Computer system, maintenance determination assistance method, and system

Abstract

To assist in maintenance determination for a facility.SOLUTION: A computer system is connected to a facility comprising a plurality of devices, and retains threshold value information for managing threshold values for classifying, on the basis of failure probabilities of the devices, the devices into one of a first state indicating a state in which maintenance is required, a second state indicating a state in which it is necessary to determine whether or not to perform maintenance, and a third state indicating a state in which maintenance is not required. The threshold value information includes, for each device, a maintenance threshold value specifying a boundary between the first state and the second state, and an abnormality threshold value specifying a boundary between the second state and the third state. The computer system performs processing of determining the maintenance threshold value for each device, and processing of calculating the failure probability of the device, classifying the device into one of the first state, the second state, and the third state on the basis of the failure probability of the device and the threshold value information, and outputting determination assistance information including classification results.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a technology for supporting judgment as to whether or not maintenance of equipment constituting a facility is required. [Background technology]

[0002] Facilities in the electric power and railway industries are becoming more and more sophisticated year by year. In particular, advances in sensor and IoT technologies in recent years have made it possible to monitor the condition of the equipment that makes up the facilities with high accuracy. As a result, facility maintenance is beginning to shift from time-based maintenance (TBM), which periodically maintains facilities, to condition-based maintenance (CBM), which maintains facilities according to their condition, in order to reduce maintenance costs.

[0003] As a technology to support maintenance decisions based on the idea of ​​CBM, for example, a technology has been proposed that considers the trade-off between the failure risk and maintenance cost of devices that make up the equipment to be maintained, and creates an optimal maintenance plan in terms of both cost and risk (see, for example, Patent Document 1). The technology described in Patent Document 1 indicates whether the state of each device falls into an area requiring preventive maintenance or an area not requiring preventive maintenance at the time of inspection.

[0004] Moreover, Patent Document 2 discloses a technique for selecting devices as candidates for preventive maintenance, calculating an evaluation value for each selected device, and prioritizing the maintenance. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2005-182465 A [Patent Document 2] JP 2020-4403 A Summary of the Invention [Problem to be solved by the invention]

[0006] When there is a large number of devices to be maintained, the technology described in Patent Document 1 results in very high maintenance costs, while the technology described in Patent Document 2 limits the number of devices to be maintained, which may hinder safe operation.

[0007] An object of the present invention is to provide a system and method for presenting information that supports a decision on whether or not to perform maintenance on equipment, taking into account operational risks and maintenance costs. [Means for solving the problem]

[0008] A representative example of the invention disclosed in the present application is as follows: That is, a computer system including at least one computer is connected to a facility consisting of at least one device, and holds threshold information for managing thresholds for classifying a state into a first state indicating a state in which maintenance is required, a second state indicating a state in which it is necessary to determine whether or not maintenance is performed, and a third state indicating a state in which maintenance is not required based on a failure probability of the device, the threshold information including a maintenance threshold specifying a boundary between the first state and the second state and an abnormality threshold specifying a boundary between the second state and the third state for each of the devices, and the computer system executes a first process of determining the maintenance threshold for each of the devices, and a second process of calculating a failure probability of the device, classifying the device into the first state, the second state, or the third state based on the failure probability of the device and the threshold information, and outputting decision support information including the result of the classification. Effect of the Invention

[0009] According to the present invention, it is possible to present information that supports the decision of whether or not to perform maintenance of equipment, taking into account operational risks and maintenance costs. Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 illustrates an example of a system configuration according to a first embodiment. [Diagram 2] 1 is a diagram illustrating an example of a hardware configuration of a computer that configures the maintenance decision support system of the first embodiment. [Diagram 3] FIG. 4 is a diagram illustrating an example of a data structure of a device parameter DB according to the first embodiment. [Figure 4] FIG. 2 is a diagram illustrating an example of a data structure of a facility schedule DB according to the first embodiment. [Figure 5A] FIG. 4 is a diagram illustrating an example of a data structure of a device state DB according to the first embodiment. [Figure 5B] FIG. 4 is a diagram illustrating an example of a data structure of a device state DB according to the first embodiment. [Figure 6] 11 is a flowchart illustrating an example of a threshold calculation process executed by the maintenance decision support system according to the first embodiment. [Figure 7] 11 is a flowchart illustrating an example of an equipment model generation process executed by an equipment model generation unit according to the first embodiment. [Figure 8] FIG. 4 is a diagram illustrating an example of data processing in the equipment model generation processing of the first embodiment. [Figure 9] 11 is a flowchart illustrating an example of a maintenance determination model generation process executed by a maintenance determination model generation unit according to the first embodiment. [Figure 10] FIG. 11 is a diagram illustrating an example of data processing in the maintenance determination model generation processing in the first embodiment. [Figure 11A] 11 is a flowchart illustrating an example of a maintenance simulation executed by a simulation unit according to the first embodiment. [Figure 11B] 11 is a flowchart illustrating an example of a maintenance simulation executed by a simulation unit according to the first embodiment. [Figure 12] FIG. 13 is a diagram illustrating an example of a data structure of a simulation result DB according to the first embodiment. [Figure 13] FIG. 4 is a diagram illustrating a specific example of an internal process executed by a simulation unit according to the first embodiment. [Figure 14] 11 is a flowchart illustrating an example of an analysis process executed by an analysis unit according to the first embodiment. [Figure 15A]FIG. 11 is a diagram illustrating an example of a change in a maintenance threshold of the device according to the first embodiment. [Figure 15B] FIG. 11 is a diagram illustrating an example of a change in a maintenance threshold of the device according to the first embodiment. [Figure 16A] FIG. 11 is a diagram illustrating an example of data processing in the analysis process of the first embodiment. [Figure 16B] FIG. 11 is a diagram illustrating an example of data processing in the analysis process of the first embodiment. [Figure 16C] FIG. 11 is a diagram illustrating an example of data processing in the analysis process of the first embodiment. [Figure 17A] FIG. 13 is a diagram illustrating an example of a data structure of an analysis result DB according to the first embodiment. [Figure 17B] FIG. 13 is a diagram illustrating an example of a data structure of an analysis result DB according to the first embodiment. [Figure 18] 11 is a flowchart illustrating an example of a presentation process of decision support information executed by an analysis unit according to the first embodiment. [Figure 19] FIG. 11 is a diagram illustrating an example of a screen displayed on the terminal of the first embodiment. [Figure 20] FIG. 2 is a sequence diagram showing a method of using the maintenance decision support system of the first embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. However, the present invention should not be interpreted as being limited to the description of the embodiment shown below. It will be easily understood by those skilled in the art that the specific configuration can be changed without departing from the concept or purpose of the present invention.

