JP7690406B2 - Maintenance improvement support device and maintenance improvement support method - Google Patents

Description

The present invention relates to a maintenance improvement support device and a maintenance improvement support method that present improvement measures for maintenance work.

In many fields, including infrastructure, railways, industrial equipment, and medical equipment, maintenance such as repairs and regular inspections is continuously performed after assets such as facilities and equipment are installed. Continuing maintenance allows assets to maintain their specified functions and performance. Maintenance requires analyzing the condition of the target asset and the status of maintenance execution, and planning and executing appropriate maintenance work. Furthermore, to ensure that maintenance work is carried out as planned in the actual business, designs and plans must include the maintenance organization and maintenance equipment.

However, in the maintenance business, various maintenance tasks must be performed on numerous and diverse assets, or on assets located over a wide area, so designing the appropriate size and layout of the maintenance organization is not easy. In particular, in recent years, it has become necessary to introduce maintenance processes that are different from the conventional ones, taking into account the effects and operation of IoT (Internet of Things) technology. As a result, there have been major changes in the number of personnel, required skills, tools, and other maintenance work resources, and the complexity of maintenance work design is increasing.

If proper maintenance work design is not planned, necessary work may not be performed due to a shortage of workers, inappropriate maintenance intervals, or inappropriate diagnosis of asset conditions, which could result in safety or operational problems. Alternatively, excessive maintenance could increase costs, reducing business profitability and ultimately making it impossible to continue using assets. For this reason, there is a need for technology to support appropriate maintenance work design.

Examples of technologies that support the creation of business improvement plans for asset operation and maintenance and the estimation of the effects are described in Patent Documents 1 and 2. Patent Document 1 describes a plant operation management support system that uses past operation data to simulate future operating costs and other factors to create an asset introduction plan for a plant and uses the simulation results in the creation of the plant introduction plan. Patent Document 2 describes a plant lifetime cost management support system that predicts operating costs and maintenance costs based on a combination of operating conditions and operators for operating assets in a plant and presents the optimal combination.

JP 2005-258816 A JP 2006-244288 A

In designing maintenance operations, it is necessary to determine the specific number of personnel in the maintenance organization, the items and details of the maintenance work, the content and scale of the IoT implementation, etc., at an appropriate scale that ensures profitability while ensuring that maintenance work is carried out reliably. However, in maintenance work, there is a high degree of contingency in the occurrence of breakdowns and delays during maintenance work, and the effect of the introduced IoT is determined by the balance between many tasks, so it is difficult to actually design appropriate operations.
Therefore, it is possible to build a simulation model of the current maintenance work to enable the reproduction of the business, and then incorporate future business plans and IoT introduction plans into the simulation model to verify whether the business is designed appropriately. In order to simulate maintenance work, settings are created to reproduce various elements such as the assets that are the subject of the work, the workers in charge of the work, and IoT. In addition, the relationships between assets, work contents, and workers must be set one by one, which makes it difficult and time-consuming to create the settings.

One possible approach would be to generate simulation settings from existing maintenance work (maintenance task) logs (history) and equipment and personnel ledgers. However, because maintenance work logs contain missing records or the recording period is short, there are cases where the simulation settings cannot be generated properly. This can lead to problems such as the risk of an inaccurate simulation being performed or a large number of assets being extracted, resulting in a huge amount of calculations, even though the individual maintenance tasks are small. For this reason, it is desirable to extract and improve the missing and inefficient simulation settings and set them up appropriately.

The present invention was made in light of this background, and aims to provide a maintenance improvement support device and a maintenance improvement support method that can appropriately simulate maintenance work.

In order to solve the above-mentioned problems, a maintenance improvement support device according to the present invention includes an asset that is a target of maintenance work in a maintenance work, a worker in charge of the maintenance work, and a task that is the maintenance work performed by the worker on the asset, and includes a simulation setting generation unit that generates setting information for simulating the maintenance work based on a maintenance work log database that is a record of the execution of the maintenance work, a simulation unit that executes a simulation of the maintenance work using an asset agent that is an agent representing the asset, and a worker agent that is an agent representing the worker, and a reproducibility improvement processing unit that detects a state of the asset agent or the worker agent that violates a rule in a work knowledge database that includes rules for performing the maintenance work , and corrects the setting information, and the simulation setting generation unit generates task setting information from the maintenance work log database that is setting information that associates the asset, a pattern of failures that occur in the asset, and the task performed by the worker corresponding to the failure. and generates, from the maintenance work log database, work responsibility setting information which is setting information associating the worker, the asset which is the target of the task performed by the worker, and the task, and the simulation unit executes a simulation in which a failure occurs in the asset agent according to a failure pattern included in the task setting information, and an allocation of a task corresponding to the failure is determined to an asset included in the work responsibility setting information which corresponds to the asset agent and a worker agent which corresponds to the task and is associated with the task, and the worker agent performs the task on the asset agent, and the reproducibility improvement processing unit includes an excessive work elimination processing unit which, if the activity amount of the worker agent during the simulation is greater than a specified value contrary to a specification in the business knowledge database, adds the task performed by the worker corresponding to the worker agent included in the work responsibility setting information and the asset which is the target of the task to the work responsibility setting information of a worker different from the worker and who has an activity amount less than the specified value.
A maintenance improvement support device according to the present invention includes an asset that is a target of maintenance in a maintenance operation, a worker in charge of the maintenance operation, and a task that is the maintenance operation performed by the worker on the asset, and includes a simulation setting generation unit that generates setting information for simulating the maintenance operation based on a maintenance operation log database that is a record of the execution of the maintenance operation; a simulation unit that executes a simulation of the maintenance operation using an asset agent that is an agent representing the asset, and a worker agent that is an agent representing the worker; and a reproducibility improvement processing unit that detects a state of the asset agent or the worker agent that violates a rule in a task knowledge database that includes rules for executing the maintenance operation, and corrects the setting information. The simulation setting generation unit generates task setting information from the maintenance operation log database, which is setting information that associates the asset, a pattern of failures that occur in the asset, and the task performed by the worker corresponding to the failure, and calculates task setting information from the maintenance operation log database. The simulation unit generates work responsibility setting information which is setting information associating the worker, the asset which is the target of the task performed by the worker, and the task, and the simulation unit determines an assignment of a task corresponding to the failure to an asset included in the work responsibility setting information which corresponds to the asset agent and a worker agent which corresponds to the task, and executes a simulation in which the worker agent performs the task for the asset agent, in accordance with a failure pattern included in the task setting information, and determines an assignment of the task corresponding to the failure to an asset included in the work responsibility setting information which corresponds to the asset agent and a worker agent which corresponds to the task, and executes a simulation in which the worker agent performs the task for the asset agent, and the reproducibility improvement processing unit includes a work deficiency response processing unit which, when a period during which a corresponding task is not performed after a failure in the task setting information for the asset agent during the simulation occurs is greater than a specified period in the business knowledge database, searches for an agent which represents the worker who will perform the task, and adds the task and the asset which corresponds to the asset agent to the work responsibility setting information of the worker corresponding to the agent which represents the worker.

