Industrial equipment operation support system

Abstract

We support timely and efficient activities for a wide variety of equipment. [Solution] The industrial equipment operation support system 1 includes an information processing device 6. The information processing device 6 includes an information storage platform 37 that stores acquired environmental information and operational information, a condition index preparation engine 32 that extracts or generates condition indexes for industrial equipment in a time series from at least one of the stored environmental information and operational information, an inference value storage engine 33 that infers the probability of malfunction events occurring by inputting the condition indexes into a trained inference model and stores the obtained inferred values ​​of occurrence in the information storage platform 37, and a task generation engine 34 that generates maintenance recommendation tasks based on the results of data analysis using the time series data of the stored inferred values ​​of occurrence.

Description

[Technical Field]

[0001] The technology disclosed herein relates to an industrial equipment operation support system. [Background technology]

[0002] In the technical field related to industrial equipment, maintenance and management methods such as those disclosed in Patent Document 1 are known. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2002-298270 [Overview of the project] [Problems that the invention aims to solve]

[0004] When operating industrial equipment, it is necessary to carry out various activities such as daily or periodic planning, inspection, testing, adjustment, and replacement to maintain equipment performance, which include functions such as preventing equipment deterioration, measuring deterioration, and restoring deterioration. However, it is difficult to carry out these activities in a timely and efficient manner for a wide variety of equipment, and there is a demand for technologies that can support the operation of industrial equipment.

[0005] The technology disclosed herein provides an industrial equipment operation support system that can support timely and efficient equipment maintenance activities for a wide variety of equipment. [Means for solving the problem]

[0006] This specification provides an industrial equipment operation support system. The industrial equipment operation support system comprises a plurality of environmental sensors placed on the industrial equipment, a plurality of controllers equipped on the industrial equipment, and a plurality of information processing devices configured to acquire and store environmental information detected by the environmental sensors and operational information generated by the controllers, and having a hierarchical structure for transmitting the environmental information and operational information from downstream to upstream. The plurality of information processing devices each have, in any one of the layers, an information storage platform for storing acquired environmental information and operational information, a condition index preparation engine for extracting or generating condition indexes for the industrial equipment in a time series from at least one of the environmental information and operational information stored on the information storage platform, an inference value storage engine having a trained inference model that infers the probability of occurrence of a malfunction event by inputting the condition index and stores the obtained inferred value of the occurrence probability on the information storage platform, and a task generation engine that generates maintenance recommendation tasks based on the results of data analysis using the time series data of the inferred value of the occurrence probability stored on the information storage platform. [Effects of the Invention]

[0007] The technology disclosed herein can support timely and efficient equipment maintenance activities for a wide variety of equipment. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 is a schematic diagram illustrating an industrial equipment operation support system according to an embodiment. [Figure 2] Figure 2 is a schematic diagram showing industrial equipment and demand equipment according to the embodiment. [Figure 3] Figure 3 is a hardware configuration diagram showing an information processing device according to an embodiment. [Figure 4] Figure 4 is a diagram illustrating an information processing device according to an embodiment. [Figure 5] Figure 5 is a functional block diagram showing an information processing device according to an embodiment. [Figure 6] Figure 6 shows an example of an information storage platform according to this embodiment. [Figure 7] Figure 7 shows an example of a trained inference model according to the embodiment. [Figure 8] Figure 8 shows an example of a database according to this embodiment. [Figure 9] Figure 9 is a flowchart showing a maintenance method for an industrial equipment operation support system according to an embodiment. [Figure 10] Figure 10 is a schematic diagram illustrating an example of the operation of an industrial equipment operation support system according to an embodiment. [Modes for carrying out the invention]

[0009] [1] Industrial equipment operation support system Figure 1 is a schematic diagram showing an industrial equipment operation support system 1 according to an embodiment. The industrial equipment operation support system 1 acquires and stores environmental information and operational information of industrial equipment 2 to support the operation of industrial equipment 2. The industrial equipment operation support system 1 generates useful information that contributes to equipment operation using the environmental information and operational information of industrial equipment 2. The industrial equipment operation support system 1 supports the operation of industrial equipment 2 by providing the user of industrial equipment 2 with useful information that contributes to equipment operation.