[0012] In the configurations of the invention described below, the same or similar configurations or functions are given the same reference numerals, and duplicated explanations are omitted.

[0013] In this specification, the terms "first," "second," "third," and the like are used to identify components and do not necessarily limit the number or order.

[0014] In this specification, equipment maintenance is a concept that includes replacement and repair of equipment or parts that constitute the equipment. Maintenance performed before a failure occurs is described as preventive maintenance, and maintenance performed after a failure occurs is described as corrective maintenance. When there is no distinction between preventive maintenance and corrective maintenance, they are referred to as maintenance. EXAMPLES

[0015] Fig. 1 is a diagram illustrating an example of a system configuration of the embodiment 1. Fig. 2 is a diagram illustrating an example of a hardware configuration of a computer that configures the maintenance decision support system of the embodiment 1.

[0016] The system includes a maintenance decision support system 100, a plurality of terminals 101, and a plurality of facilities 102. The maintenance decision support system 100, the plurality of terminals 101, and the plurality of facilities 102 are connected to each other via a network 103.

[0017] The terminal 101 is a terminal used by a user such as a worker on-site, etc. The terminal 101 includes a processor, a memory, an input device, an output device, and a communication device, all of which are not shown.

[0018] The facility 102 is composed of multiple devices (not shown). The facility 102 includes a sensor device 104 that monitors the status of the devices. The present invention is not limited to the type and size of the facility 102, and the type and number of devices that constitute the facility.

[0019] The maintenance decision support system 100 supports a decision on whether or not to perform maintenance of equipment. The maintenance decision support system 100 is configured with, for example, a computer 200 as shown in FIG. 2. The computer 200 includes a processor 201, a main memory device 202, an auxiliary memory device 203, a communication device 204, an input device 205, and an output device 206. Each hardware element is connected via an internal bus or the like. Note that the computer 200 does not necessarily have to include the input device 205 and the output device 206.

[0020] The input device 205 is a keyboard, a mouse, a touch panel, etc., and accepts input of information, commands, etc. The output device 206 is, for example, a display, etc., and outputs various types of information.

[0021] The communication device 204 communicates with external devices. The communication device 204 receives information from the terminal 101 and the facility 102, and stores the information in the auxiliary storage device 203.

[0022] The processor 201 executes a program stored in the main memory device 202. The processor 201 executes processing according to the program, thereby operating as a functional unit (module) that realizes a specific function. In the following explanation, when a process is explained with a functional unit as the subject, it indicates that the processor 201 is executing a program that realizes the functional unit.

[0023] The main storage device 202 is a memory or the like, and stores the programs executed by the processor 201. The main storage device 202 is also used as a work area. The main storage device 202 stores programs that realize the equipment model generation unit 110, the maintenance judgment model generation unit 111, the simulation unit 112, the analysis unit 113, and the maintenance threshold calculation unit 114.

[0024] The auxiliary storage device 203 is a hard disk drive (HDD) or a solid state drive (SSD), etc., and permanently stores data. The auxiliary storage device 203 stores an equipment parameter DB 120, an equipment schedule DB 121, an equipment status DB 122, a simulation result DB 123, and an analysis result DB 124.

[0025] The information stored in the auxiliary storage device 203 may be stored in the main storage device 202. The programs stored in the main storage device 202 may be stored in the auxiliary storage device 203.

[0026] FIG. 3 is a diagram illustrating an example of a data structure of the device parameter DB 120 according to the first embodiment.

[0027] The equipment parameter DB 120 stores a table 300 for each piece of equipment 102. The table 300 stores entries including an equipment ID 301, a group ID 302, a failure cost 303, a maintenance cost 304, an equipment model parameter 305, a maintenance judgment model parameter 306, and an abnormality threshold 307. One entry exists for one piece of equipment.

[0028] The equipment ID 301 is a field that stores an ID for uniquely identifying equipment that constitutes the equipment. The group ID 302 is a field that stores an ID for uniquely identifying a group of equipment. In this embodiment, equipment is managed in group units, and equipment is treated as being composed of groups. The group ID 302 in FIG. 2 stores an ID that can identify the equipment 102 and the group. Specifically, "INS001" represents the equipment 102, and "EQG001" represents the group.

[0029] The breakdown cost 303 is a field for storing the cost incurred when a breakdown occurs during operation. The cost is set for each group, and includes the cost of loss due to the stoppage of operation of the equipment and the cost of corrective maintenance.

[0030] The maintenance cost 304 is a field for storing the cost required for maintaining the device. The cost is set for each device.

[0031] The equipment model parameters 305 are a group of fields that store parameters of a model (equipment model) for estimating the lifespan of a device.

[0032] The maintenance judgment model parameters 306 are a group of fields that store parameters of a model for estimating the probability of performing maintenance.

[0033] The abnormality threshold 307 is a field for storing the threshold of the failure probability (abnormality threshold) at which maintenance must be performed. The abnormality threshold is a value determined according to the material and type of the equipment, and is set in advance in the equipment. The abnormality threshold may be obtained directly from the equipment, or may be input by the user.

[0034] Values ​​set in advance by a user or the like are stored in device ID 301, group ID 302, failure cost 303, maintenance cost 304, and abnormality threshold 307. The stored values ​​do not change unless the configuration of the equipment is changed. On the other hand, the information stored in equipment model parameters 305 and maintenance judgment model parameters 306 may be updated by processing in equipment model generation unit 110 and maintenance judgment model generation unit 111.

[0035] Here, the facility model and the maintenance judgment model will be described.

[0036] An equipment model is a model for calculating the failure probability over time, and in the field of reliability engineering, Weibull distribution estimation is often used. This embodiment will also be described using failure probability estimation based on Weibull distribution. In general, if it is assumed that the life of a certain device follows the Weibull distribution, the equipment model is given by equation (1). Here, m represents the Weibull coefficient, and η represents the scale.

[0037]

number

[0038] The equipment model of the equipment is characterized by the Weibull coefficient m and the scale η. Therefore, the equipment model parameters 305 of the second embodiment include fields for storing the Weibull coefficient m and the scale η.

[0039] In the first embodiment, the failure probability of the device is defined by the formula (2).

[0040]

number

[0041] The failure probability shown in formula (2) indicates the failure probability of the device with device ID "j" from time t to time t+dt. F j (t) represents the probability that a device with ID "j" will fail by time t, and is defined by equation (3). Here, m j , η j represents the parameters of the equipment model of the device with device ID “j”.

[0042]

number

[0043] In addition, since there are time periods when the equipment 102 is not in operation, the time t in formula (1) may be corrected to the time during operation. In addition, since the load on each device differs depending on the method of operation, when the load is high, the time t may be multiplied by a coefficient of 1 or more.

