WO2019234913A1 - 支援装置、学習装置、及びプラント運転条件設定支援システム - Google Patents
支援装置、学習装置、及びプラント運転条件設定支援システム Download PDFInfo
- Publication number
- WO2019234913A1 WO2019234913A1 PCT/JP2018/022018 JP2018022018W WO2019234913A1 WO 2019234913 A1 WO2019234913 A1 WO 2019234913A1 JP 2018022018 W JP2018022018 W JP 2018022018W WO 2019234913 A1 WO2019234913 A1 WO 2019234913A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- state
- unit
- value
- control target
- learning
- Prior art date
Links
Images
Classifications
-
- 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
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
-
- 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
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/048—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators using a predictor
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Definitions
- the present invention relates to a support apparatus for supporting setting of plant operating conditions, a plant operating condition setting support system, and a learning apparatus usable in a plant operating condition setting support system.
- a series of processes are executed by a large number of devices such as reactors and heating furnaces, and a large amount of operation is required to control each of the large numbers of devices. Operating conditions are set.
- a plant where a multi-stage process is executed many manipulated variables can interact in a complex manner, so it is not easy to predict the effects of changes in manipulated variables. The plant is in operation.
- Patent Documents 1 and 2 the conventional risk evaluation technology is mainly used in the stage of designing a plant, the stage of reviewing the operation conditions of the plant, and the like.
- the recent decrease in the number of skilled operators has progressed, and the present inventors need technology that supports the safe and stable operation of the plant by applying the results of risk assessment even during operation of the plant. It was recognized as a problem and the present invention was conceived.
- the present invention has been made in view of such circumstances, and an object of the present invention is to provide a technology that supports setting of operating conditions capable of realizing a suitable operation of a plant.
- a support device includes a state value acquisition unit that acquires a plurality of state values indicating states of a plurality of control target devices during operation of the plurality of control target devices; Based on each of a plurality of state values acquired by the state value acquisition unit, a prediction unit that estimates a predicted value of each state value at a predetermined future time point, and a predetermined time point acquired by the state value acquisition unit The index calculated based on the difference between the respective state values at and the predicted values estimated by the prediction unit at a predetermined time point or a time point after the predetermined time or the change rate of the difference is predetermined. And a notification unit for notifying that when the above condition is met.
- the prediction unit may estimate the predicted value by an estimation algorithm acquired by machine learning based on each past actual value of the plurality of state values. According to this aspect, it is possible to improve the accuracy of estimating the predicted value of the state value at a predetermined time in the future, and therefore it is possible to more accurately detect a behavior different from the past driving performance.
- the index calculation algorithm may be learned based on the history of the difference value or the actual value of the change rate of the difference calculated for each state value and the evaluation of the driving behavior of the specific control target device. According to this aspect, the influence on the operation behavior of the specific control target device can be more accurately indexed and presented to the operator, so that the setting of suitable operation conditions of the plant similar to the skilled operator can be more accurately set in real time. Can help.
- a learning unit that machine-learns the calculation algorithm may be further provided based on the history of the difference value calculated for each state value or the history value of the change rate of the difference and the evaluation of the driving behavior of the specific control target device. .
- the influence on the operation behavior of the specific control target device can be more accurately indexed and presented to the operator, so that the setting of suitable operation conditions of the plant similar to the skilled operator can be more accurately set in real time. Can help.
- the index may include an importance level indicating the magnitude of the influence on the specific control target device and an urgency level indicating the urgency of the influence on the specific control target device.
- the influence on the operation behavior of a specific control target device can be divided into importance and urgency and can be indexed and presented to the operator. Can be supported more accurately.
- a display unit for displaying the importance level and the urgency level on the display device may be further provided.
- the display unit may display a matrix in which importance and urgency are plotted on the vertical axis and the horizontal axis. According to this aspect, the importance and urgency of the influence on the operation behavior of the specific control target device can be presented to the operator in a manner that is easy to visually understand. It is possible to more accurately support setting of operating conditions.
- the display unit may display the difference calculated for each of the plurality of state values or the change rate of the difference on the display device. According to this aspect, since it is possible to present to the operator a state value that can affect the operation behavior of a specific control target device, it is possible to more accurately support the setting of suitable operation conditions in the same plant as a skilled operator in real time. can do.
- This device includes an evaluation acquisition unit that acquires a history of predicted values and actual values of a plurality of state values indicating states of a plurality of control target devices, and an evaluation of driving behavior of a specific control target device, and a plurality of control targets.