[0010] The user of the industrial equipment 2 refers to a person who owns or leases the industrial equipment 2. The maintenance contractor of the industrial equipment 2 provides the user with comprehensive equipment operation support services through the industrial equipment operation support system 1 based on the contract concluded with the user of the industrial equipment 2. The comprehensive equipment operation support services aim to create an environment in which the user can concentrate on the core business (the main business related to the production of goods or the provision of services) by having the maintenance contractor undertake various construction works for equipment maintenance and substitute for equipment operation, and further maximize the performance of equipment operation. Incidentally, equipment maintenance (Productive Maintenance) is a general term for various activities such as daily, regular or situation-based planning, inspection, examination, adjustment, repair, replacement, etc., which are responsible for functions such as preventing equipment deterioration, measuring deterioration and restoring deterioration in order to maintain equipment performance.

[0011] The comprehensive equipment operation support services may include, for example, the following services. (1) Provision of hardware and software for connecting the industrial equipment 2 and the maintenance contractor with information and communication technology (ICT: Information and Communication Technology). (2) Provision of a user interface (UI: User Interface) for connecting the user and the maintenance contractor with ICT. (3) Provision of visualizable content of the operation performance information (operation performance by the maintenance contractor) of the industrial equipment 2. (4) Provision of visualizable content of the condition information (current status / history) of the industrial equipment 2. (5) Provision of before maintenance (condition-based maintenance / time-based maintenance) and after maintenance for the industrial equipment 2. (6) Provision of notifications regarding the daily management of the industrial equipment 2. (7) Professional proposals regarding the smart operation (labor saving / autonomization) and added value improvement (low carbonization / SDGs) of the industrial equipment 2.

[0012] Figure 2 is a schematic diagram showing industrial equipment 2 and demand equipment 8 according to an embodiment. Equipment refers to machinery installed in buildings such as factories, or in vehicles, ships, etc. Machinery is a general term for machinery, instruments, and equipment. Industrial equipment 2 refers to machinery and equipment used for the production of goods or the provision of services. Industrial equipment 2 includes utility conversion machinery and equipment that converts primary utilities into secondary utilities usable by demand equipment 8, a utility transmission and distribution network that transmits and distributes secondary utilities, and demand-end machinery and equipment that utilize primary or secondary utilities. Demand equipment 8, consisting of demand-end machinery and equipment, is one embodiment of industrial equipment 2.

[0013] A utility refers to an energy source or fluid necessary for industrial activities. Examples of primary utilities input to utility conversion machinery include fuel (gas, oil), electricity, and raw water. Examples of secondary utilities output from industrial equipment include heat transfer fluids (steam, heat transfer oil, hot water, chilled water), compressed air, electricity, and treated water.

[0014] Demand equipment 8 utilizes secondary utilities output from industrial equipment 2. Demand equipment 8 uses heat transfer fluid as a heat source for various production processes or air conditioning. Demand equipment 8 uses compressed air as a power source for pneumatic equipment or pneumatic tools. Demand equipment 8 uses electricity as a power source for electric equipment, power tools, or lighting. Demand equipment 8 uses treated water as process water for food, cosmetics, pharmaceuticals, or semiconductor manufacturing.

[0015] Examples of utility conversion machinery and equipment include thermal equipment, air compressors, generators, and water treatment equipment. Examples of thermal equipment include combustion steam boilers, electric heater steam boilers, heat recovery steam boilers, combustion heat transfer boilers, combustion hot water boilers, electric heater hot water boilers, heat recovery hot water boilers, electric heat pumps, electric chillers, electric heat pump chillers, and flash steam generators. Examples of air compressors include electric air compressors, heat recovery electric air compressors, and steam-driven air compressors. Examples of generators include monogenerators and cogeneration generators. Examples of water treatment equipment include reverse osmosis membrane systems, hard water softeners, deoxygenation systems, and various filtration systems.

[0016] Industrial equipment 2 converts primary utilities into secondary utilities usable by demand equipment 8. Primary utilities are supplied to industrial equipment 2 via supply route 21. Secondary utilities discharged from industrial equipment 2 are delivered to demand equipment 8 via transport route 22. Secondary utilities that have passed through demand equipment 8 are discharged via discharge route 23. Transport route 22 may be formed as a transport piping network. Auxiliary equipment for medium storage (such as steam headers, hot water tanks, and chilled water tanks) may also be provided along transport route 22.

[0017] Furthermore, industrial equipment 2 may also include medical machinery and equipment used in a series of processes from receiving to discharging items to be washed and sterilized, laundry machinery and equipment used in a series of processes from collecting to shipping laundry, food and beverage manufacturing machinery and equipment used in a series of processes from receiving raw materials to storing products, vehicles for unmanned transport of goods between multiple points set up within the business premises, navigation machinery and equipment and cargo handling machinery and equipment installed on ships such as cargo ships.