[0044] The maintenance decision model is a model that calculates the probability of performing maintenance according to the failure probability of equipment. The simplest maintenance decision model is one in which the lower the failure probability, the lower the probability of performing preventive maintenance, and the higher the failure probability, the higher the probability of performing maintenance. This can be defined in a form obtained by modifying the sigmoid function shown in equation (4).

[0045]

number

[0046] P j (t, dt) is the formula defined in formula (2). a is a parameter that determines the shape of the sigmoid function; the larger the value, the steeper the rise of the model. b is an offset term, a parameter used to adjust the position of the failure probability where the value of formula (3) becomes 0.5.

[0047] In other words, the maintenance decision model is a model that represents the probability that a maintenance worker will perform maintenance for each device failure probability, and the M of devices that do not require maintenance is j (t, dt) is a value that can be approximated to 0, and M j (t, dt) is a continuous value ranging from 0 to 1, and M j (t,dt) is a value that can be approximated to 1.

[0048] FIG. 4 is a diagram illustrating an example of a data structure of the facility schedule DB 121 according to the first embodiment.

[0049] The facility schedule DB 121 stores a table 400 for each facility 102. The table 400 is a table for managing the inspection and operation schedule of the facility 102, and stores entries including a start date and time 401, an end date and time 402, an action 403, and a load 404. One entry exists for the schedule of one action.

[0050] Start date / time 401 and end date / time 402 are fields for storing the start date / time and end date / time of the schedule. Action 403 is a field for storing the type of action. Load 404 is a field for storing the load of the equipment. For entries where action 403 is "inspection", load 404 is blank. The value of load 404 can be used as a value by which time t in equation (1) is multiplied.

[0051] 5A and 5B are diagrams illustrating an example of a data structure of the device state DB 122 according to the first embodiment.

[0052] The device status DB 122 stores a table 500 and a table 510 .

[0053] The table 500 is a table for managing measurement information received from the sensor device 104, and stores entries including a device ID 501, an elapsed time 502, and a measurement value 503. There is one entry for one device.

[0054] The device ID 501 is the same field as the device ID 301. The elapsed time 502 is a field that stores the time that has elapsed since the device was installed in the facility. The measured value 503 is a field that stores the measured value or a value calculated from the measured value.

[0055] In this embodiment, the sensor device 104 uses a measurement value that indicates the shape, state, etc. of the device to calculate an abnormal value that is in the range from 0 to 1, and transmits measurement information including the elapsed time and the abnormal value. The abnormal value is stored in the measurement value 503.

[0056] Table 510 is a table for managing device information received from terminal 101, and stores entries including a device ID 511, operation start date and time 512, maintenance date and time 513, failure date and time 514, and measurement value 515. One entry exists for one device on which maintenance has been performed by a worker.

[0057] The device ID 511 is the same field as the device ID 301. Since the table 510 stores information about devices that have been maintained or replaced, multiple entries for devices with the same ID may be stored. The operation start date and time 512 is a field that stores the operation start date and time of the device. The maintenance date and time 513 is a field that stores the date and time when preventive maintenance was performed. A value is stored only in the maintenance date and time 513 of an entry corresponding to a device that has been subjected to preventive maintenance. The failure date and time 514 is a field that stores the date and time of failure of a device that has been subjected to corrective maintenance. A value is stored only in the failure date and time 514 of an entry corresponding to a device that has been subjected to corrective maintenance. The measurement value 515 is a field that stores the measurement result of the sensor device 104 when maintenance was performed. An abnormal value is stored in the measurement value 515.

[0058] FIG. 6 is a flowchart illustrating an example of a threshold calculation process executed by the maintenance decision support system 100 according to the first embodiment.

[0059] When the maintenance decision support system 100 receives a processing request including a simulation condition, it starts the processing described below.

[0060] The maintenance threshold calculation unit 114 acquires the simulation conditions included in the processing request (step S101). The simulation conditions include, for example, a simulation period, a target facility, hyperparameters of a Monte Carlo tree search described later, etc. Note that information specifying a device to be simulated may also be included.

[0061] The maintenance threshold calculation unit 114 calls the equipment model generation unit 110. The equipment model generation unit 110 executes an equipment model generation process (step S102). Details of the equipment model generation process will be described with reference to FIG.

[0062] After completing the equipment model generation process, the maintenance threshold calculation unit 114 calls the maintenance judgment model generation unit 111. The maintenance judgment model generation unit 111 executes the maintenance judgment model generation process (step S103). The details of the maintenance judgment model generation process will be described with reference to FIG. 9.

[0063] The maintenance threshold calculation unit 114 initializes the simulation result DB 123 (step S104).

[0064] The maintenance threshold calculation unit 114 calls the simulation unit 112. The simulation unit 112 executes a maintenance simulation of the equipment (step S105). Details of the maintenance simulation will be described with reference to Figs. 11A and 11B.

[0065] The maintenance threshold calculation unit 114 determines whether or not a convergence condition is satisfied (step S106), where the convergence condition is that the number of times the maintenance simulation is executed is greater than a threshold, or that the rate of change of the maintenance threshold is smaller than a threshold.

[0066] If the convergence condition is not satisfied, the maintenance threshold calculation unit 114 updates the maintenance threshold using the simulation result DB 123 that stores the results of the maintenance simulation (step S107), and then returns to step S105.

[0067] The maintenance threshold calculation unit 114 updates the maintenance threshold using, for example, the Markov-Chain Monte-Carlo method, etc. In the MCMC method, for example, the maintenance threshold is updated so as to maximize the value of equation (5).

[0068]

number

[0069] Here, θ i (i is 1 to n) represents the maintenance threshold of the equipment i. δ represents the hyperparameter. Δ is a value for evaluating the results of the maintenance simulation, and is defined, for example, as in equation (6).

[0070]

number

[0071] Here, Cost budget represents the budget amount, and Cost SIM represents the maintenance cost calculated by the maintenance simulation. Formula (6) is the value when the constraint condition is set that the maintenance cost does not exceed the budget amount. Note that the value may also take into account the constraint condition that the difference between the maintenance cost and the budget amount is equal to or less than a certain value, and the constraint condition that the maintenance cost over multiple periods is within a certain budget amount.

[0072] In addition, a constraint may be added such that the cost when a device in each device group fails, as shown in equation (7), is always equal to or less than a certain value.

[0073]

number

[0074] In addition, a constraint may be added that, based on the failure probability of the equipment in each equipment group shown in equation (7), the cost in the event of equipment failure shown in equation (8) is always kept below a certain value.