- a learning unit that performs machine learning on the calculation algorithm based on the history and the evaluation.
- the learning device includes an evaluation acquisition unit that acquires a history of predicted values and actual values of a plurality of state values indicating states of a plurality of control target devices, and an evaluation of driving behavior of a specific control target device, and a history and evaluation A learning unit that learns an index calculation algorithm based on the learning unit.
- FIG. 1 shows an overall configuration of a plant operation condition setting support system according to an embodiment.
- the plant operating condition setting support system 1 includes a plant 3 for producing chemical products, industrial products, and the like, and a learning device 4 for learning an algorithm used to support setting of operating conditions in the plant 3.
- the plant 3 and the learning device 4 are connected by an arbitrary communication network 2 such as the Internet or an in-house connection system, and are operated in an arbitrary operation mode such as on-premises or edge computing.
- the plant 3 is learned by the learning device 4, the control target device 10 such as a reactor or a heating furnace installed in the plant 3, the control device 20 for setting the operation amount for controlling the operation condition of the control target device 10, and the learning device 4. And an operation condition setting support device 30 that supports the setting of the operation conditions of the plant 3 using the generated algorithm.
- the control target device 10 such as a reactor or a heating furnace installed in the plant 3
- the control device 20 for setting the operation amount for controlling the operation condition of the control target device 10
- the learning device 4 for setting the operation amount for controlling the operation condition of the control target device 10
- an operation condition setting support device 30 that supports the setting of the operation conditions of the plant 3 using the generated algorithm.
- the plant operation condition setting support system 1 incorporates the experience of such a skilled operator and appropriately supports the setting of operation conditions for operating the plant 3 safely and stably.
- an index indicating the importance and urgency of the abnormality is calculated and presented to the operator in real time.
- a state quantity to be monitored among a plurality of control target devices 10 operated in the plant 3, it has been stopped or has failed due to some factor.
- a fault tree analysis in which the identified important factor is an upper event is performed, and the cause, the occurrence route, and the occurrence probability of the lower event that causes the upper event are grasped.
- An index calculation algorithm is created in which the difference between the state quantity of the past and the change rate of the difference is weighted.
- the index calculation algorithm calculates an index indicating the degree of importance and urgency of the influence on the important device from the difference between the past behavior and the current behavior of each state quantity or the change rate of the difference.
- the estimated value of each state quantity in is estimated, and the difference between the predicted value of each state quantity and the current value at a predetermined time point or a time point after the predetermined time point is calculated.
- the value of the state quantity changes due to the change of the type of the control target device 10 constituting the plant 3, the type of process, the rate of change of the state quantity, and the value of the controlled quantity It may be determined depending on the speed to be performed, but may be after several seconds to several minutes, for example.
- an estimation algorithm acquired by machine learning based on each past actual value of a plurality of state values is used.
- the predicted value may be calculated using a mathematical formula, a database, or the like set based on the past actual value.
- this embodiment uses artificial intelligence that systematically learns the personal experience of a skilled operator through machine learning to quickly and accurately detect behaviors that differ from past driving performance. Indices indicating the importance and urgency of the influence of the behavior on the important device can be presented to the operator in real time. Thereby, it is possible to provide accurate support so that conditions for operating the plant 3 safely and stably can be appropriately set regardless of the skill and experience of the operator.
- FIG. 2 shows the configuration of the learning device according to the embodiment.
- the learning device 4 includes a performance value acquisition unit 41, an estimation algorithm learning unit 50, an index evaluation acquisition unit 44, an index calculation algorithm learning unit 51, and a provision unit 49.
- these configurations are realized by a CPU of a computer, a memory, a program loaded in the memory, and the like, but here, functional blocks realized by their cooperation are illustrated. Accordingly, those skilled in the art will understand that these functional blocks can be realized in various forms by hardware only, software only, or a combination thereof.
- the estimation algorithm learning unit 50 includes a plurality of estimation algorithms 43a, 43b,... For calculating respective predicted values of a plurality of state quantities that can cause a stop or failure of an important device, and a plurality of estimation algorithms 43a, 43b,... (Hereinafter collectively referred to as “estimation algorithm 43”), a plurality of estimation algorithm learning units 42a, 42b,... (Hereinafter collectively referred to as “estimation algorithm learning unit 42”). Including.
- the estimation algorithm 43 is used to estimate a predicted value of each state value at a predetermined future time point based on each of a plurality of state values indicating the state of the plant 3.
- the actual value acquisition unit 41 acquires the past actual values of the plurality of state values and the set values of the plurality of control amounts from the plant 3.
