Device, method and program for searching for return operation
The apparatus and method address the challenge of automatic recovery from equipment abnormalities by using machine learning to detect issues, set targets based on normal data, and adaptively search for recovery operations, enhancing efficiency and reducing the need for exhaustive simulations.
Patent Information
- Application Number
- JP2023087188
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-26
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2043-05-26
AI Technical Summary
Existing technologies struggle to automatically identify and recover from equipment abnormalities in complex systems like plants without pre-defining all possible abnormal conditions, especially in environments where various conditions interact intricately.
An apparatus and method that acquires state data, detects abnormalities, determines target data based on normal conditions, evaluates trial operations, and searches for recovery operations through trial and error using machine learning and evaluation models to restore equipment to normalcy.
Enables efficient and adaptive recovery from various abnormal states in equipment by objectively setting targets based on normal data, evaluating trial operations, and sequentially changing operation parameters, reducing the need for exhaustive simulations and human intervention.
Smart Images

Figure 2024170171000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to an apparatus, a method, and a program for searching for a return operation. [Background technology]
[0002] Patent Document 1 states that "a plant operation device is obtained that can identify the cause of an abnormality in a plant and can automatically perform recovery operations from an abnormal state without the intervention of an operator." [Prior art document] [Patent documents] [Patent Document 1] JP 2014-229109 A Summary of the Invention
[0003] A first aspect of the present invention provides an apparatus for searching for a return operation, the apparatus including: a state data acquisition unit that acquires state data indicating a state of equipment, an evaluation unit that evaluates a trial operation based on a change in the state data when a control target is controlled according to a trial operation that is attempted to return the equipment to normal from an abnormality, and a search unit that searches for a return operation for returning the equipment to normal from an abnormality based on a result of the evaluation.
[0004] The apparatus may further include a target data determination unit that determines target data indicating a target state of the equipment for returning the equipment from an abnormality to normal, and the evaluation unit may evaluate the trial operation based on a change in distance between the state data and the target data.
[0005] In any of the above devices, the target data determination unit may determine the target data based on a plurality of normal data indicating the status data under normal circumstances.
[0006] In any of the above devices, the target data determination section may determine the target data based on statistics of the plurality of normal data.
[0007] In any of the above devices, the target data determination unit may determine the target data based on a time series of the plurality of normal data.
[0008] Any of the devices may further include an index acquisition unit that acquires an evaluation index that evaluates the state of the equipment in response to inputting the state data into an evaluation model, and the evaluation unit may evaluate the trial operation based on a change in the evaluation index.
[0009] Any of the devices may further comprise the evaluation model.
[0010] In any of the devices, the search section may sequentially change one operation parameter out of a plurality of operation parameters included in the trial operation when searching for the return operation.
[0011] In any of the above devices, the search section may omit changing the one operation parameter to some of a plurality of candidate operation amounts.
[0012] Any of the devices may further include a detection unit that detects an abnormality in the equipment based on the status data.
[0013] Any of the devices may further include a warning unit that issues a warning when the equipment does not return to normal from the abnormality within a predetermined condition even after the searched recovery operation.
[0014] In any of the devices, the warning unit may issue a warning when an elapsed time until recovery to normality in response to detection of an abnormality in the equipment exceeds a predetermined threshold.
[0015] In any of the devices, the warning unit may issue a warning when the number of times the control target is controlled in accordance with the trial operation exceeds a predetermined threshold.
[0016] Any of the devices may further include a control unit that controls the control target in accordance with the trial operation.
[0017] In a second aspect of the present invention, there is provided a method for searching for a return operation, the method including the steps of: acquiring status data indicating a status of equipment; evaluating a trial operation based on a change in the status data when a control target is controlled according to the trial operation to return the equipment to normal from an abnormality; and searching for a return operation to return the equipment to normal from an abnormality based on a result of the evaluation.
[0018] In a third aspect of the present invention, there is provided a program for searching for a return operation, which is executed by a computer and causes the computer to function as a status data acquisition unit that acquires status data indicating a status of equipment, an evaluation unit that evaluates a trial operation based on a change in the status data when a control target is controlled in accordance with a trial operation attempted to return the equipment to normal from an abnormality, and a search unit that searches for a return operation to return the equipment to normal from an abnormality based on a result of the evaluation.
[0019] The above summary of the invention does not list all of the features of the present invention. Also, subcombinations of these features may also be inventions. [Brief description of the drawings]
[0020] [Figure 1] 1 shows an example of a block diagram of a control system 1 in which an apparatus 1000 according to the present embodiment may be included. [Diagram 2] 1 shows a first example of determining target data. [Diagram 3] A second example of determining target data is shown. [Figure 4] 1 shows an example of a flow diagram of a method that may be performed by the device 1000 according to the present embodiment. [Diagram 5]1 shows another example of a block diagram of a control system 1 in which the device 1000 according to the present embodiment may be included. [Figure 6] FIG. 1 shows an example of a block diagram of a control system 1 that may include an apparatus 1000 according to a modified embodiment of the present invention. [Figure 7] An example of a flow diagram of a method that may be executed by the device 1000 according to a variation of this embodiment is shown. [Figure 8] FIG. 13 shows another example of a block diagram of a control system 1 that may include a device 1000 according to a modified example of the present embodiment. [Figure 9] An example of a computer 9900 in which aspects of the present invention may be embodied, in whole or in part, is shown. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0021] The present invention will be described below through embodiments of the invention, but the following embodiments do not limit the scope of the invention according to the claims. Furthermore, not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention.
[0022] FIG. 1 shows an example of a block diagram of a control system 1 that may include a device 1000 according to this embodiment. These blocks are functionally separated functional blocks and may not necessarily match the actual device configuration. That is, even if one block is shown in this diagram, it does not necessarily have to be composed of one device. Also, even if separate blocks are shown in this diagram, they do not necessarily have to be composed of separate devices. The same applies to the subsequent block diagrams. The control system 1 may include a facility 10, a control device 20, and a device 1000.
[0023] The facility 10 is a device that manufactures a product from raw materials. The facility 10 may be a single device, or may be a composite device that combines multiple devices. The facility 10 may be, for example, an entire plant or factory, or a group of devices in a segment of a plant or factory. Examples of plants include industrial plants such as chemical and bio plants, as well as plants that manage and control wellheads and surrounding areas of gas fields and oil fields, plants that manage and control power generation such as hydroelectric, thermal, and nuclear power, plants that manage and control environmental power generation such as solar and wind power, and plants that manage and control water supply and sewage systems, dams, etc.
[0024] The equipment 10 may be provided with one or more sensors (not shown) capable of measuring various conditions (physical quantities) inside and outside the equipment 10. The sensors may output measured values PV (Process Variables) measuring temperature, pressure, flow rate, etc. at various points of the equipment 10. Status data indicating the status of the equipment 10 may include such measured values PV. The status data may further include manipulated variables MV (Manipulated Variables) indicating the opening and closing degree of a valve, etc. The status data may further include consumption data indicating the amount of energy consumed in the equipment 10 or the amount of raw material consumed, etc. The status data may further include environmental data indicating physical quantities that may act as disturbances in the equipment 10.
[0025] The facility 10 may be provided with a control target 15. In this figure, for convenience of explanation, only one control target 15 is shown, but the facility 10 may be provided with one or a plurality of control targets 15.