[0075]

number

[0076] The above-mentioned method of updating the maintenance threshold is merely an example and is not intended to be limiting. The above is an explanation of step S107.

[0077] If the convergence condition is satisfied, the maintenance threshold calculation unit 114 calls the analysis unit 113. The analysis unit 113 executes an analysis process (step S108). Details of the analysis process will be described with reference to FIG. 14. After the analysis process is completed, the maintenance threshold calculation unit 114 ends the series of processes.

[0078] The equipment model generation process and the maintenance judgment model generation process may be omitted. Furthermore, the equipment model generation process and the maintenance judgment model generation process may be executed at different execution timings.

[0079] As described later, the maintenance decision support system 100 of the first embodiment classifies the equipment into one of a state in which maintenance is required (first state), a state in which it is necessary to determine whether or not to perform maintenance (second state), and a state in which maintenance is not required (third state) using a maintenance threshold and an abnormality threshold. Specifically, if the failure probability is smaller than the maintenance threshold, it is classified into the third state, if the failure probability is equal to or greater than the maintenance threshold and smaller than the abnormality threshold, it is classified into the second state, and if the failure probability is equal to or greater than the abnormality threshold, it is classified into the third state. In the following description, the equipment in the first state is referred to as a target equipment, the equipment in the second state is referred to as a candidate equipment, and the equipment in the third state is referred to as a non-target equipment.

[0080] Fig. 7 is a flowchart illustrating an example of the equipment model generation process executed by the equipment model generating unit 110 according to the first embodiment. Fig. 8 is a diagram illustrating an example of data processing in the equipment model generation process according to the first embodiment.

[0081] The equipment model generation unit 110 selects one entry from the equipment parameter DB 120 (step S201). In the following, the equipment corresponding to the selected entry is referred to as a target equipment.

[0082] The equipment model generation unit 110 acquires an entry related to a failure of the target equipment from the table 510 of the equipment status DB 122 (step S202). Specifically, the equipment model generation unit 110 acquires an entry in which the ID of the target equipment is stored in the equipment ID 511 and the date and time is stored in the failure date and time 514.

[0083] The equipment model generation unit 110 determines whether the number of acquired entries is greater than a threshold value (step S203). The threshold value is, for example, 20.

[0084] If the number of acquired entries is equal to or less than the threshold, the equipment model generation unit 110 proceeds to step S206.

[0085] If the number of acquired entries is greater than the threshold value, the equipment model generation unit 110 calculates the parameters of the equipment model using the acquired entries (step S204).

[0086] Specifically, for each entry, the equipment model generation unit 110 calculates the operation years by subtracting the operation start date and time 512 of the entry from the failure date and time 514. The equipment model generation unit 110 generates a histogram of the operation years, and calculates the Weibull coefficient m and scale η to statistically represent the histogram.

[0087] 8, the solid line 801 statistically represents the histogram. Therefore, the Weibull coefficient m and the scale η are calculated to represent the solid line 801.

[0088] The equipment model generation unit 110 updates the equipment model parameters 305 of the entry corresponding to the target equipment in the equipment parameter DB 120 based on the calculation result (step S205). After that, the equipment model generation unit 110 proceeds to step S206.

[0089] In step S206, the equipment model generation unit 110 determines whether or not the processing has been completed for all entries in the equipment parameter DB 120 (step S206).

[0090] If the processing has not been completed for all entries in the equipment parameter DB 120, the equipment model generation unit 110 returns to step S201. If the processing has been completed for all entries in the equipment parameter DB 120, the equipment model generation unit 110 ends the equipment model generation processing.

[0091] Fig. 9 is a flowchart illustrating an example of the maintenance judgment model generation process executed by the maintenance judgment model generation unit 111 according to the first embodiment. Fig. 10 is a diagram illustrating an example of data processing in the maintenance judgment model generation process according to the first embodiment.

[0092] The maintenance judgment model generating unit 111 selects one entry from the device parameter DB 120 (step S301). In the following, the device corresponding to the selected entry is referred to as a target device.

[0093] The maintenance judgment model generation unit 111 acquires an entry related to preventive maintenance of the target equipment from the table 510 of the equipment status DB 122 (step S302). Specifically, the equipment model generation unit 110 acquires an entry in which the ID of the target equipment is stored in the equipment ID 511 and the date and time is stored in the maintenance date and time 513.

[0094] The maintenance judgment model generating unit 111 judges whether the number of acquired entries is greater than a threshold value (step S303). The threshold value is, for example, 20.

[0095] If the number of acquired entries is greater than the threshold, the maintenance determination model generation unit 111 calculates parameters of the maintenance determination model using the acquired entries (step S304), after which the maintenance determination model generation unit 111 proceeds to step S306.

[0096] Specifically, the maintenance judgment model generation unit 111 calculates the failure probability of the target device during maintenance for the acquired entry by using an equipment model of the target device. The maintenance judgment model generation unit 111 generates a histogram of the failure probability of the target device, and calculates parameters a and b to express the histogram.

[0097] In the case of the histogram of the failure probability of the target device shown in Fig. 10, the solid line 1001 statistically represents the histogram. Therefore, parameters a and b are calculated so as to represent the solid line 1001. Note that the range used in the processing may be set with the abnormality threshold as the upper limit. In Fig. 10, the histogram of the range 1002 is used.

[0098] If the number of acquired entries is equal to or less than the threshold, the maintenance determination model generation unit 111 calculates parameters of the maintenance determination model according to the rules (step S305), after which the maintenance determination model generation unit 111 proceeds to step S306.

[0099] As a rule in step S305, for example, the maintenance judgment model generating unit 111 sets a very large value such as 1e99 to parameter a, and sets a value obtained by multiplying the abnormality threshold by 0.5 to parameter b. By setting the parameters as described above, a step function is formed in which the probability changes from 0 to 1 when the failure probability is half the abnormality threshold.

[0100] In step S306, the maintenance determination model generating unit 111 updates the maintenance determination model parameters 306 of the entry corresponding to the target device in the device parameter DB 120 based on the calculation result (step S306).

[0101] The maintenance judgment model generating unit 111 judges whether or not the processing has been completed for all the entries in the device parameter DB 120 (step S307).

[0102] If the processing has not been completed for all entries in the equipment parameter DB 120, the maintenance judgment model generation unit 111 returns to step S301. If the processing has been completed for all entries in the equipment parameter DB 120, the maintenance judgment model generation unit 111 ends the maintenance judgment model generation process.

[0103] Although the failure probability of the equipment is calculated using the number of years of operation, it may be calculated using the measurement value 515 in the table 510 of the equipment status DB 122.