- the estimation algorithm learning unit 42 learns the estimation algorithm 43 by machine learning based on the past actual values of the plurality of state values acquired by the actual value acquisition unit 41 and the set values of the plurality of control amounts.
- the estimation algorithm learning unit 42 inputs a plurality of state values at a certain point of time and actual values of a plurality of manipulated variable setting values to the estimation algorithm 43, the estimation algorithm learning unit 42
- the estimation algorithm 43 is learned so that a value close to the actual value is calculated.
- the estimation algorithm learning unit 42 may learn the estimation algorithm 43 by supervised learning using the past actual value acquired by the actual value acquisition unit 41 as teacher data, or by any other known machine learning technique.
- the estimation algorithm 43 may be learned.
- the accuracy of the estimation algorithm 43 can be improved by learning the estimation algorithm 43 using a large number of actual values, it is possible to calculate the predicted value of the state value indicating the state of the plant 3 more accurately.
- An estimation algorithm 43 can be generated. Further, since it is not necessary to develop an advanced simulator for reproducing a complicated process, the time and load required to generate the estimation algorithm 43 can be greatly reduced. In addition, since elements that cannot be reproduced by the simulator can be taken into account, the predicted output value can be calculated more accurately.
- the index calculation algorithm is based on the difference between the predicted value of each state value at a predetermined time point or a time point after the predetermined time point and the actual state value at the predetermined time point, or the change rate of the difference. It is used to calculate an index representing the magnitude and urgency of the influence on the driving behavior of the vehicle.
- the index includes an importance indicating the magnitude of the influence on the important device, and an urgency indicating the urgency of the influence on the important device.
- the importance mainly reflects the difference between the predicted value of each state value and the actual state value and the magnitude of the rate of change of the difference.
- the urgency level mainly reflects each state value. The magnitude of the rate of change of the difference between the current predicted value and the current actual state value, and the difference between the future predicted value of each state value and the current actual state value or the rate of change of the difference are reflected.
- the index evaluation acquisition unit 44 acquires the predicted value calculated for each state value, the history of the actual state value, and the evaluation of the operation behavior of the important device.
- the evaluation of the driving behavior of the important device may be calculated from a history of state values or may be input by an operator.
- the index evaluation acquisition unit 44 acquires a prediction value calculated for each state value, a history of actual value of the actual state value, and an evaluation for the index calculated by the importance calculation algorithm 46 and the urgency calculation algorithm 48. May be.
- the evaluation with respect to the index may be a result of evaluation by the operator.
- the index calculation algorithm learning unit 51 includes an importance calculation algorithm 46 for calculating the importance indicating the magnitude of the influence on the important device, an importance calculation algorithm learning unit 45 for learning the importance calculation algorithm 46, and an important device.
- An urgency calculation algorithm 48 for calculating the urgency indicating the urgency of the influence to be given, and an urgency calculation algorithm learning unit 47 for learning the urgency calculation algorithm 48 are included.
- the importance calculation algorithm learning unit 45 and the urgency calculation algorithm learning unit 47 are based on the prediction value and the history value of the state value acquired by the index evaluation acquisition unit 44 and the evaluation of the driving behavior of the important device. Each of the degree calculation algorithm 46 and the urgency degree calculation algorithm 48 is learned.
- the importance level calculation algorithm learning unit 45 and the urgency level calculation algorithm learning unit 47 are based on the prediction value and actual value history acquired by the index evaluation acquisition unit 44 and the evaluation of the driving behavior of the important device at a certain time or thereafter.
- the index calculated by substituting the difference between the predicted value and the state value at the point of time or the rate of change of the difference into the importance calculation algorithm 46 and the urgency calculation algorithm 48 affects the operation behavior of the important device.
- the importance calculation algorithm learning unit 45 and the urgency calculation algorithm learning unit 47 calculate the importance so that an index with a worse value is calculated when the subsequent driving behavior of the important device is worse than a predetermined evaluation.
- the algorithm 46 and the urgency calculation algorithm 48 are learned, and the subsequent operation behavior of the important device is better than a predetermined evaluation, the importance calculation algorithm 46 and the urgency level are calculated so that a better index is calculated.
- the calculation algorithm 48 is learned.
- the importance calculation algorithm learning unit 45 and the urgency calculation algorithm learning unit 47 include an importance calculation algorithm by supervised learning using the evaluation of the driving behavior of the important device acquired by the index evaluation acquisition unit 44 as teacher data. 46 and the urgency calculation algorithm 48 may be learned, or the importance calculation algorithm 46 and the urgency calculation algorithm 48 may be learned by any other known machine learning technique.