[0026] The controlled object 15 is a device to be controlled. The controlled object 15 may be an actuator, such as a valve, heater, motor, fan, or switch, that controls at least one physical quantity, such as the amount, temperature, pressure, flow rate, speed, or pH, of an object related to the process of the facility 10, that is, an operating end, and may execute a given operation according to the operating amount. However, the controlled object 15 is not limited to this. The controlled object 15 may also be a controller that controls the operating end. In other words, the term "control" used in this specification may be broadly interpreted to include indirect control of the operating end via a controller in addition to direct control of the operating end.
[0027] The control device 20 is a device that controls the control target 15. The control device 20 may acquire status data indicating the status of the equipment 10, and control the control target 15 based on the status data. In this case, the control device 20 may control the control target 15 by PID (Proportional Integral Differential) control, or may control the control target 15 by AI (Artificial Intelligence) control.
[0028] The AI control is a control using an operation model generated by machine learning. Such an operation model may be generated by reinforcement learning in which an evaluation index that evaluates the state of the equipment 10 is used as at least a part of the reward. The evaluation index used in such reinforcement learning may be output from an evaluation model that has been machine-learned to output an evaluation index in response to input of state data.
[0029] An abnormality may occur in the equipment 10 under the control of the controlled object 15 by the control device 20. In such a case, it is desirable to automatically restore the equipment 10 by controlling the controlled object 15 according to a restoration operation. However, the equipment 10 may be in various abnormal states depending on various conditions. In order to deal with such cases, it is conceivable to define a restoration operation for each abnormal state in a database or the like. However, it is practically impossible to consider all abnormal states in advance, and such a problem becomes particularly noticeable in complex equipment such as a plant where various conditions affect each other.
[0030] Therefore, the device 1000 according to the present embodiment searches for a return operation for returning the equipment 10 to normal from an abnormality through trial and error using an actual machine. As a result, the device 1000 according to the present embodiment can search for a return operation through trial and error even when the equipment 10 is in various abnormal states. This will be described in detail. The device 1000 may include a state data acquisition unit 110, a detection unit 120, a target data determination unit 130, an evaluation unit 170, a search unit 180, and a warning unit 190.
[0031] The status data acquisition unit 110 acquires status data indicating the status of the equipment 10. The status data acquisition unit 110 may be provided mainly by a communication interface, and may acquire the status data from the equipment 10 in time series via a network. However, this is not limited to this. The status data acquisition unit 110 may acquire the status data from a device other than the equipment 10, or may acquire the status data via other means other than a network (such as various memory devices or user input). The status data acquisition unit 110 supplies the acquired status data to the detection unit 120.
[0032] The detection unit 120 detects an abnormality in the equipment 10 based on the status data. The detection unit 120 may be provided mainly by executing a program by a CPU or other processor, and may detect an abnormality in the equipment 10 based on the status data acquired by the status data acquisition unit 110. In this case, the detection unit 120 may use a machine learning model that classifies the equipment 10 into a normal state and an abnormal state based on the status data. When it is determined that an abnormality in the equipment 10 has not been detected, the detection unit 120 attaches a normal label to the status data and supplies it to the target data determination unit 130. When it is determined that an abnormality in the equipment 10 has been detected, the detection unit 120 attaches an abnormal label to the status data and supplies it to the target data determination unit 130 and the evaluation unit 170. In addition, the detection unit 120 notifies the search unit 180 and the warning unit 190 that an abnormality in the equipment 10 has been detected.
[0033] The target data determination unit 130 determines target data indicating a target state of the equipment 10 for returning the equipment 10 from an abnormality to normal. The target data determination unit 130 may be provided mainly by a CPU, and may determine the target data based on a plurality of state data labeled as normal by the detection unit 120, i.e., a plurality of normal data indicating state data in a normal state. This will be described in detail later. The target data determination unit 130 supplies the determined target data to the evaluation unit 170.
[0034] The evaluation unit 170 evaluates the trial operation based on a change in status data when the control target 15 is controlled according to a trial operation attempted to restore the equipment 10 to normal from an abnormality. The evaluation unit 170 may be provided mainly by a CPU, and may evaluate the trial operation based on a change in distance between the status data labeled as abnormal by the detection unit 120 and the target data determined by the target data determination unit 130. The evaluation unit 170 supplies the evaluated result to the search unit 180.
[0035] The search unit 180 searches for a return operation for returning the equipment 10 to normal from the abnormality based on the evaluation result. The search unit 180 may be mainly provided by a communication interface and a CPU, and may supply the trial operation to the control device 20 via a network. Then, the control device 20 may control the control target 15 according to the trial operation. This causes the state of the equipment 10 to change. The state data acquisition unit 110 may acquire state data again after the control according to the trial operation. Also, the evaluation unit 170 may evaluate the trial operation based on the change in the state data. Then, the search unit 180 may sequentially change one of the multiple operation parameters included in the trial operation based on the evaluation result. By repeatedly executing such processing, the search unit 180 can search for a return operation by trial and error based on the result evaluated by the evaluation unit 170.
[0036] The warning unit 190 issues a warning when the equipment 10 does not return to normal from the abnormality within the predetermined conditions even by the searched return operation. The warning unit 190 may be provided mainly by a display device, and may display a warning screen indicating that the equipment 10 does not return to normal from the abnormality within the predetermined conditions even by the searched return operation.
[0037] The device 1000 having such a functional unit may be a computer such as a PC (personal computer), a tablet computer, a smartphone, a workstation, a server computer, or a general-purpose computer, or may be a computer system to which a plurality of computers are connected. Such a computer system is also a computer in a broad sense. The device 1000 may be implemented by a virtual computer environment in which one or more programs can be executed in a computer. Alternatively, the device 1000 may be a dedicated computer designed to search for a return operation, or may be dedicated hardware realized by a dedicated circuit. In addition, when the device 1000 can be connected to the Internet, the device 1000 may be realized by cloud computing.
[0038] Such a computer may include a memory for storing a program and a processor for executing the program, and the processor may execute the program to implement the functions of the device 1000. That is, a program may be provided that is executed by a computer and causes the computer to function as a status data acquisition unit 110 that acquires status data indicating the status of the equipment 10, an evaluation unit 170 that evaluates a trial operation based on a change in the status data when the control target 15 is controlled according to a trial operation that is attempted to return the equipment 10 to normal from the abnormality, and a search unit 180 that searches for a return operation for returning the equipment 10 to normal from the abnormality based on the evaluation result.
[0039] FIG. 2 shows a first example of determining target data. This figure shows a feature space defined by state data. In this figure, a circle (◯) indicates normal data, i.e., state data labeled as normal by the detection unit 120, and a cross (×) indicates abnormal data, i.e., state data labeled as abnormal by the detection unit 120. Reference numeral 210 is a visualization of the decision boundary. The machine learning model that may be used by the detection unit 120 may be tuned by machine learning in this way to determine the boundary for classifying the equipment 10 as normal or abnormal.
[0040] Reference numeral 220 indicates the target abnormal data. When the target abnormal data is supplied from the detection unit 120, the target data determination unit 130 may determine target data indicating a target state of the equipment 10 for returning the equipment 10 from the abnormality to normal. At this time, the target data determination unit 130 may determine the target data based on a plurality of normal data. The target data determination unit 130 can determine such target data using various methods.