[0018] Examples of medical equipment include washers and sterilizers. Examples of washers include vacuum boiling washers and ultrasonic washers. Examples of sterilizers include steam sterilizers and gas sterilizers. Medical equipment is installed in the central sterile supply department of a medical institution.

[0019] Examples of laundry machinery include washing machines, dryers, and finishing machines. Examples of washing machines include continuous washing machines, water washing machines, and dry cleaning machines. Examples of dryers include gas dryers and steam dryers. Examples of finishing machines include gas roll ironers and steam roll ironers. Laundry machinery is installed in laundry factories.

[0020] Examples of machinery and equipment for food and beverage manufacturing include thawers, cooking machines, coolers, and sterilizers. Examples of thawers include vacuum steam thawers, microwave thawers, high-frequency thawers, and running water thawers. Examples of cooking machines include steam kneaders, steam kettles, and saturated steam cookers. Examples of coolers include vacuum coolers, chilled water coolers, and cold air coolers. Examples of sterilizers include retort sterilizers and pasteurizers. Machinery and equipment for food and beverage manufacturing are installed in food and beverage factories.

[0021] Examples of autonomously operating unmanned transport vehicles (AGVs) include trolley-type vehicles, forklift-type vehicles, and towed vehicles. AGVs are used in various manufacturing plants.

[0022] Examples of navigational machinery include main engines (single-fuel / dual-fuel diesel engines), turbochargers (auxiliary equipment), exhaust gas economizers (steam generators), shaft generators, steam turbine generators, binary generators, desalination plants, boilers for BOG combustion, and exhaust gas cleaning equipment. Examples of cargo handling machinery include diesel generators, cranes, derricks, ballast water pumps, and ballast water treatment equipment.

[0023] Returning to Figure 1, Establishment 3 refers to an individual place where the production of goods or the provision of services is carried out as a business. Industrial equipment 2 is installed in Establishment 3. Factory 4 is provided in Establishment 3 where goods are produced, etc. Examples of factory 4 include food factories, beverage factories, metal product factories, plastic product factories, textile factories, and laundry factories. Industrial equipment 2 is installed in factory 4.

[0024] Furthermore, a factory 4 is not required to be provided at the business establishment 3 that provides the service. The business conducted at business establishment 3 may include public health services. Examples of public health services include hospitals, clinics, and health centers. Business establishment 3 may also include a catering center. Instead of business establishment 3, industrial equipment 2 may be installed on a vessel used for maritime transport services.

[0025] In the example shown in Figure 1, a certain business operator has three business establishments 3. Business establishments 3 include the first business establishment 3A, the second business establishment 3B, and the third business establishment 3C. A factory 4 is located in the second business establishment 3B. The factory 4 located in the second business establishment 3B includes the first factory 4A and the second factory 4B. Alternatively, a factory 4 may be located in either the first business establishment 3A or the third business establishment 3C, or both.

[0026] Industrial equipment 2 is installed in both the first factory 4A and the second factory 4B. Data collection terminals 7 are installed in both the first factory 4A and the second factory 4B. Data collection terminals 7 is a general term for devices used for data collection at the business establishment 3.

[0027] The industrial equipment operation support system 1 comprises a plurality of environmental sensors 5 placed on the industrial equipment 2, a plurality of controllers 61 equipped on the industrial equipment 2, and a plurality of information processing devices 6 configured to acquire and store environmental information detected by the environmental sensors 5 and operational information generated by the controllers 61.

[0028] The Environmental Sensor 5 detects environmental information of the industrial equipment 2. Environmental information of the industrial equipment 2 refers to the environmental state or conditions of the space in which the industrial equipment 2 operates. This includes environmental information of the business establishment 3 (factory 4) where the industrial equipment 2 is installed. The environmental information includes physical parameters of the industrial equipment 2 and its surroundings. Some of the detection data from the Environmental Sensor 5 is used for the operation or control of the industrial equipment 2. Examples of Environmental Sensor 5 include temperature sensors, humidity sensors, pressure sensors, water level sensors, flow rate sensors, electrical conductivity sensors (EC sensors), power sensors, distance sensors, image sensors, and force sensors.

[0029] The environmental sensor 5 is connected to the controller 61 of the industrial equipment 2 and the controller 61 of the data acquisition terminal 7, respectively.