- the index evaluation acquisition unit 44 may acquire the index value calculated or evaluated by the operator himself / herself as teacher data without using the importance calculation algorithm 46 and the urgency calculation algorithm 48.
- the importance calculation algorithm learning unit 45 and the urgency calculation algorithm learning unit 47 calculate the difference between the predicted value and the state value at a certain time point or a subsequent time point, or the change rate of the difference, and the importance calculation algorithm 46 and the emergency value.
- the importance calculation algorithm 46 and the urgency calculation algorithm 48 are learned so that the value of the index acquired by the index evaluation acquisition unit 44 when being input to the degree calculation algorithm 48 is calculated.
- the learning device 4 is shown as a single device, but the learning device 4 is realized by a plurality of servers using cloud computing technology, distributed processing technology, or the like. May be.
- the estimation algorithm 43, the importance calculation algorithm 46 , and the urgency calculation algorithm 48 can be greatly reduced.
- FIG. 3 shows the configuration of the operating condition setting support device and the control device according to the embodiment.
- the control device 20 includes a control unit 21, an operation panel 22, and a result value storage unit 29.
- the operation panel 22 displays various state values indicating the operation state of the plant 3, various operation amount set values set by the control device 20, output values indicating the operation results of the plant 3, and the like on the display device. At the same time, input of set values of various manipulated variables is accepted from the operator.
- the control unit 21 includes an operation amount setting unit 23, a state value acquisition unit 24, a state value transmission unit 25, an actual value transmission unit 26, an index evaluation acquisition unit 27, and an index evaluation transmission unit 28.
- These functional blocks can also be realized in various forms by hardware only, software only, or a combination thereof.
- the operation amount setting unit 23 sets various operation amount setting values received from the operator through the operation panel 22, controls the control target device 10, and displays it on the display device of the operation panel 22.
- the state value acquisition unit 24 acquires various state values indicating the operation state and operation result of the plant 3 from various sensors and measuring devices provided in the control target device 10 and the like, and displays them on the display device of the operation panel 22. While being displayed, it is recorded in the result value storage unit 29.
- the state value transmission unit 25 transmits the state value acquired by the state value acquisition unit 24 to the driving condition setting support device 30.
- the actual value transmission unit 26 transmits the operation amount value set by the operation amount setting unit 23 and the state value stored in the actual value storage unit 29 to the learning device 4.
- the index evaluation acquisition unit 27 acquires an evaluation for the index displayed on the operation panel 22 from the operator.
- the index evaluation acquisition unit 27 may acquire an evaluation indicating that the index displayed on the operation panel 22 is too large or too small from the operator, or obtain a correction value of the index displayed on the operation panel 22 from the operator. You may get it.
- the index evaluation transmission unit 28 transmits the evaluation for the index acquired by the index evaluation acquisition unit 27 from the operator to the learning device 4.
- the evaluation for the index is used for learning the importance calculation algorithm 46 and the urgency calculation algorithm 48 in the learning device 4 as described above.
- the driving condition setting support device 30 includes a control unit 31, an estimation algorithm 43, an importance calculation algorithm 46, and an urgency calculation algorithm 48.
- the control unit 31 includes a state value acquisition unit 32, prediction units 33a, 33b,... (Hereinafter collectively referred to as “prediction unit 33”), an index calculation unit 34, a notification unit 37, a display unit 38, and a learning unit 39. Is provided. These functional blocks can also be realized in various forms by hardware only, software only, or a combination thereof.
- the estimation algorithm 43, the importance calculation algorithm 46, and the urgency calculation algorithm 48 are acquired from the learning device 4 and stored in the storage device.
- the state value acquisition unit 32 acquires a plurality of state values from the state value transmission unit 25 of the control device 20.
- the prediction unit 33 uses the estimation algorithm 43 to calculate a predicted value of each state quantity at a predetermined future time point from the plurality of state values acquired by the state value acquisition unit 32, and stores it in the storage device.
- the index calculator 34 includes an importance calculator 35 and an emergency calculator 36.
- the importance level calculation unit 35 and the urgency level calculation unit 36 are the difference or difference between the state value acquired by the state value acquisition unit 32 and the predicted value of the state value calculated by the prediction unit 33 and stored in the storage device.
- the change rate is calculated, and using the importance calculation algorithm 46 and the urgency calculation algorithm 48, an index indicating the importance and the urgency is calculated.