[0041] This figure shows an example of a method for determining target data based on statistics of multiple normal data. Reference numeral 230 indicates the center of gravity of multiple normal data. The target data determination unit 130 may calculate the sum of squares of distances from multiple normal data, and define the point at which the sum of squares of distances is minimum as the center of gravity 230. In this case, the target data determination unit 130 may select multiple normal data to be used in the statistical processing according to a predetermined rule. For example, when the target abnormal data is supplied from the detection unit 120, the target data determination unit 130 may select m normal data supplied from the detection unit 120 in order of newest, and use the selected m normal data for the statistical processing.
[0042] The target data determination unit 130 may calculate a vector 240 that points from the target abnormal data 220 to the center of gravity 230 of the multiple normal data. Then, the target data determination unit 130 may determine the intersection 250 of the vector 240 and the decision boundary 210 as the target data. In the above description, the intersection 250 of the vector 240 and the decision boundary 210 is set as the target data as an example, but the target data determination unit 130 may determine the center of gravity 230 itself as the target data, or may determine any point on the vector 240 between the intersection 250 and the center of gravity 230 as the target data. In the above description, the center of gravity is used as a statistical quantity as an example, but the target data determination unit 130 may determine the target data based on other statistical quantities such as the average value, the median value, or the mode value. For example, in this manner, the target data determination unit 130 can determine the target data based on the statistical quantities of multiple normal data.
[0043] Figure 3 shows a second example of determining target data. In this figure, the same reference numerals are used for components having the same functions and configurations as those in Figure 2, and explanations will be omitted except for the differences. In this figure, the numbers in the circles (○) indicate the chronological order of the normal data, with smaller values indicating newer data and larger values indicating older data.
[0044] This figure shows an example of a method for determining target data based on a time series of multiple normal data. Reference numeral 310 indicates the latest normal data among the selected m normal data (m=10 in this figure). The target data determination unit 130 may calculate a vector 320 that points from the target abnormal data 220 to the latest normal data 310. Then, the target data determination unit 130 may determine an intersection 330 between the vector 320 and the decision boundary 210 as the target data. In the above description, the intersection 330 between the vector 320 and the decision boundary 210 is set as the target data, but the target data determination unit 130 may determine the latest normal data 310 itself as the target data, or may determine an arbitrary point between the intersection 330 on the vector 320 and the latest normal data 310 as the target data. The target data determination unit 130 may determine target data based on a time series of multiple normal data in this way, for example.
[0045] Although the method of determining target data based on statistics in a plurality of normal data and the method of determining target data based on a time series in a plurality of normal data have been described as separate methods, the two may be combined. For example, the target data determination unit 130 may calculate a weighted average for a plurality of normal data by setting weights according to the order of the time series so that newer data has a larger contribution and older data has a smaller contribution. Then, the target data determination unit 130 may determine target data based on the weighted average. For example, the target data determination unit 130 may determine target data based on statistics and time series in a plurality of normal data in this way.
[0046] 4 shows an example of a flow diagram of a method that may be executed by the device 1000 according to the present embodiment. Each step in the diagram may be executed by the device 1000, i.e., a computer. However, as long as the computer is the overall executing entity in each step, it may also include cases where a part that is not the main part is executed by something other than a computer.
[0047] In step S1400, the computer executes initialization. For example, the computer may initialize a counter that counts the number of times k that the control target 15 has been controlled according to the trial operation. That is, the computer may set k=0. The computer may also initialize a timer that measures an elapsed time t in response to detection of an abnormality in the equipment 10. That is, the computer may set t=0.
[0048] In step S1410, the computer acquires status data indicating the status of the equipment 10. For example, the status data acquisition unit 110, which may be implemented in the computer, may acquire status data S indicating the status of the equipment 10 from the equipment 10 in time series via a network. As an example, when k=0, the status data acquisition unit 110 may acquire status data S0. The status data acquisition unit 110 supplies the acquired status data S to the detection unit 120.
[0049] In step S1420, the computer determines whether or not an abnormality in the equipment 10 has been detected based on the status data. For example, the detection unit 120, which may be implemented in the computer, may determine whether or not an abnormality in the equipment 10 has been detected based on the status data S acquired in step S1410. As an example, when k=0, the detection unit 120 may determine whether or not an abnormality in the equipment 10 has been detected based on the status data S0. At this time, the detection unit 120 may use a machine learning model that classifies the equipment 10 into a normal state and an abnormal state based on the status data S, as described above.
[0050] When the machine learning model classifies the equipment 10 as being in a normal state in response to the input of the status data S, the detection unit 120 may determine that an abnormality in the equipment 10 has not been detected (No). In this case, the detection unit 120 assigns a normal label n to the status data S and supplies it to the target data determination unit 130. Then, the computer returns the process to step S1400 and continues the flow.
[0051] On the other hand, when the machine learning model classifies the equipment 10 into an abnormal state in response to the input of the status data S, the detection unit 120 may determine that an abnormality in the equipment 10 has been detected (Yes). In this case, the detection unit 120 attaches an abnormality label a to the status data S and supplies it to the target data determination unit 130 and the evaluation unit 170. In addition, the detection unit 120 notifies the search unit 180 and the warning unit 190 that an abnormality in the equipment 10 has been detected. Then, the computer advances the process to step S1430.
[0052] In step S1430, the computer determines whether or not the predetermined condition is satisfied. For example, the warning unit 190 that may be implemented in the computer may determine whether or not the number of times k is within a predetermined threshold K (where K is an integer of 1 or 2 or more). The warning unit 190 may also determine whether or not the elapsed time t is within a predetermined threshold T. However, this is not limited to this. Various conditions for canceling, midway through, or terminating the search for the return operation may be predefined. If it is determined that the predetermined condition is satisfied (Yes), the computer advances the process to step S1440.
[0053] In step S1440, the computer determines whether or not the return process is being executed. For example, the computer may determine whether or not the number of times k is 0. If k=0, the computer may determine that the return process is not being executed (No). If it is determined that the return process is not being executed, the computer proceeds to step S1445.
[0054] In step S1445, the computer starts a timer that measures the elapsed time t in response to the detection of the abnormality of the equipment 10. Here, the computer may start the timer from a specific start point that is determined in advance in response to the detection of the abnormality of the equipment 10. For example, the computer may measure the elapsed time t from the detection of the abnormality of the equipment 10 by starting the timer at the timing when the abnormality of the equipment 10 is detected (after YES in S1420). Alternatively, the computer may measure the elapsed time t from the determination or execution of the first trial operation after the abnormality of the equipment 10 is detected (S480). The computer may use another start point that can measure the time required for the equipment 10 to recover from the abnormality. Then, the computer proceeds to step S1450.
[0055] In step S1450, the computer determines target data indicating a target state of the equipment 10 for returning the equipment 10 from an abnormality to normal. For example, the target data determination unit 130 that may be implemented in a computer may determine the target data St based on the state data S to which the normal label n has been attached in step S1420, i.e., the normal data Sn indicating the state data S in a normal state.
[0056] At this time, the target data determination unit 130 may determine the target data St based on a plurality of normal data Sn1, Sn2, ..., Snm indicating the status data S in a normal state. Here, the normal data Sn1, Sn2, ..., Snm are m pieces of normal data Sn selected in order of newness before the time point when the abnormality of the equipment 10 was detected. m indicates the order of the normal data Sn in time series, with a smaller value indicating newer data and a larger value indicating older data.