[0104] In addition, since the equipment failure probability calculated using formulas (1), (2), and (3) is a theoretical value, it may be corrected using the measurement value of the sensor device 104. For example, if it is assumed that the value of the measurement value 515 is in the vicinity of the abnormality threshold, the equipment failure probability can be estimated from the value of the measurement value 515. Using this, the calculated equipment failure probability may be corrected.

[0105] Fig. 11A and Fig. 11B are flowcharts for explaining an example of a maintenance simulation executed by the simulation unit 112 of the embodiment 1. Fig. 12 is a diagram showing an example of a data structure of the simulation result DB 123 of the embodiment 1. Fig. 13 is a diagram showing a specific example of an internal process executed by the simulation unit 112 of the embodiment 1.

[0106] In the maintenance simulation, the execution dates and times of inspection and operation actions are set as time steps, and the necessity and state of equipment maintenance at each time step are simulated.

[0107] If the simulation period is from "2021 / 12 / 10" to "2021 / 12 / 18", according to table 400 shown in Figure 4, time step T = 0 is "2021 / 12 / 10 13:00" and time step T = 1 is "2021 / 12 / 16 09:00".

[0108] The simulation unit 112 refers to the facility schedule DB 121, sets a time step in a simulation period, and initializes a variable T representing the time step to 0 (step S401). At this time, the simulation unit 112 also initializes a preventive maintenance cost and a corrective maintenance cost to 0.

[0109] The simulation unit 112 records the maintenance threshold value in the simulation result DB 123 (step S402).

[0110] 12, the simulation result DB 123 stores a table 1200. The table 1200 stores entries including a simulation ID 1201, a preventive maintenance cost 1202, a corrective maintenance cost 1203, and a maintenance threshold value 1204. One entry exists for one maintenance simulation.

[0111] The simulation ID 1201 is a field for storing the ID of a maintenance simulation. The preventive maintenance cost 1202 and the corrective maintenance cost 1203 are fields for storing the preventive maintenance cost and the corrective maintenance cost calculated by the maintenance simulation. The maintenance threshold 1204 is a field group for storing the maintenance threshold of the equipment.

[0112] In step S402, the simulation unit 112 adds an entry to the table 1200, sets an ID in the simulation ID 1201 of the added entry, and sets the security threshold of each device in the security threshold 1204. When information on the security threshold is stored in the analysis result DB 124, the simulation unit 112 may set the security threshold based on the information. Also, the security threshold may be set randomly.

[0113] The simulation unit 112 calculates the failure probability of the equipment in the time step T (step S403). Specifically, the simulation unit 112 calculates the operation years from the date and time corresponding to the time step T and the installation date and time of the equipment, and calculates the failure probability of the equipment by substituting the operation years for the variable t in the formula (2).

[0114] The simulation unit 112 determines whether the action for the time step T is “operation” (step S404). Specifically, the simulation unit 112 makes the determination based on the action 403 of the entry corresponding to the time step T in the table 400.

[0115] When the action for the time step T is "operation", the simulation unit 112 determines whether or not there is any equipment in which a failure has occurred (step S405).

[0116] Specifically, the simulation unit 112 generates a random number ranging from 0 to 1 for each device, and determines whether the random number is greater than the device's failure probability. If the random number is greater than the device's failure probability, the simulation unit 112 determines that the device is faulty.

[0117] If there is no device in which a malfunction has occurred, the simulation unit 112 proceeds to step S410.

[0118] If a faulty device is present, the simulation unit 112 updates the installation date and time of the device to the date and time corresponding to the time step T, and also updates the corrective maintenance cost (step S406). After that, the simulation unit 112 proceeds to step S410.

[0119] Specifically, the simulation unit 112 adds the total value of the maintenance cost 304 of the device in which the failure has occurred and the failure cost 303 of the device group to which the device belongs to, to the corrective maintenance cost.

[0120] In step S404, if the action for the time step T is "inspection", the simulation unit 112 determines whether or not one or more target devices and candidate devices exist (step S407).

[0121] Specifically, the simulation unit 112 classifies the devices based on the failure probability of each device, and determines whether or not there is at least one device in the first state and at least one device in the second state.

[0122] If only non-target devices exist, the simulation unit 112 proceeds to step S410.

[0123] When there is one or more target devices or candidate devices, the simulation unit 112 determines whether or not to implement device maintenance for each candidate device (step S408).

[0124] When maintenance of equipment is performed at a certain time step, the number of years of operation of the equipment is initialized, which also affects the decision on whether or not to perform maintenance at the next time step. When considering how to bring the total maintenance cost after M steps closer to the budgeted amount, the decision on whether or not to perform maintenance on the equipment at the current time step requires considering the combination of equipment to be maintained at each time step. If the number of equipment is N, the number of combinations mentioned above is 2 N×M This requires a large amount of calculation. Therefore, in this embodiment, a Monte Carlo tree search is used to efficiently search for combinations. Note that if the Monte Carlo tree calculation is performed an infinite number of times, the same result as that obtained by calculating combinations of maintenance implementation / non-implementation of all devices at all time steps can be obtained.

[0125] 13 shows an overview of the Monte Carlo tree search in the first embodiment. The nodes include operation nodes, maintenance nodes, and state nodes. The operation nodes represent the operation of the equipment, and the maintenance nodes represent the inspection of the equipment. The state represents the state of the equipment during operation or inspection.

[0126] The numerical values ​​of the state nodes in Fig. 13 represent a failure probability index and a total maintenance cost index. For ease of explanation, a failure probability index that is an integer between 0 and 10 is used instead of the failure probability, and a total maintenance cost index that is an integer greater than or equal to 0 is used.

[0127] In the following description, it is assumed that state node 1301 at time step T is (1,0).

[0128] The simulation unit 112 calculates the maintenance probability for the maintenance node 1302 by inputting the failure probability of the equipment at time step T+1 into the maintenance decision model. The simulation unit 112 selects either an action of performing or not performing maintenance based on the maintenance probability. When the action 1352 of not performing maintenance is selected, a transition is made to state node 1304 in which 1 has been added to the failure probability index of state node 1301. When the action 1351 of performing maintenance is selected, a transition is made to state node 1303 in which the failure probability index of state node 1301 has been added to 1. In reality, the maintenance cost is added to the total maintenance cost.

[0129] At time step (T+1), the state is operated in states of state node 1303 and state node 1304 .

[0130] The simulation unit 112 determines either a normal or failure action for the operation node 1305 based on the failure probability of the equipment at time step (T+2). When the normal action 1354 is selected, a transition is made to state node 1307 in which 1 has been added to the failure probability index of state node 1304. When the failure action 1353 is selected, a transition is made to state node 1306 in which the failure probability index is initialized to 0 and 1 has been added to the total maintenance cost index of state node 1304. In practice, the corrective maintenance cost is added to the total maintenance cost.