- the notification unit 37 When the index calculated by the index calculation unit 34 meets a predetermined condition, the notification unit 37 notifies that fact.
- the notification unit 37 may display an indicator calculated constantly during operation of the plant 3 on the operation panel 22 to notify the operator, or when the indicator is a value worse than a predetermined value, that fact. May be displayed on the operation panel 22 to notify the operator. Thereby, it is possible to notify the operator that a state that may affect the important device has occurred.
- the display unit 38 displays on the operation panel 22 a matrix in which the importance level and the urgency level calculated by the index calculation unit 34 are plotted on the vertical axis and the horizontal axis. Thereby, the importance and urgency of the influence of the current state on the important device can be presented to the operator in an easily understandable manner.
- the display unit 38 displays the difference between the predicted value and the state value calculated for each of the plurality of state values or the change rate of the difference on the operation panel 22.
- which state value indicates an abnormality and the degree of abnormality of the state value can be presented to the operator.
- Information to be referred to in order to change the set value of the operation amount can be appropriately provided.
- the learning unit 39 learns the estimation algorithm 43, the importance calculation algorithm 46, or the urgency calculation algorithm 48.
- the learning unit 39 uses the estimation algorithm 43, the importance calculation algorithm learning unit 45, or the emergency algorithm in the same manner as the estimation algorithm learning unit 42, the importance calculation algorithm learning unit 45, or the urgency calculation algorithm learning unit 47 of the learning device 4.
- the degree calculation algorithm learning unit 47 may be relearned.
- the learning device 4 re-learns the estimation algorithm 43, the importance calculation algorithm 46, or the urgency calculation algorithm 48, the learning unit 39 may not be provided.
- FIG. 4 shows an example of a display screen displayed on the display device of the operation panel.
- the display screen displays a process flow diagram of the plant 3, a plurality of state values, predicted values after a predetermined time of those state values, a matrix in which importance and urgency are plotted, and state value transitions. ing.
- the operator determines the set value of the operation amount with reference to the presented information and inputs it to the operation panel 22.
- the operation amount setting unit 23 controls the control target device 10 based on the input set value.
- the index calculation unit 34 calculates the index at a predetermined interval, and the display unit 38 plots the importance and urgency of the index calculated at the predetermined interval in a matrix. That is, the importance and urgency matrix is updated in real time and indicates the state of the plant 3 at that time.
- Patent Documents 1 and 2 also disclose a matrix display. In either case, one axis has importance or influence, and the other axis has an occurrence frequency, and is displayed on the operation panel 22 of the plant 3. It is not updated in real time. Since the matrix display of the present embodiment and the transition of the state value and the predicted value are always displayed on the operation panel 22, even if an abnormality occurs, the operator can change the matrix display and the state value by the change. And while confirming the change of the predicted value in real time, the operation value of the plant 3 can be adjusted by changing the set value of the manipulated variable.
- the technology of the present invention can be applied to both a continuous process plant and a batch process plant.
- 1 plant operation condition setting support system 3 plant, 4 learning device, 10 control target device, 20 control device, 21 control unit, 22 operation panel, 23 operation amount setting unit, 24 status value acquisition unit, 25 status value transmission unit, 26 actual value transmission unit, 27 index evaluation acquisition unit, 28 index evaluation transmission unit, 29 actual value storage unit, 30 operating condition setting support device, 31 control unit, 32 state value acquisition unit, 33 prediction unit, 34 index calculation unit, 37 notification section, 38 display section, 39 learning section, 41 actual value acquisition section, 42 estimation algorithm learning section, 43 estimation algorithm, 44 index evaluation acquisition section, 45 importance calculation algorithm learning section, 46 importance calculation algorithm, 47 emergency Degree calculation algorithm learning unit, 48 Urgency calculation algorithm, 49 Providing unit, 50 Algorithm learning unit, 51 index calculation algorithm learning unit.
- the present invention can be used for a support device for supporting the setting of operation conditions of a plant.