[0057] At this time, the target data determination unit 130 may determine the target data St based on the statistics of the normal data Sn1, Sn2, ..., Snm, for example, using the method of Fig. 2. The target data determination unit 130 may also determine the target data St based on the time series of the normal data Sn1, Sn2, ..., Snm, for example, using the method of Fig. 3. The target data determination unit 130 may also determine the target data St based on the statistics and time series of the normal data Sn1, Sn2, ..., Snm, for example, by combining the method of Fig. 2 and the method of Fig. 3. The target data determination unit 130 supplies the determined target data St to the evaluation unit 170.
[0058] In step S1455, the computer supplies the control device 20 with a trial operation A0 to be tried in order to restore the equipment 10 to normal from the abnormality. For example, the search unit 180, which may be implemented in a computer, may supply the trial operation A0 to the control device 20 via a network. Note that, in step S1455, since this is the stage at which the trial operation A is supplied for the first time after the abnormality of the equipment 10 is detected and before an evaluation of the trial operation A is obtained, the trial operation A0 may be set to an initial value.
[0059] Here, the trial operation A may include a plurality of operation parameters. Each of the plurality of operation parameters may be selected from a plurality of candidate operation amounts. Here, as an example, the trial operation A is assumed to be composed of a set of a plurality of operation parameters [Aa, Ab, Ac, Ad, ...]. The operation parameter Aa is a parameter for controlling the lighting of the light, and is assumed to be selected from the candidate operation amounts {OFF (off), ON (on)}. The operation parameter Ab is a parameter for controlling the supply of the additive, and is assumed to be selected from the candidate operation amounts {OFF (shut off), ON (supply)}. The operation parameter Ac is a parameter for controlling the opening of the valve, and is assumed to be selected from the candidate operation amounts {-5 (close 5 degrees), 0 (no change), +5 (open 5 degrees)}. The operation parameter Ad is a parameter for controlling the temperature of the heater, and is assumed to be selected from the candidate operation amounts {-3 (lower 3 degrees), 0 (no change), +3 (raise 3 degrees)}.
[0060] In such a case, the trial operation A0 may be set to, for example, [OFF, OFF, 0, 0, ...]. The search unit 180 may, for example, supply such a trial operation A0 to the control device 20. In response to this, the control device 20 may control the control target 15 according to the trial operation A0. Then, the computer proceeds to step S1460.
[0061] In step S1460, the computer executes an increment process. For example, the computer may set k=k+1. As an example, when k=0, the computer may set k=1. Then, the computer returns the process to step S1410 to continue the flow.
[0062] That is, when k=1, the status data acquisition unit 110 acquires status data S1 indicating the state of the equipment 10 after the control object 15 is controlled according to the trial operation A0, and the detection unit 120 may determine whether or not an abnormality in the equipment 10 has been detected based on the status data S1.
[0063] Here, if k=1, in step S1440, the computer may determine that the return process is being executed (Yes) since k≠0. If it is determined that the return process is being executed, the computer proceeds to step S1470.
[0064] In step S1470, the computer evaluates the trial operation based on a change in the status data when the control target 15 is controlled according to the trial operation attempted to restore the equipment 10 from the abnormality to normal. For example, the evaluation unit 170 that may be implemented in a computer may evaluate the trial operation based on a change in the distance between the status data labeled as abnormal in step S1420 and the target data determined in step S1450.
[0065] As an example, when k=1, the evaluation unit 170 may calculate the distance D0 between the state data S0 and the target data St. In this case, the evaluation unit 170 may calculate at least one of the Euclidean distance, the Manhattan distance, the Chebyshev distance, and the Mahalanobis distance. Similarly, the evaluation unit 170 may calculate the distance D1 between the state data S1 and the target data St. Then, the evaluation unit 170 may calculate the difference d01 by subtracting the distance D0 from the distance D1.
[0066] When the difference d01 is greater than 0, that is, when the distance D1 is smaller than the distance D0, it means that the state of the equipment 10 has come closer to the target state by the trial operation A0. In this case, the evaluation unit 170 may give the trial operation A0 an evaluation value greater than 0. In this case, the evaluation unit 170 may give an evaluation value that is greater as the absolute value of the difference d01 is greater.
[0067] If the difference d01 is 0, that is, if the distance D0 and the distance D1 are the same, it means that the trial operation A0 has not changed the state of the equipment 10. In this case, the evaluation unit 170 may give an evaluation value of 0 to the trial operation A0.
[0068] When the difference d01 is smaller than 0, that is, when the distance D1 is greater than the distance D0, it means that the state of the equipment 10 has moved away from the target state due to the trial operation A0. In this case, the evaluation unit 170 may give the trial operation A0 an evaluation value smaller than 0. In this case, the evaluation unit 170 may give an evaluation value that is smaller as the absolute value of the difference d01 is larger.
[0069] In this manner, for example, the evaluation section 170 may evaluate the trial operation A based on the change in the distance D between the state data S and the target data St. The evaluation section 170 supplies the evaluated result to the search section 180.
[0070] In step S1475, the computer searches for a return operation for returning the equipment 10 from the abnormality to normal based on the evaluation result. For example, the search unit 180, which may be implemented in the computer, may search for a return operation based on the evaluation result in step S1470. In this case, the search unit 180 may sequentially change one operation parameter among a plurality of operation parameters included in the trial operation when searching for a return operation. As an example, assume that the evaluation value of the trial operation A0 is 0 when k=1. In this case, the search unit 180 may randomly select one operation parameter among the plurality of operation parameters and change the one operation parameter to another candidate operation amount. As an example, the search unit 180 may change the operation parameter Aa of the trial operation A0 [OFF, OFF, 0, 0, ...] from "OFF" to "ON" and set the trial operation A1 to [ON, OFF, 0, 0, ...].
[0071] In step S1480, the computer supplies the control device 20 with a trial operation A to be tried in order to restore the equipment 10 to normal from the abnormality. For example, the search unit 180, which may be implemented in the computer, may supply the trial operation A changed in step S1475 to the control device 20 via a network. As an example, when k=1, the search unit 180 may supply the trial operation A1[ON, OFF, 0, 0, ...] to the control device 20. In response to this, the control device 20 may control the control target 15 according to the trial operation A1. Then, the computer advances the process to step S1460, and repeatedly executes the restoration process.
[0072] That is, when k=2, the state data acquiring unit 110 may acquire state data S2 indicating the state of the equipment 10 after the control target 15 is controlled according to the trial operation A1, and the evaluating unit 170 may calculate a distance D2 between the state data S2 and the target data St. Then, the evaluating unit 170 may calculate a difference d12 by subtracting the distance D1 from the distance D2, and give an evaluation value to the trial operation A1 according to the difference d12.
[0073] Here, suppose that the evaluation value of the trial operation A1 is 0 when the control target 15 is controlled according to the trial operation A1. The search unit 180, which has obtained such an evaluation result, can recognize that the change in the operation parameter Aa does not contribute to the return operation (i.e., the operation parameter Aa is an insensitive parameter). In this case, the search unit 180 may determine the operation parameter Aa to be an arbitrary candidate operation amount (e.g., "OFF"). Then, the search unit 180 may select one undetermined operation parameter from among the multiple operation parameters, and change it to another candidate operation amount. As an example, the search unit 180 may change the operation parameter Ab from "OFF" to "ON" and set the trial operation A2 to [OFF, ON, 0, 0, ...].