The present invention provides a maintenance improvement support device and a maintenance improvement support method that appropriately simulate maintenance work. Problems, configurations, and effects other than those described above will become clear from the description of the following embodiments.

1 is a functional block diagram of a maintenance improvement support device according to an embodiment of the present invention. FIG. 2 is a data configuration diagram of a maintenance work log database according to the embodiment. FIG. 2 is a data configuration diagram of a business knowledge database according to the embodiment. FIG. 2 is a data configuration diagram of a plan database according to the embodiment. 11 is a flowchart of a simulation setting generation process according to the present embodiment. FIG. 4 is a screen configuration diagram of a simulation setting screen according to the embodiment. 4 is a data configuration diagram of asset setting information according to the embodiment. FIG. 4 is a data configuration diagram of task setting information according to the embodiment. FIG. FIG. 4 is a data configuration diagram of worker setting information according to the embodiment. FIG. 4 is a data configuration diagram of worker setting information according to the embodiment. 5 is a diagram showing a data configuration of agent data according to the embodiment; FIG. 13 is a flowchart of a process for dealing with a worker vacancy according to the present embodiment. 11 is a flowchart of an agent merging process according to the present embodiment. 13 is a flowchart of an excessive work resolution process according to the present embodiment.

<Overview of the maintenance improvement support device>
A maintenance improvement support device in a form (embodiment) for carrying out the present invention will be described below. The maintenance improvement support device generates setting information required for executing a simulation of a maintenance work from a maintenance work log database that is a history of maintenance work (maintenance work). The setting information includes asset setting information, task setting information, worker setting information, and work-person setting information. Next, the maintenance improvement support device executes an agent-based simulation (multi-agent simulation) based on the setting information.

The maintenance work log database does not necessarily contain all information related to real assets and workers, and the setting information contains missing or incomplete information. While executing a simulation, the maintenance improvement support device detects conditions that violate maintenance work regulations, such as a long period of asset failure or workers being concentrated on work, and corrects the setting information. By using the corrected setting information, the maintenance work simulation can be executed appropriately. In addition, the corrections to the setting information include an increase in the number of workers and maintenance work (tasks) that workers will need to be responsible for in the future, and can be used as reference information for planning maintenance work.

<Configuration of the maintenance improvement support device>
1 is a functional block diagram of a maintenance improvement support device 100 according to this embodiment. The maintenance improvement support device 100 is a computer, and includes a control unit 110, a storage unit 120, and an input/output unit 180. User interface devices such as a display, a keyboard, and a mouse are connected to the input/output unit 180. The input/output unit 180 may include a communication device and may be capable of transmitting and receiving data to and from other devices. A media drive may also be connected to the input/output unit 180, and data may be exchanged using a recording medium.

≪Maintenance improvement support device: Storage section≫
The storage unit 120 includes storage devices such as a read only memory (ROM), a random access memory (RAM), a solid state drive (SSD), etc. The storage unit 120 stores a maintenance work log database 130, a work knowledge database 140, a plan database 150, agent data 160, setting information 300, and a program 128.

The program 128 includes a description of the procedures for the process of dealing with a vacancy in charge of work (see FIG. 12), the process of combining agents (see FIG. 13), and the process of eliminating excessive work (see FIG. 14). The agent data 160 and the setting information 300 will be described later together with the control unit 110. The maintenance work log database 130, the work knowledge database 140, and the plan database 150 will be described below.

<Memory section: Maintenance work log database>
2 is a data configuration diagram of the maintenance work log database 130 according to this embodiment. The maintenance work log database 130 records maintenance work (maintenance work) such as troubleshooting and regular inspections performed on assets such as facilities and equipment. The maintenance work log database 130 is, for example, data in a table format, where one row (record) indicates a maintenance work (task) and includes columns (attributes) of identification information (referred to as ID in FIG. 2), work type, task, asset, model, location, worker, base, start date and time, and end date and time.

Identification information is an identification number assigned to each individual maintenance task. Work type is the type of maintenance task. Task is the name of the maintenance task and is a detail of the work type. Asset is the identification information of the asset that is the subject of the maintenance task, model is the model of the asset, and location is the location where the asset is installed. Worker is the identification information of the worker who performed the maintenance task, and base is the base to which the worker belongs. Start date and time is the start date and time of the maintenance task, and end date and time is the end date and time of the maintenance task.

The two records with the identification number "2" indicate that "Worker 2" and "Worker 3" based in "Office A" performed maintenance work (task) identified as "Failure Response 2" on an asset located at "Site A" whose model is "ATM-A" and whose identification number is "ATM2" between 6:00 p.m. and 8:00 p.m. on January 1, 2020.

<Memory section: Business knowledge database>
3 is a data configuration diagram of the business knowledge database 140 according to this embodiment. The business knowledge database 140 stores rules and policies for carrying out maintenance work, and information (regulations) for judging the status of assets and workers.
The simulation of the maintenance work is executed so as to satisfy the regulations in the work knowledge database 140. In other words, if a state that does not satisfy the regulations occurs during the execution of the simulation, a process is started to resolve the state (see FIGS. 12 to 14 described later).

<<Memory section: Plan database>>
4 is a data configuration diagram of the plan database 150 according to this embodiment. The plan database 150 includes plans for adding and discarding assets in the future, plans for increasing and decreasing the number of workers, and plans for changing the maintenance work that workers are responsible for.

<Maintenance improvement support device: control unit>
Returning to Fig. 1, the explanation of the control unit 110 will continue. The control unit 110 is configured to include a CPU (Central Processing Unit) and has a simulation setting generation unit 111, an agent generation unit 112, a simulation unit 113, a worker vacancy response processing unit 114, an agent combination processing unit 115, and an excessive work resolution processing unit 116. Each of these will be explained in order below.

In this embodiment, a multi-agent simulation is used as a method for simulating maintenance work. In detail, agents are generated that individually simulate the behavior of objects that exist in the real world, such as assets and workers. Each agent performs its own specified behavior while evolving over time in relation to other agents to reproduce the maintenance work.

<Maintenance improvement support device: Control unit: Simulation setting generation unit>
The simulation setting generation unit 111 generates setting information 300 (see FIG. 1 ) required for executing a simulation of a maintenance operation by referring to the maintenance operation log database 130. The setting information 300 includes asset setting information 320 (see FIG. 7 ), task setting information 330 (see FIG. 8 ), worker setting information 340 (see FIG. 9 ), and worker setting information 350 (see FIG. 10 ), which will be described later.

5 is a flowchart of the simulation setting generation process according to this embodiment. The process of the simulation setting generation unit 111 will be described with reference to FIG.
In step S11, the simulation setting generation unit 111 displays a simulation setting screen 310 (see FIG. 6 described later) on a display connected to the input/output unit 180, and obtains from the user the range of the maintenance work log database 130 to be referenced and the range of the simulation.

