third and first mandatory step 54, the prediction is performed in the simplest case by linear extrapolation taken into account the most recent readings from the signal 24 and predicting future values and checking if and when the respective consequence threshold 28 or alarm set point 26 is reached and/or crossed, wherein the prediction firstly may be triggered by alarm activation
In the fourth and second mandatory step 56, the alarm display or list is updated accordingly. The prediction process performed by the prediction device may be executed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds. The cycle length may be predefined and the cycle length or period may be from a few seconds to minutes.
Furthermore, alternatively or in combination the prediction may be executed on an event triggered basis, for example an alarm, a change of the alternating rate and/or change rate of the respective process variable, in particular of temperature, pressure, or flow, occurs and/or is detected and/or determined, and/or by request.
For example, in case of a high filling level alarm:
Supposing that a container must not overflow in any case, the filling height is calculated in such a way for the triggering of the alarm that at a maximum filling speed the operator disposes of a certain minimum response time, e.g. three minutes. The required response time is usually covered by the alarm priority, e.g. priority "red" (high) because a quick reaction is necessary and the potential damage quite high. If the alarm is triggered, the assigned priority is based on the highest assumed filling speed. The actual filling speed however may be much lower in that specific case, insofar the indicated priority is considered "wrong" It is important to know that from the alarm it cannot be deduced what may be the possible time period until the overflow happens, which however is very important for the operator.
In the simplest case it could be evaluated as follows: t(overflow) = h(current distance from the edge) / v(current filling speed).
It would also be interesting to indicate this time in the alarm. If the plant operator needs to decide on which of the two alarms of equal priority should be handled first, the invention actually helps to identify the more "acute" alarm.
In figure 4 an exemplary prioritization matrix for alarm assessment and handling is presented. It shows the current state of the art how static off-line configuration of alarm priority is currently done. The table of consequences on the left shows an example how four different levels of potential damage can be defined. The first row shows the most severe potential consequences in case the operator does not react to the alarm. In this example this means either that at least one person will die, or that more than fifty barrel of crude oil will be released into the ocean or that a financial damage bigger than five million euro will occur. In addition to these four levels of severity, the matrix on the right introduces three time ranges in which the operator has to react (immediate, prompt, soon). The matrix determines the priority for the operator: for example an alarm requiring immediate response and most severe potential consequences gets the priority "emergency". If such an alarm occurs, the operators know that they should focus on this alarm first.
In figure 5 an exemplary list of alarm messages is disclosed comprising predicted trends and time schedules for the respective setpoint and/or threshold for several alarms and/or alarm messages referring to different measured process temperature and pressure variables, in particular of sensors of different technical equipment installed in the plant. The difficult question for the operator is which alarm to handle first. In this case there are three alarms with priority "critical" which usually means that these alarms need immediate response and that a severe damage can occur if no action is taken. The additional information from this invention shows that "p123" is rapidly rising and that a trip is predicted in five minutes, whereas for the other two critical alarms more time is predicted (30 minutes and more than 60 minutes) It is rather obvious that the operator should focus on "p123" first. In a conventional system there would be no easy way to know on which of the three critical alarms to focus first.