[0074] Next, suppose that the evaluation value of the trial operation A2 is greater than 0 when the control target 15 is controlled according to the trial operation A2. The search unit 180, which has obtained such an evaluation result, can recognize that the state of the equipment 10 has improved by changing the operation parameter Ab. In this case, the search unit 180 may determine the operation parameter Ab as a candidate operation amount after the change (here, "ON"). Then, the search unit 180 may select one of the undetermined operation parameters among the multiple operation parameters, and change it to another candidate operation amount. As an example, the search unit 180 may change the operation parameter Ac from "0" to "+5" and set the trial operation A3 to [OFF, ON, +5, 0, ...].
[0075] Next, suppose that when the control target 15 is controlled according to the trial operation A3, the evaluation value of the trial operation A3 is smaller than 0. The searching unit 180, which has obtained such an evaluation result, can recognize that the condition of the equipment 10 has deteriorated due to the change in the operation parameter Ac. In this case, the searching unit 180 may change the operation parameter Ac to another candidate operation amount (here, "-5") and set the trial operation A4 to [OFF, ON, -5, 0, ...].
[0076] Next, suppose that the evaluation value of the trial operation A4 is greater than 0 when the control target 15 is controlled according to the trial operation A4. The search unit 180, which has obtained such an evaluation result, can recognize that the state of the equipment 10 has improved by changing the operation parameter Ac. In this case, the search unit 180 may determine the operation parameter Ac to be a candidate operation amount after the change (here, "-5"). Then, the search unit 180 may select one of the operation parameters that has not been determined yet, and change it to another candidate operation amount. As an example, the search unit 180 may change the operation parameter Ad from "0" to "+3" and set the trial operation A5 to [OFF, ON, -5, +3, ...].
[0077] Next, suppose that the evaluation value of the trial operation A5 is greater than 0 when the control target 15 is controlled according to the trial operation A5. The search unit 180, which has obtained such an evaluation result, can recognize that the state of the equipment 10 has improved by changing the operation parameter Ad. Here, the operation parameter Ad can also select "-3" as another candidate operation amount. However, since the state of the equipment 10 has improved as a result of setting the operation parameter Ad to "+", if the operation parameter Ad is changed to "-3" which has a different polarity from "+3", it is expected that the state of the equipment 10 will worsen. In such a case, the search unit 180 may determine the operation parameter Ad as the candidate operation amount after the change (here, "+3") without trying a trial operation in which the operation parameter Ad is changed to "-3". For example, in this way, the search unit 180 can omit changing one operation parameter to some of the multiple candidate operation amounts.
[0078] The computer may search for a restoration operation by trial and error using the actual device in this manner until, in step S1420, the computer no longer detects an abnormality in the equipment 10, or until, in step S1430, a predetermined condition is exceeded.
[0079] During such a search for a return operation, if the number of times k exceeds the threshold value K in step S1430, or if the elapsed time t from when the timer is started in S1445 in response to the detection of an abnormality in the equipment 10 exceeds the threshold value T, the warning unit 190 may determine that the predetermined condition is not satisfied (No). In this case, the computer advances the process to step S1490.
[0080] In step S1490, the computer issues a warning when the equipment 10 does not return to normal from the abnormality within a predetermined condition. For example, the warning unit 190, which may be implemented in the computer, may display a warning screen indicating that the number of times k exceeds a threshold K or that the elapsed time t exceeds a threshold T. At this time, the warning unit 190 may display the execution history of the search on the screen. As an example, the warning unit 160 may display the trial operation A and the transition of the state data S on the screen using a graph or the like. The warning unit 190 may issue a warning, for example, in this way when the number of times (e.g., k) the control target 15 is controlled according to the trial operation exceeds a predetermined threshold (e.g., K). The warning unit 190 may also issue a warning when the elapsed time (e.g., t) until the equipment 10 returns to normal in response to the detection of an abnormality in the equipment 10 exceeds a predetermined threshold (e.g., T).
[0081] In this way, a method may be provided in which a computer acquires status data indicating the status of the equipment 10, evaluates the trial operation based on a change in the status data when the control object 15 is controlled according to the trial operation attempted to return the equipment 10 to normal from the abnormality, and searches for a return operation to return the equipment 10 to normal from the abnormality based on the evaluated result.
[0082] Conventionally, the content of the restoration operation for an abnormal state has been determined by referring to a restoration operation determination database in which the content of the restoration operation for an abnormal state is predefined. However, it is practically impossible to consider various abnormal states in advance, and such a problem is particularly prominent in complex facilities such as plants where various conditions affect each other.
[0083] In contrast, the device 1000 according to the present embodiment evaluates a trial operation based on a change in state data when the control target 15 is controlled according to the trial operation, and searches for a return operation based on the evaluation result. As a result, the device 1000 according to the present embodiment searches for a return operation for returning the equipment 10 to normal from an abnormality by trial and error using an actual machine, so that it is possible to try to return from various abnormal states without having to prepare in advance a database for determining a return operation or a simulator that simulates the equipment 10.
[0084] Furthermore, the device 1000 according to the present embodiment may determine target data and evaluate the trial operation based on the change in the distance between the state data and the target data. As a result, the device 1000 according to the present embodiment can give a higher evaluation to a trial operation that brings the state of the equipment 10 closer to a target state.
[0085] Furthermore, the device 1000 according to the present embodiment may determine target data based on multiple normal data sets. As a result, the device 1000 according to the present embodiment can objectively set a target to be aimed at when returning the equipment 10 from an abnormality to normality, based on multiple normal data sets, rather than setting the target at random.
[0086] In this case, the device 1000 according to the present embodiment may determine the target data based on statistics of a plurality of normal data, or may determine the target data based on a time series of a plurality of normal data. There are many different types of abnormalities, including static abnormalities with no sequential relationship between data, and dynamic abnormalities with sequential relationship between data. According to the device 1000 according to the present embodiment, even for such a wide variety of abnormalities, it is possible to appropriately set a target to be aimed at when returning the equipment 10 to normal from an abnormality.
[0087] Furthermore, when searching for a return operation, the device 1000 according to the present embodiment may sequentially change one of a plurality of operation parameters included in the trial operation. As a result, the device 1000 according to the present embodiment can search for a return operation while sequentially evaluating each operation parameter one by one, even when it is unclear which operation parameter contributes to the return operation.
[0088] At this time, the device 1000 according to the present embodiment may omit changing one operation parameter to some of the multiple candidate operation amounts. As a result, the device 1000 according to the present embodiment can efficiently search for a return operation without exhaustively trying all combinations of the candidate operation amounts.
[0089] Moreover, the device 1000 according to the present embodiment may further include a function of detecting an abnormality in the equipment 10 based on the status data. As a result, the device 1000 according to the present embodiment can realize an abnormality detection function and a recovery operation search function in a single device. Moreover, the device 1000 according to the present embodiment can autonomously detect an abnormality in the equipment 10 without human intervention, and can automatically sort normal data and abnormal data required for searching a recovery operation by assigning a normal label and an abnormal label to the status data.
[0090] Moreover, the device 1000 according to the present embodiment may further include a function of issuing a warning when the equipment 10 does not return to normal from an abnormality within a predetermined condition. As a result, according to the device 1000 according to the present embodiment, when the equipment 10 does not return to normal from an abnormality within the condition despite the search for a return operation, it is possible to inform the user of this fact. Therefore, according to the device 1000 according to the present embodiment, even when it is difficult to automatically return the equipment 10, it is possible to prompt the user to switch to manual return, etc., without unnecessarily repeating the search process for a return operation.