6 is a diagram showing the screen configuration of a simulation setting screen 310 according to this embodiment. In an area 311, a period for which the maintenance work log database 130 is referenced is specified. In an area 312, a type of asset to be simulated is specified. In an area 313, a region in which the asset to be simulated is installed is specified. In an area 314, a period for which the simulation is to be performed is specified. When the "Start" button is pressed, the simulation begins after the setting information 300 is generated.
6, the type of asset is specified in the area 312, but the identification information of the individual asset may be used. Also, instead of the area specified in the area 313, the name of the worker or the company to which the worker belongs (the name of the maintenance organization) may be specified.

Returning to FIG. 5, in step S12, the simulation setting generation unit 111 accesses the maintenance work log database 130 (see FIG. 2) and obtains records (maintenance work records, maintenance work history) for the period specified in area 311 (see FIG. 6).

In step S13, the simulation setting generation unit 111 generates asset setting information 320 (see FIG. 7 described later). FIG. 7 is a data structure diagram of asset setting information 320 according to this embodiment. The asset setting information 320 is, for example, data in a tabular format, where one row (record) indicates an asset and includes columns (attributes) of asset identification information, model, and installation location. The simulation setting generation unit 111 extracts the attributes of the asset, model, and location from the maintenance work log database 130 (see FIG. 2) to generate asset setting information 320.

Returning to FIG. 5, in step S14, the simulation setting generation unit 111 generates task setting information 330 (see FIG. 8 described later). FIG. 8 is a data configuration diagram of task setting information 330 according to this embodiment. Task setting information 330 is, for example, data in a tabular format, where one row (record) indicates a task (maintenance work), and includes columns (attributes) of the model of the asset to be maintained, the work type, the task (task name), the occurrence pattern, and the work time.

The simulation setting generation unit 111 extracts the attributes of the model, work type, and task from the maintenance work log database 130 (see FIG. 2) and generates the model, work type, and task attributes of the task setting information 330. The simulation setting generation unit 111 also extracts the attributes of the start date and time and the end date and time from the maintenance work log database 130 to calculate the work time, and generates the work time attributes of the task setting information 330.

The simulation setting generation unit 111 generates a task occurrence pattern by aggregating the frequency of tasks for each asset. When looking at individual assets, for example, it may be difficult to accurately estimate the occurrence pattern of a task with a low occurrence rate due to a low failure rate, for example, because there are not enough samples within the period of the maintenance work record being referenced. For this reason, the simulation setting generation unit 111 may obtain an accurate pattern by taking aggregators for each asset model or installation area category. The simulation setting generation unit 111 may also obtain an accurate pattern by accessing an information source related to assets other than the maintenance work log database 130 and taking aggregators for each user category, asset usage period, and the like.

As an occurrence pattern, for example, the frequency of occurrence of a task marked as "Failure response 1" in the maintenance work record is counted and set as a random task occurrence. For example, the first record indicates that for an asset whose model is "ATM-A," the task "Failure response 1" occurs on average once every 2000 hours, indicating that the mean time between failures in question is 2000 hours. In addition, in the case of regular inspections, the period is the average inspection cycle. Note that if the frequency of occurrence of failures or the regular inspection cycle changes depending on the period (number of years) since installation, an occurrence pattern according to the number of years since installation may be included.

Returning to FIG. 5, in step S15, the simulation setting generation unit 111 generates worker setting information 340 (see FIG. 9 described later). FIG. 9 is a data configuration diagram of the worker setting information 340 according to this embodiment. The worker setting information 340 is, for example, data in a table format, where one row (record) indicates a worker and includes identification information of the worker and a column (attributes) of the base. The simulation setting generation unit 111 extracts the attributes of the worker and base from the maintenance work log database 130 to generate the worker setting information 340.

Returning to FIG. 5, in step S16, the simulation setting generation unit 111 generates worker setting information 350 (see FIG. 10 described later). FIG. 10 is a data configuration diagram of worker setting information 350 according to this embodiment. Worker setting information 350 is, for example, data in a table format, where one row (record) indicates a task (maintenance work) for which a worker is responsible, and includes columns (attributes) of identification information for the worker, assets, and tasks. The simulation setting generation unit 111 extracts the attributes of workers, assets, and tasks from the maintenance work log database 130 to generate worker setting information 350.

<Maintenance improvement support device: control unit: agent generation unit>
Returning to FIG. 1, the agent generation unit 112 generates agents 200 (see FIG. 11 described later) for each asset and each worker in the agent data 160 in accordance with the setting information 300 generated by the simulation setting generation unit 111 .

Figure 11 is a diagram showing the data configuration of the agent data 160 according to this embodiment. The agent generation unit 112 generates an asset agent 210 corresponding to each asset shown in the asset setting information 320 (see Figure 7). The agent generation unit 112 also generates a worker agent 220 corresponding to each worker shown in the worker setting information 340 (see Figure 9). The agent generation unit 112 also generates a work plan agent 230 that assigns tasks to workers, and a regular inspection work generation agent 240 that generates regular inspection work requests for each asset. The asset agent 210, worker agent 220, work plan agent 230, and regular inspection work generation agent 240 are collectively referred to as agent 200. Details of the configuration of the agent 200 will be described later.

The agent generation unit 112 adds, deletes, or modifies asset agents 210 and worker agents 220 in accordance with plans for adding/deleting assets and workers, or changing responsibilities, in the plan database 150 (see FIG. 4). This is performed at the occurrence date and time in the plan database 150, based on the time within the simulation in the simulation unit 113, which will be described later. Note that the addition, deletion, or modification of asset agents 210 and worker agents 220 after the start of simulation execution may also be performed by the simulation unit 113.

<Maintenance improvement support device: Control unit: Simulation unit>
The simulation unit 113 operates the agents 200 in the agent data 160 generated by the agent generation unit 112 to execute a simulation of a maintenance operation. The simulation unit 113 executes a time update process for each agent 200. Driven by this, processes such as updating the internal state of the agent 200 and interactions between the agents 200 are executed.

Although the agent 200 is data stored in the agent data 160, it can be operated by the simulation unit 113 to change its internal state and act on other agents 200. For this reason, the process performed by the simulation unit 113 may be described as a process performed primarily by the agent 200. For example, "the asset agent 210 requests a task to the work plan agent 230 to handle a failure" or "the asset agent 210 status update unit 211 updates its internal state." This description is similar to the work person vacancy response processing unit 114, the agent merging processing unit 115, and the excessive work resolution processing unit 116 described below, and the processes performed by these functional units may be described as processes performed by the agent 200. In addition, the assets and asset agents 210 included in the setting information 300 may be described as being the same, and the workers and worker agents 220 may be described as being the same.