Landscapes
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Automation & Control Theory (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Data Mining & Analysis (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
Description
Claims (11)
- 複数の制御対象装置の運転中に、前記複数の制御対象装置の状態を示す複数の状態値を取得する状態値取得部と、
前記状態値取得部により取得された複数の状態値のそれぞれに基づいて、将来の所定の時点におけるそれぞれの状態値の予測値を推定する予測部と、
前記状態値取得部により取得された、前記所定の時点におけるそれぞれの状態値と、前記予測部により推定された、前記所定の時点又は前記所定の時点よりも後の時点におけるそれぞれの予測値との間の差又は差の変化率に基づいて算出される指標が所定の条件に合致した場合に、その旨を報知する報知部と、
を備えることを特徴とする支援装置。 - 前記予測部は、前記複数の状態値のそれぞれの過去の実績値に基づく機械学習により獲得された推定アルゴリズムにより前記予測値を推定することを特徴とする請求項1に記載の支援装置。
- それぞれの状態値について算出された前記差又は差の変化率が特定の制御対象装置の運転挙動に与える影響の大きさに基づいてそれぞれの差又は差の変化率に重みが付された所定の指標算出アルゴリズムにより前記指標を算出する算出部を更に備えることを特徴とする請求項1又は2に記載の支援装置。
- 前記指標算出アルゴリズムは、それぞれの状態値について算出された前記予測値及び実績値の履歴と、前記特定の制御対象装置の運転挙動の評価とに基づいて学習されたものであることを特徴とする請求項3に記載の支援装置。
- それぞれの状態値について算出された前記予測値及び実績値の履歴と、前記特定の制御対象装置の運転挙動の評価とに基づいて、前記算出アルゴリズムを機械学習する学習部を更に備えることを特徴とする請求項4に記載の支援装置。
- 前記指標は、前記特定の制御対象装置に与える影響の大きさを示す重要度と、前記特定の制御対象装置に与える影響の緊急性を示す緊急度とを含むことを特徴とする請求項3から5のいずれかに記載の支援装置。
- 前記重要度と前記緊急度を表示装置に表示する表示部を更に備えることを特徴とする請求項6に記載の支援装置。
- 前記表示部は、前記重要度と前記緊急度を縦軸及び横軸に取ってプロットしたマトリクスを表示することを特徴とする請求項7に記載の支援装置。
- 前記表示部は、前記複数の状態値のそれぞれについて算出された前記差又は差の変化率をそれぞれ前記表示装置に表示することを特徴とする請求項7又は8に記載の支援装置。
- 複数の制御対象装置の状態を示す複数の状態値の予測値及び実績値の履歴と、特定の制御対象装置の運転挙動の評価とを取得する評価取得部と、
前記複数の制御対象装置の運転中に、前記複数の状態値の予測値と実績値との間の差又は差の変化率が前記特定の制御対象装置の運転挙動に与える影響の大きさを表す指標を算出するための指標算出アルゴリズムを、前記履歴及び前記評価に基づいて機械学習する学習部と、
を備えることを特徴とする学習装置。 - プラントの運転条件の設定を支援する支援装置と、
前記支援装置において使用される指標算出アルゴリズムを学習する学習装置と、
を備え、
前記支援装置は、
前記プラントに含まれる複数の制御対象装置の運転中に、前記複数の制御対象装置の状態を示す複数の状態値を取得する状態値取得部と、
前記状態値取得部により取得された複数の状態値のそれぞれに基づいて、将来の所定の時点におけるそれぞれの状態値の予測値を推定する予測部と、
前記状態値取得部により取得された、前記所定の時点におけるそれぞれの状態値と、前記予測部により推定された、前記所定の時点又は前記所定の時点よりも後の時点におけるそれぞれの予測値との間の差又は差の変化率に基づいて、前記指標算出アルゴリズムにより算出される指標が所定の条件に合致した場合に、その旨を報知する報知部と、
を備え、
前記学習装置は、
複数の制御対象装置の状態を示す複数の状態値の予測値及び実績値の履歴と、特定の制御対象装置の運転挙動の評価とを取得する評価取得部と、
前記履歴及び前記評価に基づいて前記指標算出アルゴリズムを学習する学習部と、
を備えることを特徴とするプラント運転条件設定支援システム。
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2018558257A JP6529690B1 (ja) | 2018-06-08 | 2018-06-08 | 支援装置、学習装置、及びプラント運転条件設定支援システム |
PCT/JP2018/022018 WO2019234913A1 (ja) | 2018-06-08 | 2018-06-08 | 支援装置、学習装置、及びプラント運転条件設定支援システム |
AU2018426458A AU2018426458B2 (en) | 2018-06-08 | 2018-06-08 | Assistance device, learning device, and plant operation condition setting assistance system |
TW108110188A TWI710873B (zh) | 2018-06-08 | 2019-03-25 | 支援裝置、學習裝置以及廠房運轉條件設定支援系統 |
US17/113,400 US20210088986A1 (en) | 2018-06-08 | 2020-12-07 | Assistance device, learning device, and plant operation condition setting assistance system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2018/022018 WO2019234913A1 (ja) | 2018-06-08 | 2018-06-08 | 支援装置、学習装置、及びプラント運転条件設定支援システム |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US17/113,400 Continuation US20210088986A1 (en) | 2018-06-08 | 2020-12-07 | Assistance device, learning device, and plant operation condition setting assistance system |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2019234913A1 true WO2019234913A1 (ja) | 2019-12-12 |
Family
ID=66821623