[0091] In this case, the device 1000 according to the present embodiment may issue a warning when the number of times the control target 15 has been controlled according to the trial operation exceeds a predetermined threshold, or may issue a warning when the time that elapses until the equipment 10 returns to normal in response to the detection of an abnormality exceeds a predetermined threshold. Since the time constants and settling times vary depending on the equipment 10, the speed at which the state of the equipment changes when the control target 15 is controlled according to the trial operation varies. According to the device 1000 according to the present embodiment, even in such a case, the timing to stop, middle, or end the search for a return operation can be set to an optimal timing for each equipment 10.
[0092] FIG. 5 shows another example of a block diagram of a control system 1 that may include the device 1000 according to this embodiment. In this diagram, the same reference numerals are used for components having the same functions and configurations as those in FIG. 1, and descriptions thereof will be omitted except for the differences. Up to this point, a case in which the control device 20 and the device 1000 are provided as separate, independent devices has been shown as an example. However, the control device 20 and the device 1000 may be provided as a single device in which a part or all of them are integrated. In this diagram, a case in which the device 1000 also provides the functions of the control device 20 is shown.
[0093] In this diagram, device 1000 further includes a control unit 510. In this diagram, state data acquisition unit 110 may supply the acquired state data to control unit 510 in addition to detection unit 120. Furthermore, search unit 180 may supply the trial operation to control unit 510 instead of control device 20.
[0094] The control unit 510 may be mainly provided by a CPU, and may be implemented in the device 1000 as a functional unit similar to the control device 20. That is, when no abnormality occurs in the equipment 10, the control unit 510 may control the control target 15 based on the state data acquired by the state data acquisition unit 110. At this time, the control unit 510 may control the control target 15 by PID control, or may control the control target 15 by AI control. On the other hand, when an abnormality occurs in the equipment 10, the control unit 510 may control the control target 15 according to the trial operation supplied from the search unit 180. In this case, the control unit 510 may switch the control mode depending on whether or not an abnormality in the equipment 10 is detected by the detection unit 120.
[0095] In this manner, the device 1000 according to the present embodiment may further include the control unit 510. As a result, the device 1000 according to the present embodiment can realize the function of searching for a return operation and the function of controlling the control target 15 by a single device. Therefore, the device 1000 according to the present embodiment does not need to supply a trial operation from the device 1000 to the control device 20, and can eliminate the risk that the trial operation is tampered with in the communication between the device 1000 and the control device 20, thereby protecting the facility 10 from security threats.
[0096] Fig. 6 shows an example of a block diagram of a control system 1 that may include an apparatus 1000 according to a modified example of this embodiment. In this figure, the same reference numerals are used for components having the same functions and configurations as those in Fig. 1, and descriptions thereof will be omitted hereinafter except for differences. In the above description, a case in which the apparatus 1000 evaluates a trial operation using target data is shown as an example. However, in this modified example, the apparatus 1000 evaluates a trial operation using an evaluation index.
[0097] In this modification, the control system 1 may further include an evaluation model 40. The evaluation model 40 outputs an evaluation index that evaluates the state of the equipment 10 in response to input of status data indicating the state of the equipment 10. In generating such an evaluation model 40, as an example, labeling data may be generated based on the operation target of the equipment 10 (such as a plant KPI (Key Performance Indicator)), status data indicating the state of the equipment 10, and teacher labels. Such an operation target may be an operation target of the entire plant or the entire factory when the equipment 10 is the entire plant or the entire factory, may be an operation target of the segment when the equipment 10 is a segment of the plant or the factory, or may be an operation target of a single device or a single process when the equipment 10 corresponds to a single device or a single process. Then, the generated labeling data may be used as learning data to generate the evaluation model 40 by a machine learning algorithm. The generation process of the evaluation model 40 itself may be arbitrary, so further details will not be described here.
[0098] In this modified example, the control device 20 executes AI control using an operation model. In addition, in this figure, a case where the evaluation model 40 is the same as the evaluation model used in the reinforcement learning of the operation model is shown as an example. That is, the evaluation index used for evaluating the trial operation and the evaluation index used in the reinforcement learning of the operation model may be output from the same model. However, this is not limited to this. The evaluation model 40 may be different from the evaluation model used in the reinforcement learning of the operation model. That is, the evaluation index used for evaluating the trial operation and the evaluation index used in the reinforcement learning of the operation model may be output from different models.
[0099] In this modification, the device 1000 further includes an index acquisition unit 610 instead of the target data determination unit 130. When the detection unit 120 determines that an abnormality in the equipment 10 has been detected, the detection unit 120 attaches an abnormality label to the status data and supplies the status data to the index acquisition unit 610.
[0100] The index acquiring unit 610 acquires an evaluation index that evaluates the state of the equipment 10 in response to inputting the state data to the evaluation model 40. The index acquiring unit 610 may be provided mainly by a communication interface, and may supply the state data to which an anomaly label has been attached by the detection unit 120 to the evaluation model 40 via a network. The evaluation model 40 may output an evaluation index that evaluates the state of the equipment 10 in response to inputting the state data. The index acquiring unit 610 may acquire the evaluation index thus output from the evaluation model 40 via a network. The index acquiring unit 610 supplies the acquired evaluation index to the evaluation unit 170.
[0101] In this modified example, the evaluation unit 170 evaluates the trial operation based on a change in the evaluation index. The evaluation unit 170 may be mainly provided by a CPU, and may derive a change in the evaluation index by analyzing the evaluation index acquired in time series from the index acquisition unit 610. Then, the evaluation unit 170 may evaluate the trial operation based on the change in the evaluation index. The evaluation unit 170 supplies the evaluation result to the search unit 180. The search unit 180 may search for a return operation based on such an evaluation result.
[0102] 7 shows an example of a flow diagram of a method that may be executed by device 1000 according to a modification of the present embodiment. Steps S1400 to S1460 and steps S1475 to S1490 may be the same as those in FIG. 4 except that step S1450 is not included, and detailed description thereof will be omitted here. In this modification, if it is determined in step S1440 that the return process is being executed, the computer advances the process to step S1465.
[0103] In step S1465, the computer acquires an evaluation index for evaluating the state of the equipment 10 in response to inputting the state data to the evaluation model 40. For example, the index acquisition unit 610, which may be implemented in a computer, may supply the state data S to which an anomaly label has been added in step S1420 to the evaluation model 40 via a network.
[0104] The evaluation model 40 may output an evaluation index I that evaluates the state of the equipment 10 in response to the input of the state data S. As an example, when k=1, the evaluation model 40 may output an evaluation index I1 that evaluates the state of the equipment 10 before the control target 15 is controlled according to the trial operation A0 in response to the input of the state data S0. Also, the evaluation model 40 may output an evaluation index I1 that evaluates the state of the equipment 10 after the control target 15 is controlled according to the trial operation A0 in response to the input of the state data S1. In this way, the evaluation model 40 may output the evaluation indexes I0 and I1 before and after the control target 15 is controlled according to the trial operation A0. Then, the index acquisition unit 610 may acquire the evaluation indexes I0 and I1 output from the evaluation model 40 via a network.