<Simulation Department: Asset Agent>
The configuration and operation of the agent 200 (see FIG. 11) differs depending on the type, such as the asset agent 210 or the worker agent 220. The state update unit 211 of the asset agent 210 is set with a fault response task for each asset, and performs a process of generating the task probabilistically. For example, in the asset agent 210 corresponding to the model "ATM-A", a "Fault Response 1" task randomly occurs once every 2000 hours in the simulation time in the simulation unit 113 (see FIG. 8). In addition, at the same time as the task occurs, it is considered that a fault has occurred, and the state update unit 211 updates the internal state to "fault state".

The interaction unit 212 of the asset agent 210 notifies the work plan agent 230 of a response request for a task generated by the status update unit 211. Furthermore, when a worker agent 220 described below executes a task for the asset agent 210, the interaction unit 212 notifies the status update unit 211. The status update unit 211 then determines that the failure response task has been executed and that the asset is now operational, and updates the internal status to "normal state." In this way, the interaction unit 212 performs a cooperative process with other agents 200. The status monitoring unit 213 and reproducibility improvement unit 214 of the asset agent 210 will be described later.
In the above explanation, it was described that the asset agent 210 is the main body, with the status update unit 211 and the interaction unit 212 functioning, but in processing on the maintenance improvement support device 100, the simulation unit 113 accesses the data of the asset agent 210 and performs processing.

<Simulation part: Periodic inspection task generation agent>
The regular inspection work generation agent 240 performs processing to generate regular inspection tasks for each asset agent 210. In detail, the regular inspection work generation agent 240 performs processing to periodically generate regular inspection tasks for each asset agent 210 by referring to task setting information 330 (see FIG. 8 ). Next, when the regular inspection work generation agent 240 notifies the asset agent 210 of the generation of a regular inspection task, the status update unit 211 of the asset agent 210 updates the status to "waiting for regular inspection". The regular inspection work generation agent 240 also notifies the work plan agent 230 of a request to respond to the regular inspection task.

<Simulation Department: Work Planning Agent>
Based on requests for failure response tasks from the asset agent 210 and requests for regular inspection tasks from the regular inspection work generation agent 240, the work planning agent 230 assigns tasks to worker agents 220, taking into consideration constraints such as the number of people required for each task, the workers or skills in charge, and the start and end dates of execution, as well as tasks that have already been assigned to workers, and notifies the worker agents 220 of the assigned tasks.

<Simulation Department: Worker Agent>
When the state update unit 221 of the worker agent 220 receives a task notification from the work plan agent 230, it stores the task as worker information. The state update unit 221 is in a "standby/rest state" during times when no tasks are scheduled to be executed, such as failure response tasks or regular inspection tasks. The state update unit 221 transitions to a "task execution state" when executing an assigned task. In the "task execution state," it executes a task for the asset agent 210 via the interaction unit 222. The state update unit 211 of the asset agent 210 then updates the internal state from a "failure state" or "waiting for regular inspection state" to a "normal state." The state monitoring unit 223 and reproducibility improvement unit 224 of the worker agent 220 will be described later.

As explained above, a failure response task occurs in the asset agent 210, and the work planning agent 230 is notified. A regular inspection task also occurs in the regular inspection work generation agent 240, and the asset agent 210 and the work planning agent 230 are notified. The work planning agent 230 assigns the task to the worker agent 220, and the assigned worker agent 220 executes the task on the asset agent, thereby completing the failure response task and regular inspection task. The status update unit 211 of the asset agent 210 returns to the "normal state". In this way, the agent 200 operates (or more precisely, the simulation unit 113 operates the agent 200), thereby executing a simulation of the maintenance work.

In other words, the simulation performed by the simulation unit 113 proceeds as follows: (1) A failure occurs in the asset agent 210 according to the failure pattern included in the task setting information 330. (2) The assignment of the task corresponding to this failure to the asset agent 210 (asset) and the worker agent 220 (worker) set in the worker setting information 350 to be in charge of the task is determined. (3) The worker agent 220 performs the task for the asset agent 210.

However, there are several problems that must be solved to perform a highly accurate simulation without simulation defects. The main problems are those caused by missing simulation setting information 300, those related to calculation efficiency such as the amount of calculation, and excessive concentration of work on workers. Below, we will explain the process of improving reproducibility to solve these problems, including a process for dealing with missing workers, a process for combining agents, and a process for eliminating excessive work.

<Maintenance improvement support device: Control unit: Worker missing response processing unit>
The first problem is that there may be a deficiency in the worker setting information 350 (see FIG. 10). In that case, even if the asset agent 210 issues a request for a failure response task, the work plan agent 230 cannot assign a worker agent 220, and the task is not executed. In response to such a deficiency in the worker setting information 350, the worker deficiency response processing unit 114 executes a worker deficiency response process to deal with the deficiency.

12 is a flowchart of the worker vacancy response process according to this embodiment. The worker vacancy response processing unit 114 (reproducibility improvement processing unit) operates the agent 200 in parallel with the simulation unit 113 to execute the worker vacancy response process.
In step S21, the status monitoring unit 213 of the asset agent 210 monitors the internal status of the status update unit 211, and if the "failure status" or "waiting for regular inspection status" is not updated for a certain period of time (step S21→YES), the process proceeds to step S22. The internal status of the status update unit 211 is not updated in this manner when there is a defect in the worker setting information 350 (see FIG. 10) and a failure response task or regular inspection task is not executed. The certain period of time is specified as the "maximum waiting time for task execution" in the business knowledge database 140 (see FIG. 3).

In the following processing, the asset agent 210 and the worker agent 220 exchange requests and scores via the interaction units 212, 222. In the agent merging processing and excessive work resolution processing described below, the asset agent 210 and the worker agent 220 also exchange with other agents 200 via the interaction units 212, 222.
In step S22, the reproducibility improvement unit 214 of the asset agent 210 inquires of all worker agents 220 as to whether or not they are able to handle the failure response tasks and regular inspection tasks of the asset agent 210 itself.

In step S23, the reproducibility improvement unit 224 of each worker agent 220 calculates a score of the possibility of taking on the inquired task and responds to the asset agent 210. As a method of calculating the score, for example, if the tasks that the worker agent 220 has already taken on include a task of the same type as the inquired task, a high score is given. Also, if the location of the asset agent 210 (see the location of the asset setting information 320 in FIG. 7) is close to the base of the worker agent 220 (see the base of the worker setting information 340 in FIG. 9), a high score is given. In addition, the smaller the working time for the current month calculated by the status monitoring unit 223 of the worker agent 220 is than the standard working time per month in the business knowledge database 140 (see FIG. 3), the higher the score is considered to be indicative of the amount of spare capacity. Also, if the working time for the current month exceeds the sum of the standard working time per month in the business knowledge database 140 and the overtime hours per month, the lowest score is given.

In step S24, the reproducibility improvement unit 214 of the asset agent 210 assigns the task to a worker agent 220 that has a score higher than a certain value among the received scores.
In step S25, the reproducibility improvement unit 224 of the worker agent 220 adds a task to the worker information stored therein and the worker setting information 350 (see FIG. 10). This additional correction eliminates the loss of the worker setting information 350.