[0131] As described above, the maintenance node and the operation node can select either an action in which the failure probability index increases, or an action in which the failure probability index is initialized and the total maintenance cost index increases.

[0132] The simulation unit 112 similarly expands the tree up to, for example, time step (T+3), and then executes a simulation for each state node at time step (T+3) up to the terminal. The terminal is, for example, time step (T+n). The simulation unit 112 assigns an evaluation value to each state node at time step (T+3) based on the result of the simulation. For example, the average value of the difference between the total maintenance cost and the budget amount is calculated as the evaluation value.

[0133] The simulation unit 112 selects a node based on the evaluation value, and also performs expansion, simulation, and evaluation for each state node in time step (T+3). The evaluation value of each expanded state node is updated with the evaluation value of a node closer to state node 1301 than the expanded state node. For example, the simulation unit 112 adds the evaluation value assigned to state node 1307 to state node 1304, calculates the average value, and updates the evaluation value of state node 1304.

[0134] The above-described process is repeated, and an action that transitions to the node with the best evaluation value assigned to the state node 1301 is selected as the action for the maintenance node 1302.

[0135] This concludes the description of the process in step S408.

[0136] The simulation unit 112 adds the cost required for the maintenance of the target device and the candidate device for which maintenance implementation has been selected to the preventive maintenance cost (step S409), and then proceeds to step S410. At this time, the installation time of the target device and the candidate device for which maintenance implementation has been selected is set to the time step T.

[0137] In step S410, the simulation unit 112 advances the time step T by one (step S410).

[0138] The simulation unit 112 performs a simulation for a time step T END It is determined whether or not this is the case (step S411).

[0139] Time step T is the terminal time step T END If not, the simulation unit 112 returns to step S403.

[0140] Time step T is the terminal time step T END If so, the simulation unit 112 records the preventive maintenance cost and the corrective maintenance cost in the simulation result DB 123 (step S412), and then ends the maintenance simulation.

[0141] Fig. 14 is a flowchart for explaining an example of analysis processing executed by the analysis unit 113 of the first embodiment. Figs. 15A and 15B are diagrams showing an example of a change in the maintenance threshold of the device of the first embodiment. Figs. 16A, 16B, and 16C are diagrams showing an example of data processing in the analysis processing of the first embodiment. Figs. 17A and 17B are diagrams showing an example of a data structure of the analysis result DB 124 of the first embodiment.

[0142] The analysis unit 113 refers to the simulation result DB 123 and determines the maintenance threshold value for each device (step S501).

[0143] Fig. 15A shows the change in the maintenance threshold of the device. Fig. 15B is a histogram showing the distribution of the maintenance threshold. As shown in Fig. 15A and Fig. 15B, the maintenance threshold of each device converges to a certain range. The analysis unit 113 determines a statistical value such as the median, maximum likelihood value, or average value of the distribution shown in Fig. 15B as the maintenance threshold of the device.

[0144] The analysis unit 113 calculates, for each device group, an influence degree that indicates the magnitude of the influence caused by a failure of a device that constitutes the device group (step S502).

[0145] Specifically, the analysis unit 113 calculates the degree of influence, which takes a discrete value, based on the value of the equipment group's failure cost 303. For example, a method of assigning integers in descending order of the failure cost 303 is conceivable.

[0146] The relationship between the impact degree and the maintenance threshold is expressed as a scatter diagram as shown in Figure 16A. The vertical axis represents the impact degree, and the horizontal axis represents the maintenance threshold. In Figure 16A, integers from 0 to 9 are assigned to the impact degrees.

[0147] The analysis unit 113 calculates a group maintenance threshold for each device group (step S503).

[0148] Specifically, the analysis unit 113 generates a histogram representing the distribution of the maintenance thresholds of the devices for each device group. The analysis unit 113 fits the histogram to a normal distribution and calculates the average value as the group maintenance threshold.

[0149] An image of the calculation is shown in Fig. 16B. Rectangles 1601 in Fig. 16B represent bins of the histogram. Curve 1602 in Fig. 16B represents the result of fitting the histogram with a normal distribution.

[0150] The analysis unit 113 calculates a group abnormality threshold for each device group (step S504). The method of calculating the group abnormality threshold is the same as the method of calculating the group integrity threshold.

[0151] FIG. 16C shows the results of calculating the group integrity threshold and the group anomaly threshold. The dotted rectangular area is the area determined by the group integrity threshold. The white rectangular area is the area determined by the group integrity threshold and the group anomaly threshold. The vertically lined rectangular area is the area determined by the group anomaly threshold. The dotted rectangular area represents the range of the third state, the white rectangular area represents the range of the second state, and the vertically lined rectangular area represents the range of the first state.

[0152] The analysis unit 113 stores the processing result in the analysis result DB 124 (step S505), and ends the analysis processing.

[0153] The analysis result DB 124 stores a table 1700 and a table 1710 .

[0154] The table 1700 stores the security thresholds of the devices. The table 1700 stores entries including a device ID 1701 and a security threshold 1702. There is one entry for one device.

[0155] The device ID 1701 is the same field as the device ID 301. The security threshold 1702 is a field that stores the security threshold of the device calculated in step S501.

[0156] The table 1710 stores entries including a group ID 1711, an impact level 1712, a group integrity threshold 1713, and a group abnormality threshold 1714. There is one entry for one group.

[0157] The group ID 1711 is the same field as the group ID 302. The impact level 1712 is a field for storing the impact level of the group. The group integrity threshold 1713 and the group abnormality threshold 1714 are fields for storing the group integrity threshold and the group abnormality threshold.

[0158] The analysis unit 113 may execute only the process of step S501. In this case, a table 1700 is stored in the analysis result DB .

[0159] Fig. 18 is a flowchart illustrating an example of a presentation process of the decision support information executed by the analysis unit 113 according to the first embodiment. Fig. 19 is a diagram illustrating an example of a screen displayed on the terminal 101 according to the first embodiment.

[0160] When the analysis unit 113 receives a presentation request, the analysis unit 113 executes the process described below. The presentation request includes information that specifies the facility or device.

[0161] The analysis unit 113 calculates the current failure probability of each device (step S601).

[0162] The analysis unit 113 identifies, for each device, the influence of the group to which the device belongs (step S602). Specifically, the analysis unit 113 identifies the group to which the device belongs by referring to the device parameter DB 120. The analysis unit 113 identifies the influence of the group by referring to the analysis result DB 124.