[0105] Similarly, when k=2, the evaluation model 40 may output an evaluation index I2 that evaluates the state of the equipment 10 after the control target 15 is controlled according to the trial operation A1 in response to the input of the state data S2. Then, the index acquisition unit 610 may acquire the evaluation index I2 output from the evaluation model 40 via a network. In this manner, for example, the index acquisition unit 610 acquires an evaluation index I that evaluates the state of the equipment 10 in response to the input of the state data S to the evaluation model 40. The index acquisition unit 610 supplies the acquired evaluation index I to the evaluation unit 170.
[0106] In step S1470, the computer evaluates the trial operation based on the change in the evaluation index. For example, the evaluation unit 170, which may be implemented in a computer, may evaluate the trial operation A based on the change in the evaluation index I obtained in step S1465.
[0107] As an example, when k=1, the evaluation unit 170 may calculate the difference i01 by subtracting the evaluation index I1 from the evaluation index I0. When the difference i01 is smaller than 0, that is, when the evaluation index I1 has a higher evaluation than the evaluation index I0, it means that the condition of the equipment 10 has improved by the trial operation A0. In this case, the evaluation unit 170 may give the trial operation A0 an evaluation value greater than 0. At this time, the evaluation unit 170 may give an evaluation value such that the larger the absolute value of the difference i01, the larger the value becomes.
[0108] When the difference i01 is 0, that is, when the evaluation index I0 and the evaluation index I1 are the same, it means that the trial operation A0 has not changed the state of the equipment 10. In this case, the evaluation unit 170 may give an evaluation value of 0 to the trial operation A0.
[0109] When the difference i01 is greater than 0, that is, when the evaluation index I1 has a lower evaluation than the evaluation index I0, it means that the condition of the equipment 10 has deteriorated due to the trial operation A0. In this case, the evaluation unit 170 may give the trial operation A0 an evaluation value smaller than 0. In this case, the evaluation unit 170 may give an evaluation value that is smaller as the absolute value of the difference d01 is larger.
[0110] In this manner, for example, the evaluation unit 170 may evaluate the trial operation A based on a change in the evaluation index I. The evaluation unit 170 provides the evaluation result to the search unit 180. Then, the search unit 180 may search for a return operation based on, for example, such an evaluation result.
[0111] As described above, various methods are conceivable for setting a target to be aimed at when returning the equipment 10 to normal from an abnormality, and it may be unclear which method is optimal for determining target data. In response to this, the device 1000 according to this modification searches for a return operation based on the result of evaluating a trial operation in cooperation with the evaluation model 40. As a result, according to the device 1000 according to this modification, instead of determining specific target data by the target data determination unit 130 when returning the equipment 10 to normal from an abnormality and determining a return operation to approach the target data, it is possible to search for a return operation that improves the evaluation index obtained using the evaluation model 40.
[0112] In particular, when the evaluation model 40 that outputs the evaluation index is the same as the evaluation model used during reinforcement learning of the operation model, the device 1000 of this modified example can search for an operation that is consistent or has minimal contradiction with the operation under AI control as the return operation.
[0113] On the other hand, when an abnormality occurs in the equipment 10, the cause may be in the operation model or evaluation model used for AI control (for example, underlearning, overlearning, or labeling error, etc.). Therefore, when the evaluation model 40 that outputs the evaluation index is different from the evaluation model used in reinforcement learning of the operation model, the device 1000 according to this modification can search for an independent operation that is separated from the operation under AI control as a restoration operation.
[0114] FIG. 8 shows another example of a block diagram of a control system 1 that may include an apparatus 1000 according to a modified example of this embodiment. In this diagram, the same reference numerals are used for components having the same functions and configurations as those in FIG. 5 or FIG. 6, and descriptions thereof will be omitted hereinafter except for the differences. Up to this point, an example has been shown in which the evaluation model 40 is stored outside the apparatus 1000. However, the evaluation model 40 may be stored inside the apparatus 1000. In this diagram, the apparatus 1000 further includes an evaluation model 40.
[0115] In this manner, the device 1000 according to this modification may further include the evaluation model 40. As a result, the device 1000 according to this modification can realize, in one device, a function of searching for a return operation and a function of storing the evaluation model 40. Therefore, the device 1000 according to this modification does not need to communicate with the outside in acquiring the evaluation index, so that it is possible to reduce communication costs and time and prevent information leakage to the outside.
[0116] Various embodiments of the present invention may be described with reference to flow charts and block diagrams, where the blocks may represent (1) stages of a process in which operations are performed or (2) sections of an apparatus responsible for performing the operations. Particular stages and sections may be implemented by dedicated circuitry, programmable circuitry provided with computer readable instructions stored on a computer readable medium, and / or a processor provided with computer readable instructions stored on a computer readable medium. Dedicated circuitry may include digital and / or analog hardware circuitry and may include integrated circuits (ICs) and / or discrete circuits. Programmable circuitry may include reconfigurable hardware circuitry including logical AND, logical OR, logical XOR, logical NAND, logical NOR, and other logical operations, memory elements such as flip-flops, registers, field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), and the like.
[0117] A computer-readable medium may include any tangible device capable of storing instructions that are executed by a suitable device, such that the computer-readable medium having instructions stored thereon comprises an article of manufacture that includes instructions that can be executed to create means for performing the operations specified in the flowcharts or block diagrams. Examples of computer-readable media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, and the like. More specific examples of computer-readable media may include floppy disks, diskettes, hard disks, random access memories (RAMs), read-only memories (ROMs), erasable programmable read-only memories (EPROMs or flash memories), electrically erasable programmable read-only memories (EEPROMs), static random access memories (SRAMs), compact disk read-only memories (CD-ROMs), digital versatile disks (DVDs), Blu-ray (RTM) disks, memory sticks, integrated circuit cards, and the like.
[0118] The computer readable instructions may include either assembler instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk®, JAVA®, C++, etc., and conventional procedural programming languages such as the “C” programming language or similar programming languages.
[0119] Computer readable instructions may be provided to a processor or programmable circuitry of a general purpose computer, special purpose computer, or other programmable data processing apparatus, either locally or over a local area network (LAN), a wide area network (WAN), such as the Internet, etc., to execute the computer readable instructions to create means for performing the operations specified in the flowcharts or block diagrams. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, etc.
[0120] 9 illustrates an example of a computer 9900 in which aspects of the present invention may be embodied in whole or in part. Programs installed on the computer 9900 may cause the computer 9900 to function as or perform operations associated with an apparatus or one or more sections of the apparatus according to an embodiment of the present invention, and / or to perform a process or steps of a process according to an embodiment of the present invention. Such programs may be executed by the CPU 9912 to cause the computer 9900 to perform specific operations associated with some or all of the blocks of the flowcharts and block diagrams described herein.
[0121] The computer 9900 according to this embodiment includes a CPU 9912, a RAM 9914, a graphic controller 9916, and a display device 9918, which are connected to each other by a host controller 9910. The computer 9900 also includes input / output units such as a communication interface 9922, a hard disk drive 9924, a DVD drive 9926, and an IC card drive, which are connected to the host controller 9910 via an input / output controller 9920. The computer also includes legacy input / output units such as a ROM 9930 and a keyboard 9942, which are connected to the input / output controller 9920 via an input / output chip 9940.
[0122] The CPU 9912 operates according to programs stored in the ROM 9930 and the RAM 9914, thereby controlling each unit. The graphics controller 9916 retrieves image data generated by the CPU 9912 into a frame buffer or the like provided in the RAM 9914 or into itself, and causes the image data to be displayed on the display device 9918.