The fixed score value in step S24 is a preset value, but may be calculated from the scores for tasks that the worker agent 220 has already handled (tasks in the worker's work assignment setting information 350). In detail, in step S23, the reproducibility improvement unit 224 calculates and responds with scores for tasks that the asset agent 210 has already handled, in addition to the tasks in the inquiry. In step S24, the reproducibility improvement unit 214 sets the fixed value to the average value or a fixed ratio of the scores for the received tasks that have already been handled.

<Features of the Worker Loss Response Department>
As described above, when the period during which a task is not performed after a failure occurs in the task setting information 330 for the asset agent 210 during simulation is longer than the period stipulated in the business knowledge database 140, the worker absence response processing unit 114 searches for the worker agent 220 who will perform the task, and adds the task and the asset to the worker setting information 350. The reproducibility improvement units 214, 224 (more precisely, the worker absence response processing unit 114) detect the absence of the worker setting information 350, and correct the worker setting information 350.

In this way, the maintenance improvement support device 100 is equipped with a worker vacancy response processing unit 114 (reproducibility improvement processing unit) that detects a state of the asset agent 210 that violates the rules in the business knowledge database 140, which contains rules for performing maintenance work, and corrects the setting information (worker setting information 350). In this case, violating the rules in the business knowledge database 140 (see Figure 3) means that the period after the occurrence of a failure during which the corresponding task is not performed is longer than the "maximum task execution waiting time."

By using this method, even if assets and workers are frequently changed in the settings of the planning database 150, or if the quality of the maintenance work log database 130 is low and the generated work assignment setting information has many gaps and is difficult to correct, the reproducibility improvement units 214, 224 can improve the work assignment setting information 350 during simulation execution, making it possible to realize a simulation without defects.

<Maintenance improvement support device: control unit: agent combination processing unit>
The second problem occurs when the number of assets included in the asset setting information 320 (see FIG. 7) generated by the simulation setting generation unit 111 is very large. When the number of assets is large, the amount of calculation for the simulation becomes large, making it difficult to execute the simulation in a practical time and memory consumption. Such a second problem occurs in a situation where there are a large number of assets with a very low failure rate. For example, electronic terminals such as network devices and tablet devices are installed in large numbers to cover an entire office or factory, or distributed to individuals, resulting in a large number of units. Although there are many of them compared to precision machinery such as factory manufacturing equipment and ATMs, their failure rates are low, so there are few failure responses and no regular inspections are performed.

For such a large amount of assets, the agent merging processing unit 115 performs an agent merging process to reduce the number of asset agents 210. The agent merging processing unit 115 performs an agent merging process in the asset agent 210, which merges multiple asset agents 210 that are determined to be of the same type and replaces them with a single asset agent 210. This process can reduce the amount of calculation and memory consumed by the asset agent 210.

13 is a flowchart of the agent combination process according to this embodiment. The agent combination processing unit 115 (reproducibility improvement processing unit) operates the agents 200 in parallel with the simulation unit 113 to execute the agent combination process.
In step S31, the status monitoring unit 213 of the asset agent 210 monitors the status update unit 211 and calculates the total (activity amount) of the frequency of execution of failure response tasks and regular inspection tasks by the worker agent 220, and the frequency of entering a "failure state" or a "waiting for regular inspection state." If the calculated frequency is lower than the low task execution frequency threshold defined in the business knowledge database 140 (see FIG. 3) (step S31→YES), the status monitoring unit 213 proceeds to step S32. This state occurs because maintenance work is rarely performed on the asset agent 210.

In step S32, the reproducibility improvement unit 214 of the asset agent 210 sends a notification to the other asset agents 210 that it desires to undergo merging processing. If the other asset agents also desire to undergo merging processing at the same time, they will have similarly notified their desire. The notification may include identification information of the asset agent 210 and the mode (state) contained in the state update unit 211.
In step S33, if the reproducibility improvement unit 214 of the asset agent 210 receives a notification from another asset agent 210 (step S33→YES), it proceeds to step S34, and if no notification is received (step S33→NO), it terminates the agent merging process.

In step S34, the reproducibility improvement unit 214 judges whether or not merging is possible. The criteria for this judgment include that the asset agents 210 are the same model (see the model in FIG. 7) or that the modes (states) included in the state update unit 211 are the same. In addition, if the asset agents 210 are installed in the same location (see the location in FIG. 7) or are geographically close, it may be judged that merging is possible.
If the reproducibility improving unit 214 determines that the agent combination is possible (step S34→YES), the process proceeds to step S35. If the agent combination is not possible (step S34→NO), the reproducibility improving unit 214 ends the agent combination process.

In step S35, the reproducibility improvement unit 214 merges with an asset agent 210 that can be merged. In detail, the reproducibility improvement unit 214 inherits the mode in the state update unit 211 of the asset agent 210 to be merged with itself, and generates a new asset agent 210 with the total failure occurrence rate and regular inspection requirements before the merger. The reproducibility improvement unit 214 also changes the asset setting information 320 (see FIG. 7). For example, if the models are the same, the failure occurrence frequency of the failure mode may be changed to a total value (harmonic mean of the mean time between failures) before the merger. The reproducibility improvement unit 214 also rewrites the information related to the asset before the merger in the worker setting information 350 (see FIG. 10) to the new asset agent 210.

<Features of agent fusion processing>
The agent merging process reduces the amount of calculations and memory usage because the asset agents 210 to be merged become a single asset agent 210. This process is repeatedly executed as long as the total frequency of task execution and the frequency of "fault state" and "regular inspection task" is lower than the low task execution frequency threshold. As a result, the number of asset agents 210 is reduced to the minimum, reducing the amount of calculations and memory usage.

As described above, the maintenance improvement support device 100 is equipped with an agent combination processing unit 115 (reproducibility improvement unit 214, reproducibility improvement processing unit) that detects the state of an asset agent that violates the regulations in the business knowledge database, which includes regulations for performing maintenance work, and corrects the setting information (asset setting information 320, worker setting information 350). In this case, violating the regulations in the business knowledge database (see Figure 3) means that the total frequency of execution of failure response tasks and regular inspection tasks, and the frequency of entering a "failure state" or "waiting for regular inspection state" is lower than the "low task execution frequency threshold" defined in the business knowledge database 140 (see Figure 3).

<Maintenance improvement support device: Control unit: Excessive work elimination processing unit>
The third problem is that the number of worker agents 220 generated is too small, or tasks are concentrated on a small number of worker agents 220, resulting in an abnormally long task execution time for the worker agents 220 (excessive work). This problem occurs because the maintenance work log database 130 (see FIG. 2) is data for a finite period, so even if ten people are actually in charge of maintenance work on an asset, for example, less than half of the work may be recorded during that period. In general, there are legal and physical limits on the amount of work that a worker can perform, so a simulation that reproduces abnormally long work times is not accurate, and there are problems with using it for business planning.

In the case of such a small number of workers, the excessive work resolution processing unit 116 executes excessive work resolution processing to deal with the situation. The excessive work resolution processing unit 116 adds worker agents 220 to resolve the abnormally long task execution times.