[0030] As shown in Figure 2, the environmental sensor 5 is installed in the supply path 21 (primary utility side). The environmental sensor 5 is installed in the transport path 22 (secondary utility side). The environmental sensor 5 may be positioned closer to the industrial equipment 2 or closer to the demand equipment 8 in the transport path 22. The environmental sensor 5 may also be positioned in the discharge path 23 (downstream of the demand equipment 8).

[0031] In this embodiment, the environmental information of the industrial equipment 2 detected by the environmental sensor 5 includes input energy information and output energy information of the industrial equipment 2. The input energy information includes the input amount of the primary utility. The output energy information includes the output amount of the secondary utility. The environmental information includes information on the actual quality level of the medium or goods produced by the industrial equipment 2.

[0032] Returning to Figure 1, the controller 61 has the function of controlling the operation of the industrial equipment 2 and controlling data collection. The controller 61 is mainly used for controlling the operation of machinery and equipment (heat equipment, water treatment equipment, water quality measuring equipment, etc.). The controller 61 is connected to environmental sensors 5 attached to the machinery and equipment. The controller 61 may also be configured as part of a data collection terminal 7 that specializes in collecting information from environmental sensors 5 that are not attached to the machinery and equipment (environmental sensors that are retrofitted to the machinery and equipment or retrofitted to piping networks, etc.).

[0033] The controller 61 is incorporated into the industrial equipment 2 and the data acquisition terminal 7. Examples of the controller 61 incorporated into the industrial equipment 2 include a microcomputer 61A and a programmable logic controller 61B (PLC). In this embodiment, the industrial equipment 2 includes a first industrial equipment 2A where the microcomputer 61A is located and a second industrial equipment 2B where the programmable logic controller 61B is located. An example of the controller 61 incorporated into the data acquisition terminal 7 is the microcomputer 61A.

[0034] The controller 61 of the industrial equipment 2 uses environmental information collected from the environmental sensor 5 to control the operation of the industrial equipment 2 and records it for operational management. The controller 61 of the industrial equipment 2 may also receive environmental information collected by controllers 61 incorporated in other industrial equipment 2 or data collection terminals 7 and use it to control its own operation.

[0035] The controller 61 of the data collection terminal 7 is connected to each of the multiple environmental sensors 5 scattered throughout the business premises 3. The controller 61 of the data collection terminal 7 is connected to multiple environmental sensors 5 of different types.

[0036] The controller 61 generates operational information for the industrial equipment 2. The controller 61 may generate operational information for the industrial equipment 2 based on detection data from the environmental sensor 5. In this embodiment, the operational information for the industrial equipment 2 generated by the controller 61 includes the operating time of the industrial equipment 2 and whether or not there is an abnormality in the industrial equipment 2. The operational information includes actual operating time information linked to the normal state of the industrial equipment 2 and non-operating time information linked to the abnormal state. The operational information includes whether the condition of the industrial equipment 2 is good or bad. The operational information includes an abnormality notification signal transmitted by the controller 61 when a malfunction occurs in the industrial equipment 2.

[0037] The condition of the industrial equipment 2 is detected by the environmental sensor 5. For example, when the condition of the industrial equipment 2 deteriorates, the value detected by the environmental sensor 5 will often be a value that deviates from the value detected by the environmental sensor 5 when the condition of the industrial equipment 2 is normal. Therefore, multiple thresholds are set according to the level of deviation of the value detected by the environmental sensor 5 from the normal value, and the controller 61 is instructed to detect if there is a sign of an abnormality when the deviation level reaches the lower first threshold, and if an abnormality has occurred when the deviation level reaches the higher second threshold.

[0038] If industrial equipment 2 is a steam boiler, the operational information related to its condition includes scale buildup information based on water tube temperature and feedwater pump performance information based on boiler water level control. If industrial equipment 2 is a heat pump or chiller, the operational information related to its condition includes refrigerant leak detection information and differential pressure information for the condenser and evaporator. If industrial equipment 2 is an air compressor, the operational information related to its condition includes lubricant degradation detection information and differential pressure information for filters. If industrial equipment 2 is an RO membrane system, the operational information related to its condition includes permeate flux information for the membrane element and water quality information for the permeate.

[0039] [2] Information processing device The industrial equipment operation support system 1 has multiple information processing devices 6. The information processing devices 6 include a controller 61, an edge computer 62, a gateway 63, a guest computer 64, and a host computer 65.

[0040] Figure 3 is a hardware configuration diagram showing an information processing device 6 according to an embodiment. The information processing device 6 includes a computer 10. The controller 61, edge computer 62, gateway 63, guest computer 64, and host computer 65 each include the computer 10. The computer 10 has a processor 11, a storage device 12, a communication interface 13, and an input / output interface 14. The information processing device 6 also has a power supply (not shown).