[0163] The analysis unit 113 identifies the state of each device (step S603). Specifically, the analysis unit 113 classifies the state of the device based on the failure probability of the device and the group maintenance threshold and group abnormality threshold of the device group.

[0164] The analysis unit 113 calculates the urgency of each device (step S604). Specifically, the analysis unit 113 calculates the time t′ that satisfies the formula (9) as the urgency. Here, P prevention represents the abnormality threshold of the device. The urgency level represents the time from the current time until the abnormality threshold is reached.

[0165]

number

[0166] The analysis unit 113 calculates the repair efficiency for each device (step S605). Specifically, the analysis unit 113 calculates a value obtained by dividing the failure probability by the maintenance cost 304 as the repair efficiency.

[0167] The analysis unit 113 generates and outputs decision support information for displaying the processing results (step S606), after which the analysis unit 113 ends the decision support information presentation process.

[0168] After the analysis process, the analysis unit 113 may execute a decision support information presentation process. In this case, the process is executed based on the time of the process execution. The analysis unit 113 may also present only the classification result using the maintenance threshold value of each device.

[0169] Based on the decision support information, the terminal 101 displays a screen 1900 as shown in FIG.

[0170] A display area 1901 is an area for displaying the classification results of the device status. The background of the display area 1901 is displayed based on the table 1700. The points in the display area 1901 are plots of failure probabilities.

[0171] By referring to the display area 1901, the user can get an overview of the distribution of devices in each state.

[0172] A display area 1902 is an area that displays the relationship between the urgency and repair efficiency of devices. In the scatter diagram displayed in the display area 1902, devices with high urgency and repair efficiency should be given priority in maintenance.

[0173] By referring to the display area 1902, the user can select a device to be given priority in maintenance.

[0174] The display area 1903 is an area for displaying the status of the device selected in the display area 1901 based on the maintenance threshold and the abnormality threshold.

[0175] By referring to the display area 1903, the user can grasp the detailed status of the device.

[0176] The maintenance decision support system 100 determines the maintenance threshold of an equipment by performing a maintenance simulation that takes into account the budget amount. The maintenance decision support system 100 can also calculate a maintenance threshold for each equipment group (group maintenance threshold) based on the maintenance threshold of the equipment. The maintenance decision support system 100 can classify the equipment status for each equipment group by using the group maintenance threshold and the group abnormality threshold. The maintenance decision support system 100 can also support the decision on the equipment for which maintenance is to be performed by presenting information that allows an overview of the distribution of the equipment status.

[0177] FIG. 20 is a sequence diagram showing a method of using the maintenance decision support system 100 according to the first embodiment.

[0178] The sensor device 104 installed in the facility 102 transmits measurement information to the maintenance decision support system 100. The maintenance decision support system 100 stores the received measurement information in the equipment status DB 122.

[0179] The terminal 101 transmits information related to the maintenance work to the maintenance decision support system 100. The maintenance decision support system 100 stores the received information in the device status DB 122.

[0180] An administrator of the maintenance decision support system 100 or a user operating the terminal 101 requests execution of a threshold calculation process at the timing of a change in a maintenance plan, etc. The maintenance decision support system 100 updates the equipment parameter DB 120, the simulation result DB 123, and the analysis result DB 124 by executing the threshold calculation process.

[0181] The terminal 101 transmits a presentation request to the maintenance decision support system 100. When the maintenance decision support system 100 receives the presentation request, it executes presentation processing of the decision support information and transmits the decision support information to the terminal 101.

[0182] By referring to the maintenance decision support information, maintenance workers can be sure to understand the information that requires maintenance. Furthermore, by referring to the maintenance decision support information, maintenance workers can select the equipment for which maintenance should be performed from among the candidate equipment based on quantitative information. This allows maintenance of equipment to be performed within the budget while suppressing increases in operational risks. In other words, maintenance that reflects the actual conditions at the site becomes possible. This also makes it possible to improve the income and expenditure of the business operator.

[0183] The present invention is not limited to the above-mentioned embodiment, but includes various modified examples. For example, the above-mentioned embodiment describes the configuration in detail to easily explain the present invention, and the present invention is not necessarily limited to the configuration including all the described configurations. Also, it is possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

[0184] In addition, each of the above configurations, functions, processing units, processing means, etc. may be realized in part or in whole by hardware, for example, by designing them as integrated circuits. The present invention can also be realized by software program code that realizes the functions of the embodiments. In this case, a storage medium on which the program code is recorded is provided to a computer, and a processor included in the computer reads the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-mentioned embodiments, and the program code itself and the storage medium storing it constitute the present invention. Examples of storage media for supplying such program code include flexible disks, CD-ROMs, DVD-ROMs, hard disks, SSDs (Solid State Drives), optical disks, magneto-optical disks, CD-Rs, magnetic tapes, non-volatile memory cards, ROMs, etc.

[0185] Furthermore, the program code for realizing the functions described in this embodiment can be implemented in a wide range of program or script languages, such as assembler, C / C++, perl, Shell, PHP, Python, Java (registered trademark), etc.

[0186] Furthermore, the program code of the software that realizes the functions of the embodiments may be distributed over a network and stored in a storage means such as a computer's hard disk or memory, or in a storage medium such as a CD-RW or CD-R, and the processor of the computer may read out and execute the program code stored in the storage means or storage medium.

[0187] In the above-mentioned embodiment, the control lines and information lines are shown as those considered necessary for the explanation, and not all the control lines and information lines are shown in the product. All the components may be connected to each other. [Explanation of symbols]

[0188] 100 Maintenance decision support system 101 Terminal 102 Equipment 103 Network 104 Sensor device 110 Facility model generation unit 111 Maintenance judgment model generation unit 112 Simulation Department 113 Analysis Department 114 Maintenance threshold calculation unit 120 Equipment Parameter DB 121 Facility Schedule DB 122 Equipment Status DB 123 Simulation Results DB 124 Analysis result DB 200 calculator 201 Processor 202 Main storage 203 Auxiliary storage device 204 Communication Equipment 205 Input Device 206 Output Device 1900 screen

Claims

1. A computer system comprising at least one computer, Connect to a facility consisting of at least one device, retaining threshold information for managing thresholds for classifying the state into one of a first state indicating a state in which maintenance is required, a second state indicating a state in which it is necessary to determine whether or not maintenance is performed, and a third state indicating a state in which maintenance is not required, based on a failure probability of the device; the threshold information includes, for each of the devices, a maintenance threshold that specifies a boundary between the first state and the second state, and an abnormality threshold that specifies a boundary between the second state and the third state, The computer system comprises: a first process for determining the maintenance threshold value of each of the devices; and a second process of calculating a failure probability of the equipment, classifying the equipment into one of the first state, the second state, or the third state based on the failure probability of the equipment and the threshold information, and outputting decision support information including a result of the classification.