14 is a flowchart of the excessive work resolution process according to this embodiment. The excessive work resolution process unit 116 (reproducibility improvement process unit) operates the agent 200 in parallel with the simulation unit 113 to execute the excessive work resolution process.
In step S41, the state monitoring unit 223 of the worker agent 220 monitors the state update unit 221, calculates the time in the "task execution state", and if it is excessive work (step S41->YES), proceeds to step S42. One method for detecting excessive work from statistics is, for example, to calculate the amount of overtime work per month by calculating the time outside of working hours that is considered to be overtime from the task execution time, and if it exceeds the upper limit recorded in the task knowledge database 140 (see FIG. 3), it is determined to be excessive work.

In step S42, the reproducibility improvement unit 224 of the worker agent 220 inquires of the other worker agents 220 whether the task in question can be delegated. The inquiry may include the name and type of the task, the model of the asset agent 210 that is the target of the task, etc.

The repeatability improvement unit 224 of the worker agent 220 that received the inquiry in step S43 responds with a score indicating the possibility of transferring the work. One method of calculating the score is to increase the score the greater the difference between the monthly working hours and overtime hours of the worker agent 220 and the upper limit recorded in the business knowledge database 140. In addition, the closer the base to the inquired worker agent 220, the higher the score may be. Alternatively, the higher the similarity between the model and task type of the target asset agent 210, the higher the score returned. Furthermore, if the working hours for the current month exceed the sum of the monthly standard working hours and monthly overtime hours in the business knowledge database 140, the lowest score is set.

The reproducibility improvement unit 224 of the worker agent 220 inquired in step S44 judges whether the transfer is possible. More specifically, if the returned scores are none of the predetermined values, the reproducibility improvement unit 224 judges that the transfer is not possible (step S44→NO) and proceeds to step S45. If the returned scores are equal to or greater than the predetermined value, the reproducibility improvement unit 224 judges that the transfer is possible (step S44→YES) and proceeds to step S46.
The reproducibility improving unit 224 of the worker agent 220 inquired in step S45 generates a new worker agent 220. Next, the reproducibility improving unit 224 makes the worker setting information 350 of this new worker agent 220 have all or a part of its own worker setting information 350.

The reproducibility improvement unit 224 of the worker agent 220 inquired in step S46 determines the worker setting information 350 (see FIG. 10) to be added to the worker agent 220 with the highest received score (hereinafter referred to as the transfer candidate). In detail, the reproducibility improvement unit 224 compares its own worker setting information 350 with the worker setting information 350 of the transfer candidate, and if there are assets and tasks that the transfer candidate is not responsible for, determines to add to the transfer candidate the task with the largest amount of work in its own worker setting information 350 (the task setting information 330 shown in FIG. 8 has the longest work time) or a randomly selected task in the worker setting information 350.

The reproducibility improvement unit 224 of the worker agent 220 that was inquired in step S47 requests the worker agent 220 that is the transfer destination candidate to add a task.
In step S48, the reproducibility improvement unit 224 of the worker agent 220 that is the transfer destination candidate adds the task to its own worker setting information 350. As a result, the transfer destination candidate will be in charge of the task from now on.

<Features of the excessive work elimination process>
In the excessive work resolution process, when the activity amount of the worker agent 220 is greater than a specified amount in the task knowledge database 140, the task performed by the worker agent 220 (worker) and the target asset included in the worker setting information 350 are added to the worker setting information of the worker agent 220 that is different from the worker agent 220 and has an activity amount less than the specified amount. This excessive work resolution process distributes the burden of the worker agent 220 that has been performing excessive maintenance work, and the amount of work is optimized. The excessive work resolution process is repeated until an excessive amount of work is no longer detected in the worker agent 220, thereby eliminating the worker agent 220 that generates excessive work. By correcting defects in the worker setting information 340 and the worker setting information 350 caused by defects in the maintenance work log database 130, an appropriate maintenance work simulation can be performed.

In this way, the maintenance improvement support device 100 is equipped with an excessive work elimination processing unit 116 (reproducibility improvement processing unit) that detects the state of a worker agent that violates the regulations in the task knowledge database, which contains regulations for performing maintenance work, and corrects the setting information (worker setting information 350). In this case, violating the regulations in the task knowledge database (see Figure 3) means that the overtime hours calculated from the time in the "task execution state" are equal to or greater than the specified value of "overtime hours per worker per month" in the task knowledge database 140 (see Figure 3).

<Features of the maintenance improvement support device>
The simulation setting generation unit 111 generates setting information 300 necessary for executing a simulation of a maintenance operation from the maintenance operation log database 130. The setting information 300 includes asset setting information 320, task setting information 330, worker setting information 340, and operator setting information 350. Next, the agent generation unit 112 generates an agent 200 based on the setting information. Then, the simulation unit 113 runs (operates) the agent 200 to execute the simulation.

The maintenance work log database 130 does not necessarily include all information related to real assets and workers, and there may be problems with the setting information. The worker deficiency response processing unit 114 resolves the deficiency in the worker setting information 350 when the asset agent 210 is in a "failure state" or "waiting for regular inspection state" for a certain period of time. The agent merging processing unit 115 merges the asset agents 210 to reduce the amount of calculation and memory usage of the simulation when the frequency of execution of inspection tasks or the frequency of failure states in the asset agents 210 are low. The excessive work elimination processing unit 116 adds a new worker agent 220 or adds a task to the worker setting information 350 of an existing worker agent 220 when the task volume of the worker agent 220 is excessive, allowing the tasks to be distributed among the worker agents 220.

In this way, appropriate configuration information 300 is generated by repeating the process of dealing with a shortage of workers, the process of combining agents, and the process of eliminating excessive work until problems such as a failure state lasting for a certain period of time are no longer detected. As a result, an appropriate maintenance work simulation can be executed.

<<Modification: Evaluation Unit>>
The simulation unit 113 may store the start and end of a task and changes in the state of the asset agent 210 as an event log. At the same time, modifications to the setting information 300 may also be stored as an event log. The maintenance improvement support device 100 may also include an evaluation unit that performs statistical processing and visualization processing of the event log and presents it to the user. For example, the evaluation unit may calculate the time from the occurrence of a failure to the completion of a failure response task, the average overtime hours, the maximum overtime hours, the maximum continuous work hours of workers, and the like, and present them to the user. The evaluation unit may also present the modification history/modification contents (added tasks, addition history) of the worker setting information 340 and the worker setting information 350 to the user.

Modifications to the worker setting information 340 and the worker setting information 350 indicate changes in the assets and types (task types) that are the targets of the maintenance work (tasks) that the workers are responsible for, due to changes in the maintenance business situation, such as the addition or deletion of assets in the plan database 150 or the increase or decrease in the number of workers. The event log also indicates a situation in which the maintenance of an asset is no longer being performed appropriately due to a change in the maintenance business situation, and modifications to the worker setting information 340 and the worker setting information 350 present the measures required at that time. This enables users of the maintenance improvement support device 100 to plan future asset addition or disposal, and to plan the employment and training of workers.