Claims

Claims
1 . Alarm handling system in plant process automation with a data processing
device comprising
• at least one interface (10), accessing and/or processing one or more process signals and determining corresponding process variables (24),
• an alarm configuration device (20), accessing and/or providing alarm
configuration information comprising at least one setpoint for one or more determined process variables,
• a prediction device (30) determining and processing the current rate of change of at least one process variable to predict how long it will take and/or the period until and/or predict at which date and/or time a provided setpoint and/or threshold, in particular a predefined setpoint and/or threshold and in particular a consequence threshold (28), is reached and/or crossed, and/or determines whether and /or when at least one of the monitored and/or determined process variable values will cross the respective setpoint, in particular the alarm setpoint (26), for example when indicating a return-to-normal scenario.
2. Alarm handling system according ti claim 1 , characterized in that the prediction performed by the prediction device (30) is executed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds, and/or wherein the cycle length may freely be defined and in particular the cycle length or period may extend from a few seconds to minutes.
3. Alarm handling system according to one of the preceding claims, characterized in that the prediction performed by the prediction device is executed on an event triggered basis, for example when a change of the alternating rate and/or change rate occurs and/or is detected and/or determined.
4. Alarm handling system according to one of the preceding claims, characterized in that a configuration learner device (42) is provided, wherein alarm
configuration information, like for example suitable and/or appropriate process variables, corresponding alarm set point and consequence thresholds are derived and/or determined from historical data, in particular stored on at least one historical database, including for example alarm logs and process measurements.
5. Alarm handling system according to one of the preceding claims, characterized in that a prediction model learner (48) is provided, wherein a prediction model is applied and executed that takes the current process signal values as input values which are processed to predicts future values and/ or a trend for one or more process variable values.
6. Alarm handling system according to one of the preceding claims, characterized in that an alarm display device is provided, which is presenting alarm messages in list form and which is updated periodically or event triggered and wherein the predicted information like duration and/or date and/or time is disclosed and/or shown together with the correlated alarm and/or alarm message.
7. Alarm handling method in plant process automation comprising steps of
• accessing and/or processing of one or more process signals and determining corresponding process variables and/or process variable values,
• accessing and/or providing alarm configuration information comprising at least one setpoint and/or threshold for one or more determined process variables,
• determining and processing the current rate of change of at least one process variable and predicting how long it will take and/or predicting the period until and/or predicting at which date and/or time a provided setpoint and/or threshold, in particular a predefined setpoint or threshold and in particular a consequence threshold, is reached and/or crossed, and/or determines whether and /or when at least one of the monitored and/or determined process variable values will cross the respective setpoint, in particular the alarm setpoint indicating a return-to-normal scenario.
8. Alarm handling method according to claim 7, characterized in that the prediction performed is executed and /or performed on a cyclic and/or periodic basis in a predefined cycle length, in particular every 30 seconds, wherein the cycle length may freely be defined and/or the cycle length or period may extend from a few seconds up to one or more minutes.
9. Alarm handling method according to claim 7 or 8, characterized in that the
prediction performed is executed on an event triggered basis, for example in case a change of the alternating rate and/or change rate occurs and/or is detected and/or is determined, and/or by request, in particular by request of an operator.
10. Alarm handling method according to one of claims 7 to 9, characterized in that prediction can be based on a simple linear interpolation with the current rate of change, but could also be based on more complex prediction models based on historical data, e.g. Hidden Markov Models, and/or the prediction may be performed by applying a regression algorithm to the at least one respective process variable, which means that the time until the next threshold is crossed or the exceeded threshold is crossed again (return to normal) can be estimated by regression, e.g. Polynomial Regression, Gradient Boosting Trees, Stochastic gradient boosting, Artificial Neural Networks/Deep Learning, Gaussian Process, Kernel Regression, or classification, e.g. Decision Trees, Support Vector Machines, Logistics Regression, Na'ive Bayes, Random Forrests, Artificial Neural Networks/Deep Learning, or a combination of the two methods regression and classification or is performed by applying a linear regression algorithm or function to the at least one respective process variable.
PCT/EP2017/060651 2016-05-04 2017-05-04 Alarm handling system and method in plant process automation Ceased WO2017191253A1 (en)

Priority Applications (3)

Application Number Priority Date Filing Date Title
CN201780041170.6A CN109416534A (en) 2016-05-04 2017-05-04 Alarm handling system and method in factory process automation
EP17720169.6A EP3452878B1 (en) 2016-05-04 2017-05-04 Alarm handling system and method in plant process automation
US16/178,629 US10824963B2 (en) 2016-05-04 2018-11-02 Alarm handling system and method in plant process automation

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
EP16168327 2016-05-04
EP16168327.1 2016-05-04

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/178,629 Continuation US10824963B2 (en) 2016-05-04 2018-11-02 Alarm handling system and method in plant process automation

Publications (1)

Publication Number Publication Date
WO2017191253A1 true WO2017191253A1 (en) 2017-11-09

Family

ID=55919664

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2017/060651 Ceased WO2017191253A1 (en) 2016-05-04 2017-05-04 Alarm handling system and method in plant process automation

Country Status (4)

Country Link
US (1) US10824963B2 (en)
EP (1) EP3452878B1 (en)
CN (1) CN109416534A (en)
WO (1) WO2017191253A1 (en)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3287960B1 (en) * 2016-08-25 2024-05-15 ABB Schweiz AG Computer system and method to process alarm signals
JP6794919B2 (en) * 2017-04-28 2020-12-02 横河電機株式会社 Process control system and data processing method
CN110136400B (en) * 2019-03-29 2021-04-23 广州明珞汽车装备有限公司 Electrode cap replacement alarm method and system
EP3726318B1 (en) * 2019-04-17 2022-07-13 ABB Schweiz AG Computer-implemented determination of a quality indicator of a production batch-run that is ongoing
US11099529B2 (en) 2019-07-23 2021-08-24 International Business Machines Corporation Prediction optimization for system level production control
JP2021089505A (en) * 2019-12-03 2021-06-10 株式会社日立製作所 Monitoring support device and monitoring support method
CN116189399A (en) * 2022-12-22 2023-05-30 北京天河地塬安防技术服务有限公司 Alarm information management method and device, storage medium and electronic equipment

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003081183A1 (en) * 2002-03-26 2003-10-02 Endress + Hauser Gmbh + Co. Kg Arrangement for the operation of field devices