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2018/022018 WO2019234913A1 (ja) | 2018-06-08 | 2018-06-08 | 支援装置、学習装置、及びプラント運転条件設定支援システム |
Country Status (5)
Country | Link |
---|---|
US (1) | US20210088986A1 (ja) |
JP (1) | JP6529690B1 (ja) |
AU (1) | AU2018426458B2 (ja) |
TW (1) | TWI710873B (ja) |
WO (1) | WO2019234913A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP4030251A4 (en) * | 2020-08-05 | 2023-06-14 | Chiyoda Corporation | FACILITY MANAGEMENT METHOD, FACILITY DESIGN DEVICE AND FACILITY MANAGEMENT DEVICE |
JP7524569B2 (ja) | 2020-03-23 | 2024-07-30 | 横河電機株式会社 | アラーム発生システム及びアラーム発生方法 |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FI20195892A1 (en) * | 2019-10-16 | 2021-04-17 | Kemira Oyj | Diagnostic arrangement |
KR102594827B1 (ko) * | 2021-01-20 | 2023-10-26 | 한국수력원자력 주식회사 | 복수의 장치에 대한 작동예측시스템 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06347379A (ja) * | 1993-06-04 | 1994-12-22 | Hitachi Ltd | プラント監視診断システム及び非破壊検査診断システム |
JP2013145548A (ja) * | 2011-12-13 | 2013-07-25 | Yokogawa Electric Corp | アラーム表示装置およびアラーム表示方法 |
JP2016128973A (ja) * | 2015-01-09 | 2016-07-14 | 株式会社日立パワーソリューションズ | 予兆診断システム及び予兆診断方法 |
Family Cites Families (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS5864503A (ja) * | 1981-10-14 | 1983-04-16 | Hitachi Ltd | 装置の異常状態に対する原因推定方法 |
US5581459A (en) * | 1990-09-26 | 1996-12-03 | Hitachi, Ltd. | Plant operation support system |
US5764509A (en) * | 1996-06-19 | 1998-06-09 | The University Of Chicago | Industrial process surveillance system |
JP2007536634A (ja) * | 2004-05-04 | 2007-12-13 | フィッシャー−ローズマウント・システムズ・インコーポレーテッド | プロセス制御システムのためのサービス指向型アーキテクチャ |
US7248979B2 (en) * | 2005-05-09 | 2007-07-24 | International Business Machines Corporation | Apparatus employing predictive failure analysis based on in-circuit FET on-resistance characteristics |
US8396582B2 (en) * | 2008-03-08 | 2013-03-12 | Tokyo Electron Limited | Method and apparatus for self-learning and self-improving a semiconductor manufacturing tool |
TWI385492B (zh) * | 2008-12-16 | 2013-02-11 | Ind Tech Res Inst | 機台設備的維護分析系統及其方法 |
CN104583881B (zh) * | 2012-08-29 | 2017-05-24 | 丰田自动车株式会社 | 设备控制装置 |
US9465387B2 (en) * | 2015-01-09 | 2016-10-11 | Hitachi Power Solutions Co., Ltd. | Anomaly diagnosis system and anomaly diagnosis method |
WO2016151744A1 (ja) * | 2015-03-24 | 2016-09-29 | 三菱電機株式会社 | プラント監視制御装置 |
US10683844B2 (en) * | 2015-05-27 | 2020-06-16 | Vestas Wind Systems A/S | Control of a wind turbine taking fatigue measure into account |
JP6332154B2 (ja) * | 2015-06-11 | 2018-05-30 | 横河電機株式会社 | プラント運転支援装置、プラント運転支援方法、及びプログラム |
JP6682411B2 (ja) * | 2016-09-16 | 2020-04-15 | 横河電機株式会社 | プラント状態表示装置、プラント状態表示システム、及びプラント状態表示方法 |
JP6723669B2 (ja) * | 2016-09-27 | 2020-07-15 | 東京エレクトロン株式会社 | 異常検知プログラム、異常検知方法および異常検知装置 |
WO2018096682A1 (ja) * | 2016-11-28 | 2018-05-31 | 東芝三菱電機産業システム株式会社 | 圧延設備の異常診断の方法及び装置 |
JP6933899B2 (ja) * | 2017-01-12 | 2021-09-08 | 横河電機株式会社 | プラント運転支援装置、プラント運転支援方法、及びプラント運転支援プログラム |
-
2018
- 2018-06-08 WO PCT/JP2018/022018 patent/WO2019234913A1/ja active Application Filing
- 2018-06-08 AU AU2018426458A patent/AU2018426458B2/en active Active