<<Variation: Simulation Execution by Agent>>
In the above embodiment, the simulation unit 113, the worker vacancy response processing unit 114, the agent merging processing unit 115, and the excessive work resolution processing unit 116 operate the data of the agent 200 (agent data 160) to execute the simulation and the reproducibility improvement processing. Alternatively, the agent 200 may be provided with a processing module and execute the simulation and the reproducibility improvement processing. Also, other execution forms of the agent simulation may be used.

<<Variation: Modifying the setting information>>
In the above embodiment, the setting information 300 is modified by the reproducibility improvement process. Each of the asset agent 210 and the worker agent 220 may internally hold information related to itself, such as the model and tasks, and modify the information when a simulation or a reproducibility improvement process is executed. The form in which the information related to the agent is held may be an appropriate form depending on the execution form of the agent simulation.

<<Variation: Task Setting Information>>
Tasks (details of work types) are recorded in the maintenance work log database 130, and the simulation setting generator 111 generates task setting information 330 (see FIG. 8) for each task. However, when dealing with a failure, the details of the failure may not be determined at the start of the work, or there may be a large degree of uncertainty about the details of the investigation and repair/replacement work, so that the tasks (details of work types), which are detailed classification codes, may not be recorded.

In such cases, since the ultimate goal of the present invention is to simulate and evaluate the execution status of maintenance work, it is possible to improve the accuracy of the simulation by generating separate task definitions based on either or both of the occurrence pattern and the load of each task, thereby subdividing the detailed task definitions. For example, it is possible to classify tasks into those with similar durations by classifying them by short, medium, or long work time, or by clustering work times. In addition, since the number of workers required to perform a task needs to be distinguished when allocating workers in the simulation, it is necessary to subdivide the ruler tasks according to the number of personnel required.

<<Modification: Worker Setting Information>>
Although the worker setting information 350 (see FIG. 10 ) indicates the assets and tasks for which the worker is responsible, it may be simplified. For example, rather than individual assets, it may be aggregated by the location where the assets are installed, or by model. Furthermore, rather than individual tasks, it may be simplified by distinguishing by work type or skill. For example, records of worker setting information for the same or nearby asset installation locations, the same or nearby asset models, and the same or nearby task types may be merged to reduce the number of records of worker setting information. By reducing the setting information 300 in this way, the efficiency of simulation execution is improved.

Furthermore, the worker setting information 350 is generated based on the maintenance work log database 130. For this reason, it is possible that the assets and tasks that a certain worker is responsible for in the actual maintenance business design are not in the maintenance work log database 130, and worker setting may not be generated. Therefore, workers who are highly similar in the assets and tasks they are responsible for may be presumed to have the same asset and task scope in the maintenance work design, and the missing worker setting may be supplemented by finding the set sum of the assets and tasks they are responsible for. By clustering workers who have records of similar assets and tasks, it is possible to generate worker setting information 350 that is closer to reality and has no omissions.

<<Variation: Regular Inspection Task>>
In the above embodiment, the regular inspection task generation agent 240 performs the task generation process for regular inspection, but each asset agent 210 may perform this process. For example, the status update unit 211 of the asset agent 210 counts the time from the last regular inspection to the present time, and if the regular inspection period has passed, it considers that the inspection has not been completed and operation is not possible, and updates the internal status of the asset agent to "waiting for regular inspection". Furthermore, the status update unit 211 may request a regular inspection from the work plan agent 230 via the interaction unit 212.

<<Variation: Agent Combining Process>>
In the above embodiment, the agent merging processing unit 115 performs agent merging processing on asset agents 210 that perform inspection tasks or regular inspection tasks infrequently, or that enter a "failed state" or a "waiting regular inspection state" infrequently. If the frequency is lower than a predetermined value and the impact on the simulation is considered to be small, the agent merging processing unit 115 may simply delete the asset agent 210 rather than merging it, and may delete it from the asset setting information 320.

Regarding worker agents 220, the agent merging processing unit 115 may merge worker agents 220 that have a low task execution frequency (activity level) and are located at the same or nearby base (see base in the worker setting information 340 in FIG. 9) to form a single worker agent 220. Whether or not a task is executed frequently may be determined based on, for example, the time multiplied by a predetermined value of the "standard work time per worker month" or the "standard work time per worker month" in the business knowledge database 140 (see FIG. 3). In addition, if the frequency is lower than the predetermined value and the impact on the simulation is considered to be small, the agent merging processing unit 115 may simply delete the worker agent 220 from the worker setting information 340 instead of merging.

In this way, if the activity amount of an asset agent 210 or worker agent 220 during simulation execution is less than a specified value in violation of the regulations in the business knowledge database 140, the agent merging processing unit 115 merges it with another asset agent 210 or worker agent 220 whose activity amount is less than the specified amount. In addition, the agent merging processing unit 115 may delete the asset agent 210 or worker agent 220 during simulation execution if the activity amount (frequency of task execution) of the asset agent 210 or worker agent 220 during simulation execution is less than a predetermined value. By merging or deleting the asset agent 210 or worker agent 220, the amount of calculation and memory usage of the simulation can be reduced, and processing time can be shortened.

Other variations
Although several embodiments of the present invention have been described above, these embodiments are merely examples and do not limit the technical scope of the present invention. For example, the maintenance improvement support device 100 stores the maintenance work log database 130, the work knowledge database 140, and the plan database 150, but may access this information in an external device. The worker vacancy response processing unit 114, the agent combination processing unit 115, and the excessive work resolution processing unit 116 may be one functional unit (reproducibility improvement processing unit).

The present invention can take various other embodiments, and various modifications such as omissions and substitutions can be made without departing from the gist of the present invention. These embodiments and their modifications are included in the scope and gist of the invention described in this specification, etc., and are included in the scope of the invention described in the claims and their equivalents.