2. 2. The computer system of claim 1, In the first process, the computer system During a simulation period, a process is executed for each time step to determine the equipment for which maintenance is to be performed based on the failure probability of the equipment and the threshold information, a maintenance simulation is executed to calculate a cost of maintenance during the simulation period, and the cost and the maintenance threshold value of each of the equipment are recorded; determining whether or not the maintenance simulation needs to be performed again based on the cost; If it is necessary to execute the maintenance simulation again, the maintenance threshold value of each of the devices is updated, and the maintenance simulation is executed; A computer system characterized in that, when there is no need to run the maintenance simulation again, an analytical process is executed to determine the maintenance thresholds of each of the equipment based on the history of the maintenance thresholds of each of the equipment used in the maintenance simulation.

3. 3. The computer system of claim 2, In the maintenance simulation, the computer system A computer system characterized by determining whether or not to perform maintenance on the equipment classified into the second state by performing a Monte Carlo tree search using a maintenance decision model that calculates a decision probability for performing maintenance based on the failure probability of the equipment classified into the second state.

4. 4. The computer system of claim 3, The facility is managed in the unit of an equipment group consisting of one or more of the equipment, The computer system, in the analysis process, calculating a group security threshold for each of the device groups based on the security thresholds of the determined devices belonging to the device group; calculating a group abnormality threshold for each of the device groups based on the abnormality thresholds of the devices belonging to the device group; The device group, the group maintenance threshold, and the group abnormality threshold are recorded in association with each other, In the second process, classifying the devices belonging to the device group into one of the first state, the second state, and the third state using the failure probability of the devices, the group maintenance threshold, and the group abnormality threshold for each device group; and outputting the decision support information including the results of the classification.

5. 5. The computer system of claim 4, In the second process, calculating, for each of the device groups, a first index representing a degree of necessity for maintenance of the devices belonging to the device group and a second index representing an effect of the maintenance; A computer system which outputs the decision support information including the first index and the second index.

6. A maintenance decision support method executed by a computer system, comprising: The computer system comprises: At least one computer, Connect to a facility consisting of at least one device, retaining threshold information for managing thresholds for classifying the state into one of a first state indicating a state in which maintenance is required, a second state indicating a state in which it is necessary to determine whether or not maintenance is performed, and a third state indicating a state in which maintenance is not required, based on a failure probability of the device; the threshold information includes, for each of the devices, a maintenance threshold that specifies a boundary between the first state and the second state, and an abnormality threshold that specifies a boundary between the second state and the third state, The conservation decision support method includes: a first step in which the at least one computer determines the maintenance threshold for each of the devices; a second step in which the at least one computer calculates a failure probability of the equipment, classifies the equipment into one of the first state, the second state, or the third state based on the failure probability of the equipment and the threshold information, and outputs decision support information including a result of the classification.

7. The maintenance decision support method according to claim 6, The first step includes: a third step of executing a maintenance simulation in which the at least one computer executes a process for determining the equipment for which maintenance is to be performed based on the failure probability of the equipment and the threshold information for each time step during a simulation period, calculates a maintenance cost during the simulation period, and records the cost and the maintenance threshold value of each of the equipment; a fourth step of determining, by the at least one computer, whether or not it is necessary to execute the maintenance simulation again based on the cost; a fifth step of updating the maintenance threshold value of each of the devices and executing the maintenance simulation when the at least one computer needs to execute the maintenance simulation again; A maintenance decision support method characterized by including a sixth step of the at least one computer executing, if there is no need to run the maintenance simulation again, an analytical process to determine the maintenance threshold of each of the equipment based on the history of the maintenance threshold of each of the equipment used in the maintenance simulation.

8. The maintenance decision support method according to claim 7, The maintenance decision support method, wherein the third step and the fifth step include a step in which the at least one computer determines whether or not to perform maintenance on the equipment classified into the second state by executing a Monte Carlo tree search using a maintenance decision model that calculates a decision probability for performing maintenance based on a failure probability of the equipment classified into the second state.

9. The maintenance decision support method according to claim 8, The facility is managed in the unit of an equipment group consisting of one or more of the equipment, The sixth step includes: the at least one computer calculating, for each of the device groups, a group security threshold based on the security thresholds of the determined devices belonging to the device group; the at least one computer calculating, for each of the device groups, a group abnormality threshold based on the abnormality thresholds of the devices belonging to the device group; and recording, by the at least one computer, the device group, the group maintenance threshold, and the group abnormality threshold in association with each other; The second step includes: the at least one computer classifying, for each of the device groups, the devices belonging to the device group into one of the first state, the second state, or the third state using the failure probability of the devices, the group maintenance threshold, and the group abnormality threshold; and a step of outputting, by the at least one computer, the decision support information including a result of the classification.

10. The maintenance decision support method according to claim 9, The second step includes: a step of calculating, by the at least one computer, a first index representing a degree of necessity for maintenance of the devices belonging to the device group and a second index representing an effect of the maintenance, for each of the device groups; and a step of outputting, by the at least one computer, the judgment support information including the first index and the second index.

11. 1. A system comprising: The present invention comprises a plurality of terminals used by users, at least one facility including at least one device, and a maintenance decision support system, the maintenance decision support system holds threshold information for classifying a state into one of a first state indicating a state in which maintenance is required, a second state indicating a state in which it is necessary to determine whether or not maintenance is performed, and a third state indicating a state in which maintenance is not required, based on a failure probability of the equipment; the threshold information includes, for each of the devices, a maintenance threshold that specifies a boundary between the first state and the second state, and an abnormality threshold that specifies a boundary between the second state and the third state, The maintenance decision support system comprises: Acquiring measurement information indicating a state of the at least one device from the facility; during a simulation period, for each time step, a process is executed to determine the equipment for which maintenance is to be performed based on the failure probability of the equipment and the threshold information, a maintenance simulation is executed to calculate the cost of maintenance during the simulation period and record the cost and the maintenance threshold of each of the equipment, it is determined whether or not the maintenance simulation needs to be executed again based on the cost, if the maintenance simulation needs to be executed again, the maintenance threshold of each of the equipment is updated and the maintenance simulation is executed, and if it is not necessary to execute the maintenance simulation again, an analysis process is executed to determine the maintenance threshold of each of the equipment based on the history of the maintenance threshold of each of the equipment used in the maintenance simulation, A system characterized by calculating a failure probability of the equipment, classifying the equipment into one of the first state, the second state, or the third state based on the failure probability of the equipment and the threshold information, and transmitting decision support information including a result of the classification to the terminal.

Images