- 2018-06-08 JP JP2018558257A patent/JP6529690B1/ja active Active
-
2019
- 2019-03-25 TW TW108110188A patent/TWI710873B/zh active
-
2020
- 2020-12-07 US US17/113,400 patent/US20210088986A1/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06347379A (ja) * | 1993-06-04 | 1994-12-22 | Hitachi Ltd | プラント監視診断システム及び非破壊検査診断システム |
JP2013145548A (ja) * | 2011-12-13 | 2013-07-25 | Yokogawa Electric Corp | アラーム表示装置およびアラーム表示方法 |
JP2016128973A (ja) * | 2015-01-09 | 2016-07-14 | 株式会社日立パワーソリューションズ | 予兆診断システム及び予兆診断方法 |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP7524569B2 (ja) | 2020-03-23 | 2024-07-30 | 横河電機株式会社 | アラーム発生システム及びアラーム発生方法 |
EP4030251A4 (en) * | 2020-08-05 | 2023-06-14 | Chiyoda Corporation | FACILITY MANAGEMENT METHOD, FACILITY DESIGN DEVICE AND FACILITY MANAGEMENT DEVICE |
Also Published As
Publication number | Publication date |
---|---|
AU2018426458B2 (en) | 2023-12-21 |
TW202013110A (zh) | 2020-04-01 |
TWI710873B (zh) | 2020-11-21 |
US20210088986A1 (en) | 2021-03-25 |
JPWO2019234913A1 (ja) | 2020-06-18 |
AU2018426458A1 (en) | 2020-11-19 |
JP6529690B1 (ja) | 2019-06-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10809704B2 (en) | Process performance issues and alarm notification using data analytics | |
JP7504163B2 (ja) | 異常予知装置、異常予知システム、異常予知方法及び異常予知プログラム | |
JP6933899B2 (ja) | プラント運転支援装置、プラント運転支援方法、及びプラント運転支援プログラム | |
JP2018160093A (ja) | 異常検知システムおよび異常検知方法 | |
US20210088986A1 (en) | Assistance device, learning device, and plant operation condition setting assistance system | |
JP5778087B2 (ja) | プロセス監視システム及び方法 | |
JP4046309B2 (ja) | プラント監視装置 | |
EP2816431A2 (en) | Information platform for industrial automation stream-based data processing | |
JP6711323B2 (ja) | プロセスの異常状態診断方法および異常状態診断装置 | |
JP2014167706A (ja) | 監視装置および制御システム | |
JP6702297B2 (ja) | プロセスの異常状態診断方法および異常状態診断装置 | |
WO2021127646A1 (en) | Device and method for monitoring a system | |
CN109564423B (zh) | 监测和控制工业过程的方法以及过程控制*** | |
JP2022182620A (ja) | 情報処理装置、予測方法および予測プログラム | |
JP6662222B2 (ja) | 製造プロセスの異常状態診断方法及び異常状態診断装置 | |
JP6579163B2 (ja) | プロセスの状態診断方法及び状態診断装置 | |
CN115409059A (zh) | 高压电子开关设备在线故障预警方法、终端设备 | |
EP3674828B1 (en) | System and method of generating data for monitoring of a cyber-physical system for early determination of anomalies | |
CN116997867A (zh) | 用于预测技术设施的运行的方法和*** | |
CN116235121A (zh) | 识别用于执行生产过程的工业设施中的异常的装置和方法 | |
CN115427907A (zh) | 用于工业过程中的智能报警管理方法 | |
JPH04366742A (ja) | プラント監視装置および方法 | |
JP7416015B2 (ja) | 情報処理装置、アラーム予測方法およびアラーム予測プログラム | |
JP7309548B2 (ja) | 異常予兆検知装置、方法及びプログラム | |
RU2773864C1 (ru) | Устройство поддержки, устройство обучения и система поддержки настройки условий эксплуатации установки |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2018558257 Country of ref document: JP Kind code of ref document: A |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 18921650 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2018426458 Country of ref document: AU Date of ref document: 20180608 Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 18921650 Country of ref document: EP Kind code of ref document: A1 |