100 Maintenance improvement support device 111 Simulation setting generation unit 112 Agent generation unit 113 Simulation unit 114 Worker absence response processing unit (reproducibility improvement processing unit)
115 Agent combination processing unit (reproducibility improvement processing unit)
116 Excessive work elimination processing unit (reproducibility improvement processing unit)
130 Maintenance work log database 140 Work knowledge database 150 Plan database 210 Asset agent 220 Worker agent 300 Setting information 320 Asset setting information 330 Task setting information 340 Worker setting information 350 Worker setting information

Claims (9)

a simulation setting generation unit that generates setting information for simulating the maintenance work based on a maintenance work log database that includes an asset that is a maintenance target in the maintenance work, a worker in charge of the maintenance work, and a task that is the maintenance work performed by the worker on the asset, and is a record of the execution of the maintenance work;
a simulation unit that executes a simulation of the maintenance work using an asset agent that is an agent representing the asset and a worker agent that is an agent representing the worker;
a reproducibility improvement processing unit that detects a state of the asset agent or the worker agent that violates a rule in a task knowledge database that includes rules for performing the maintenance task , and corrects the setting information;
The simulation setting generation unit
generating task setting information from the maintenance work log database, the task setting information being setting information associating the asset, a failure pattern occurring in the asset, and the task performed by the worker corresponding to the failure;
generating worker setting information that associates the worker, the asset that is the target of the task performed by the worker, and the task from the maintenance work log database;
The simulation unit is
a failure occurs in the asset agent according to a failure pattern included in the task setting information;
An allocation of a task corresponding to the failure to an asset included in the worker setting information and corresponding to the asset agent and a worker agent associated with the task is determined;
The worker agent executes a simulation of the asset agent performing the task;
The reproducibility improvement processing unit includes:
and an excessive work resolution processing unit that, when the activity amount of the worker agent during the execution of a simulation is greater than a specified value in violation of a rule in the task knowledge database, adds a task and a target asset to be performed by the worker corresponding to the worker agent included in the worker setting information to the worker setting information of a worker different from the worker and having an activity amount smaller than the rule.
Maintenance improvement support equipment. a simulation setting generation unit that generates setting information for simulating the maintenance work based on a maintenance work log database that includes an asset that is a maintenance target in the maintenance work, a worker in charge of the maintenance work, and a task that is the maintenance work performed by the worker on the asset, and is a record of the execution of the maintenance work;
a simulation unit that executes a simulation of the maintenance work using an asset agent that is an agent representing the asset and a worker agent that is an agent representing the worker;
a reproducibility improvement processing unit that detects a state of the asset agent or the worker agent that violates the rule in a task knowledge database that includes the rule when performing the maintenance task , and corrects the setting information;
The simulation setting generation unit
generating task setting information from the maintenance work log database, the task setting information being setting information associating the asset, a failure pattern occurring in the asset, and the task performed by the worker corresponding to the failure;
generating worker setting information that associates the worker, the asset that is the target of the task performed by the worker, and the task from the maintenance work log database;
The simulation unit is
a failure occurs in the asset agent according to a failure pattern included in the task setting information;
An allocation of a task corresponding to the failure to an asset included in the worker setting information and corresponding to the asset agent and a worker agent associated with the task is determined;
The worker agent executes a simulation of the asset agent performing the task;
The reproducibility improvement processing unit includes:
and a task deficiency response processing unit that searches for an agent representing a worker who will perform the task when a period during which a task is not performed after a failure occurs in the task setting information for the asset agent during simulation is longer than a period specified in the business knowledge database, and adds the task and the asset corresponding to the asset agent to the worker assignment setting information of the worker corresponding to the agent representing the worker.
Maintenance improvement support equipment. The maintenance work log database includes:
The information includes at least one of an installation location of the asset, a model of the asset, and a type of the task;
The simulation setting generation unit
The maintenance improvement support device according to claim 1 or 2, wherein the worker setting information for the assets having the same installation location, the same model of the assets, and the same task type is merged to reduce the number of the worker setting information. The reproducibility improvement processing unit includes:
3. The maintenance improvement support device according to claim 1, further comprising an agent combination processing unit that deletes the agent when an activity amount of the agent during execution of a simulation is smaller than a predetermined value. The reproducibility improvement processing unit includes:
3. The maintenance improvement support device according to claim 1, further comprising an agent merging processing unit that, when an activity amount of the agent during execution of a simulation is smaller than a specified value contrary to a rule in the business knowledge database, merges the agent with another agent having an activity amount smaller than the rule. The simulation unit is
The maintenance improvement support device according to claim 1 , wherein the asset agents and the worker agents are increased or decreased based on a plan database which is a future increase or decrease plan for the assets and the workers. The maintenance improvement support device according to claim 1 or 2 , further comprising an evaluation unit that displays a history of additions to the worker setting information. The maintenance improvement support device,
generating setting information for simulating the maintenance work based on a maintenance work log database that includes an asset that is a maintenance target in the maintenance work, a worker in charge of the maintenance work, and a task that is the maintenance work performed by the worker on the asset, and is a record of the execution of the maintenance work;
executing a simulation of the maintenance work using an asset agent which is an agent representing the asset and a worker agent which is an agent representing the worker;
detecting a state of the asset agent or the worker agent that violates a rule in a task knowledge database that includes rules for performing the maintenance task , and correcting the setting information;
In the step of generating setting information for performing a simulation of the maintenance work,
generating task setting information from the maintenance work log database, the task setting information being setting information in which the asset, a failure pattern occurring in the asset, and the task performed by the worker corresponding to the failure are associated with each other;
generating worker setting information, which is setting information associating the worker, the asset that is the target of the task performed by the worker, and the task, from the maintenance work log database;
In the step of executing a simulation of the maintenance work,
a step of causing a failure in the asset agent according to a failure pattern included in the task setting information;
A step of determining an allocation of a task corresponding to the failure to an asset included in the worker setting information and corresponding to the asset agent and a worker associated with the task;
A step of executing a simulation in which the worker agent performs the task on the asset agent;
In the step of modifying the setting information,
When the activity amount of the worker agent during the simulation is larger than a specified value contrary to the specification in the task knowledge database, a step of adding a task and a target asset to be performed by the worker corresponding to the worker agent included in the worker setting information to the worker setting information of a worker different from the worker and having an activity amount smaller than the specification is executed.
Maintenance improvement support methods. The maintenance improvement support device,
generating setting information for simulating the maintenance work based on a maintenance work log database that includes an asset that is a maintenance target in the maintenance work, a worker in charge of the maintenance work, and a task that is the maintenance work performed by the worker on the asset, and is a record of the execution of the maintenance work;
executing a simulation of the maintenance work using an asset agent which is an agent representing the asset and a worker agent which is an agent representing the worker;
detecting a state of the asset agent or the worker agent that violates a rule in a task knowledge database that includes rules for performing the maintenance task , and correcting the setting information;
In the step of generating setting information for performing a simulation of the maintenance work,
generating task setting information from the maintenance work log database, the task setting information being setting information in which the asset, a failure pattern occurring in the asset, and the task performed by the worker corresponding to the failure are associated with each other;
generating worker setting information, which is setting information associating the worker, the asset that is the target of the task performed by the worker, and the task, from the maintenance work log database;
In the step of executing a simulation of the maintenance work,
a step of causing a failure in the asset agent according to a failure pattern included in the task setting information;
A step of determining an allocation of a task corresponding to the failure to an asset included in the worker setting information and corresponding to the asset agent and a worker associated with the task;
A step of executing a simulation in which the worker agent performs the task on the asset agent;
In the step of modifying the setting information,
When a period during which a task is not performed after a failure occurs in the task setting information for the asset agent during simulation is longer than a period specified in the business knowledge database, a step of searching for an agent representing the worker who will perform the task and adding the task and the asset corresponding to the asset agent to the worker setting information of the worker corresponding to the agent representing the worker is executed.
Maintenance improvement support methods.

Images