[0123] The communication interface 9922 communicates with other electronic devices over a network. The hard disk drive 9924 stores programs and data used by the CPU 9912 in the computer 9900. The DVD drive 9926 reads programs or data from the DVD-ROM 9901 and provides the programs or data to the hard disk drive 9924 via the RAM 9914. The IC card drive reads programs and data from an IC card and / or writes programs and data to an IC card.
[0124] The ROM 9930 stores therein a boot program or the like executed by the computer 9900 upon activation, and / or a program that depends on the hardware of the computer 9900. The input / output chip 9940 may also connect various input / output units to the input / output controller 9920 via a parallel port, a serial port, a keyboard port, a mouse port, etc.
[0125] The programs are provided by a computer-readable medium such as a DVD-ROM 9901 or an IC card. The programs are read from the computer-readable medium, installed in the hard disk drive 9924, the RAM 9914, or the ROM 9930, which are also examples of computer-readable media, and executed by the CPU 9912. Information processing described in these programs is read by the computer 9900, and brings about cooperation between the programs and the various types of hardware resources described above. An apparatus or method may be constructed by implementing information manipulation or processing according to the use of the computer 9900.
[0126] For example, when communication is performed between the computer 9900 and an external device, the CPU 9912 may execute a communication program loaded in the RAM 9914 and instruct the communication interface 9922 to perform communication processing based on the processing described in the communication program. Under the control of the CPU 9912, the communication interface 9922 reads transmission data stored in a transmission buffer processing area provided in the RAM 9914, the hard disk drive 9924, the DVD-ROM 9901, or a recording medium such as an IC card, and transmits the read transmission data to the network, or writes reception data received from the network to a reception buffer processing area or the like provided on the recording medium.
[0127] The CPU 9912 may also cause all or a necessary portion of a file or database stored on an external recording medium such as the hard disk drive 9924, the DVD drive 9926 (DVD-ROM 9901), an IC card, etc. to be read into the RAM 9914, and perform various types of processing on the data on the RAM 9914. The CPU 9912 then writes back the processed data to the external recording medium.
[0128] Various types of information, such as various types of programs, data, tables, and databases, may be stored in the recording medium and undergo information processing. The CPU 9912 may perform various types of processing on the data read from the RAM 9914, including various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, information search / replacement, etc., as described throughout this disclosure and specified by the instruction sequence of the program, and write back the results to the RAM 9914. The CPU 9912 may also search for information in a file, database, etc. in the recording medium. For example, when a plurality of entries each having an attribute value of a first attribute associated with an attribute value of a second attribute are stored in the recording medium, the CPU 9912 may search for an entry that matches a condition, in which an attribute value of the first attribute is specified, from among the plurality of entries, read the attribute value of the second attribute stored in the entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.
[0129] The above-described programs or software modules may be stored in a computer-readable medium on the computer 9900 or in the vicinity of the computer 9900. In addition, a recording medium such as a hard disk or a RAM provided in a server system connected to a dedicated communication network or the Internet can be used as a computer-readable medium, thereby providing the programs to the computer 9900 via the network.
[0130] Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It is clear to those skilled in the art that various modifications and improvements can be made to the above embodiments. It is clear from the description of the claims that such modifications and improvements can also be included in the technical scope of the present invention.
[0131] It should be noted that the order of execution of each process, such as operations, procedures, steps, and stages, in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not specifically stated as "before," "prior to," etc., and may be realized in any order unless the output of a previous process is used in a later process. Even if the operational flow in the claims, specifications, and drawings is explained using "first," "next," etc. for convenience, it does not mean that it is essential to perform the process in this order. [Explanation of symbols]
[0132] 10 Equipment 15 Control Object 20 Control device 40 Evaluation Models 1000 devices 110 Status data acquisition unit 120 Detection unit 130 Target Data Determination Unit 170 Evaluation Department 180 Search Department 190 Warning section 510 Control section 610 Index acquisition part 9900 Computer 9901 DVD-ROM 9910 Host Controller 9912 CPU 9914 RAM 9916 Graphics Controller 9918 Display Device 9920 Input / Output Controller 9922 Communication Interface 9924 Hard Disk Drive 9926 DVD Drive 9930 ROM 9940 Input / Output Chip 9942 Keyboard
Claims
1. A status data acquisition unit that acquires status data indicating a status of the equipment; an evaluation unit that evaluates the trial operation based on a change in the state data when a control target is controlled in accordance with a trial operation that is attempted to restore the equipment to normal from an abnormality; A search unit that searches for a recovery operation for recovering the equipment from an abnormality to a normal state based on the evaluation result; An apparatus comprising:
2. A target data determination unit determines target data indicating a target state of the equipment for returning the equipment from an abnormality to a normal state, The apparatus according to claim 1 , wherein the evaluation unit evaluates the trial operation based on a change in a distance between the state data and the target data.
3. The apparatus according to claim 2 , wherein the target data determination unit determines the target data based on a plurality of normal data indicating the status data under normal conditions.
4. The apparatus according to claim 3 , wherein the target data determination unit determines the target data based on statistics of the plurality of normal data.
5. The apparatus according to claim 3 , wherein the target data determination unit determines the target data based on a time series of the plurality of normal data.
6. An index acquisition unit that acquires an evaluation index that evaluates the state of the equipment in response to inputting the state data into the evaluation model, The apparatus according to claim 1 , wherein the evaluation unit evaluates the trial operation based on a change in the evaluation index.
7. The apparatus of claim 6 further comprising the evaluation model.
8. The device according to claim 1 , wherein the search unit sequentially changes one operation parameter out of a plurality of operation parameters included in the trial operation when searching for the return operation.
9. The apparatus according to claim 8 , wherein the search section omits changing the one operation parameter to some of a plurality of candidate operation amounts.
10. The apparatus according to claim 1 , further comprising a detection unit that detects an abnormality in the equipment based on the status data.
11. The device according to claim 1 , further comprising a warning unit that issues a warning when the equipment does not return to normal from the abnormality within a predetermined condition even by the searched recovery operation.
12. The device according to claim 11 , wherein the warning unit issues a warning when an elapsed time from when an abnormality in the equipment is detected until the equipment returns to normal exceeds a predetermined threshold.
13. The device according to claim 11 , wherein the warning unit issues a warning when the number of times the control target is controlled in accordance with the trial operation exceeds a predetermined threshold value.
14. The device according to claim 1 , further comprising a control unit that controls the control target in accordance with the trial operation.
15. The computer Obtaining status data indicative of a status of the equipment; Evaluating the trial operation based on a change in the status data when a control target is controlled according to the trial operation that is attempted to restore the equipment to normal from an abnormality; searching for a recovery operation for recovering the equipment from the abnormality to a normal state based on the evaluation result; A method comprising:
16. The method is executed by a computer, causing the computer to: A status data acquisition unit that acquires status data indicating a status of the equipment; an evaluation unit that evaluates the trial operation based on a change in the state data when a control target is controlled in accordance with a trial operation that is attempted to restore the equipment to normal from an abnormality; A search unit that searches for a recovery operation for recovering the equipment from an abnormality to a normal state based on the evaluation result; A program that makes things work.
Citation Information
Patent Citations
Plant monitoring system
JP2004133596A