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6453207B1 (en) * 1999-07-01 2002-09-17 Donald S. Holmes Method, system and article of manufacture for operating a process
CN100410825C (en) * 2004-04-22 2008-08-13 横河电机株式会社 Plant Operation Support System
JP4789277B2 (en) * 2004-04-22 2011-10-12 横河電機株式会社 Plant operation support device
EP2221696A1 (en) * 2009-02-12 2010-08-25 Siemens Aktiengesellschaft Method and device for checking a control program of an industrial device
US8818562B2 (en) * 2011-06-04 2014-08-26 Invensys Systems, Inc. Simulated fermentation process
US9529348B2 (en) * 2012-01-24 2016-12-27 Emerson Process Management Power & Water Solutions, Inc. Method and apparatus for deploying industrial plant simulators using cloud computing technologies
US10282676B2 (en) * 2014-10-06 2019-05-07 Fisher-Rosemount Systems, Inc. Automatic signal processing-based learning in a process plant
US10803636B2 (en) * 2013-03-15 2020-10-13 Fisher-Rosemount Systems, Inc. Graphical process variable trend monitoring, predictive analytics and fault detection in a process control system
WO2015104691A2 (en) * 2014-01-13 2015-07-16 Brightsource Industries (Israel) Ltd. Systems, methods, and devices for detecting anomalies in an industrial control system
US10386820B2 (en) * 2014-05-01 2019-08-20 Johnson Controls Technology Company Incorporating a demand charge in central plant optimization
CN109542740B (en) * 2017-09-22 2022-05-27 阿里巴巴集团控股有限公司 Abnormality detection method and apparatus
CN110324168A (en) * 2018-03-30 2019-10-11 阿里巴巴集团控股有限公司 Anomalous event monitoring method and device and electronic equipment
CN110489306A (en) * 2019-08-26 2019-11-22 北京博睿宏远数据科技股份有限公司 A kind of alarm threshold value determines method, apparatus, computer equipment and storage medium
CN110942210B (en) * 2019-12-11 2023-02-07 中国铁建重工集团股份有限公司 Shield TBM attitude deviation early warning method
CN111145895B (en) * 2019-12-24 2023-10-20 中国科学院深圳先进技术研究院 Abnormal data detection method and terminal equipment
CN111706499B (en) * 2020-06-09 2022-03-01 成都数之联科技有限公司 Predictive maintenance system and method for vacuum pump and automatic vacuum pump purchasing system

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003081183A1 (en) * 2002-03-26 2003-10-02 Endress + Hauser Gmbh + Co. Kg Arrangement for the operation of field devices

Also Published As

Publication number Publication date
CN109416534A (en) 2019-03-01
EP3452878B1 (en) 2025-07-09
US10824963B2 (en) 2020-11-03
US20190073609A1 (en) 2019-03-07
EP3452878A1 (en) 2019-03-13

Similar Documents

Publication Publication Date Title
US10824963B2 (en) Alarm handling system and method in plant process automation
EP3500896B1 (en) Method of monitoring and controlling an industrial process, and a process control system
US10809704B2 (en) Process performance issues and alarm notification using data analytics
US8843221B2 (en) Automation management system and method
JP6529690B1 (en) Support device, learning device, and plant operating condition setting support system
EP3771951B1 (en) Using data from plc systems and data from sensors external to the plc systems for ensuring data integrity of industrial controllers
CN115427907B (en) Intelligent alarm management methods for industrial processes
AU2010236009A1 (en) Method and apparatus for monitoring a process and/or a technical installation
JP2018139085A (en) Method, device, system, and program for abnormality prediction
EP3187950B1 (en) A method for managing alarms in a control system
JP2022182620A (en) Information processing device, prediction method and prediction program
Bao et al. Risk‐based fault diagnosis and safety management for process systems
US20190392703A1 (en) Alarm tuning using alarm and process data for condition monitoring
EP3109806A1 (en) Method for generating prognostic alerts and a human machine interface (hmi) device therefor
JP7416015B2 (en) Information processing device, alarm prediction method, and alarm prediction program
EP4312100A1 (en) Abnormality detection device, abnormality detection program, and learning device
US12061730B2 (en) Security assessment apparatus, security assessment method, and non-transitory computer readable medium
US20250111767A1 (en) AI-Based Alarm Prediction in Process Control Systems
CN117043795A (en) Method and system for event handling in industrial process automation
CN118897522B (en) A control method for intelligent alarm information management terminal based on virtual keyboard
CN110956409A (en) Community gas equipment risk control method and device
Munesawa et al. Development of fault diagnosis system using principal component analysis for intelligent operation support system
WO2025228544A1 (en) Alarm management for autonomous process plants
CN119886781A (en) Real-time monitoring and early warning system and method for process industry and storage medium

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17720169

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2017720169

Country of ref document: EP

Effective date: 20181204

WWG Wipo information: grant in national office

Ref document number: 2017720169

Country of ref document: EP