[0041] The processor 11 includes a CPU (Central Processing Unit). The processor 11 may also include a GPU (Graphics Processing Unit). The storage device 12 includes a recording medium on which computer programs and data are recorded in a readable format by the processor 11. The storage device 12 includes onboard system memory such as RAM (Random Access Memory) or ROM (Read Only Memory), high-capacity flash memory such as an SD card or USB memory, and high-capacity storage such as an HDD (Hard Disk Drive) or SSD (Solid State Drive).

[0042] The communication interface 13 communicates via a communication network. Examples of communication networks include local area networks (LANs), wide area networks (WANs), and commercial networks such as the Internet. A local area network may be a wired LAN or a wireless LAN. A wide area network may include mobile lines or satellite communication lines. Computer 10 transmits data to an external computer via the communication network. Computer 10 receives data from an external computer via the communication network. Computer 10 connects to an external device via the input / output interface 14.

[0043] The storage device 12 stores various software programs (engines and applications, described later). The processor 11 reads the software programs from the storage device 12, loads them into system memory, and executes processing according to the software programs. In other words, the processor 11 can be considered to have multiple functional units, such as engines. The functions of the engines and other components of the processor 11 are realized by software programs. The software programs may be distributed to the computer 10 via a communication network.

[0044] A software program that implements a specific function on a computer is called an engine, and the functional units of an engine are called modules. An engine may be installed on a computer as a single software package containing all its functions, but it is preferable to install it on a computer as individual software modules, each representing a functional unit of the engine. Modularizing the functional units of an engine makes it easier to update when functional modifications are made. Software programs prepared for end users (mainly customers who use equipment) to perform specific tasks on a computer are called applications, and they can be distinguished from engines in terms of their purpose.

[0045] The environmental sensor 5 is connected to the input / output interface 14 of the controller 61. Multiple environmental sensors 5 are connected to one controller 61. The controller 61 collects environmental information from the environmental sensors 5 in real time. The communication interface 13 of the controller 61 transmits real-time environmental information and real-time operational information to the edge computer 62 via the communication network (LAN).

[0046] Returning to Figure 1, the edge computer 62 is installed in factory 4. One or more edge computers 62 are installed in each factory 4. The communication interface 13 of the edge computer 62 communicates with each of the multiple controllers 61 belonging to factory 4 where the edge computer 62 is located, via a communication network (LAN). The edge computer 62 receives environmental information and operational information from the controllers 61 via the communication network.

[0047] The edge computer 62 has approximately 5GB of onboard memory as storage device 12 so that it can store a sufficient amount of information. The edge computer 62 may also have an AI engine (neural network processing unit: NPU) so that it can perform the learning and inference phases in machine learning.

[0048] Gateway 63 constitutes a connection node between the local area network and the internet, and has performance and specifications equivalent to that of the edge computer 62. Gateway 63 is installed at business establishment 3. One or more gateways 63 are installed at a single business establishment 3. The communication interface 13 of gateway 63 communicates with each of the multiple edge computers 62 belonging to business establishment 3 where gateway 63 is located, via the communication network (LAN). Gateway 63 receives environmental information and operational information from the edge computers 62 via the communication network. If the factory 4 located at business establishment 3 is a single building, the edge computer 62 and gateway 63 may be integrated and configured as a single computer.

[0049] The guest computer 64 is located outside of the business premises 3. For example, the guest computer 64 is installed at a regional base of a service provider that undertakes tasks such as monitoring the status of industrial equipment 2 installed at business premises 3 and performing maintenance inspections. One guest computer 64 is installed at each management base. The guest computer 64 includes a local server. The communication interface 13 of the guest computer 64 communicates with the gateway 63 belonging to business premises 3 via a communication network (Internet). The guest computer 64 receives environmental information and operational information from the gateway 63 via the communication network. If a guest computer 64 is not installed, the gateway 63 and the host computer 65 (described later) will be connected via the communication network.

[0050] The host computer 65 is located outside of the business premises 3. For example, the host computer 65 is installed at the central hub of a service provider. The host computer 65 includes a cloud server. The communication interface 13 of the host computer 65 communicates with each of the multiple guest computers 64 via a communication network (the Internet). The host computer 65 receives environmental information and operational information from the guest computers 64 via the communication network.

[0051] [3] Hierarchical structure Figure 4 is a diagram illustrating an information processing device 6 according to an embodiment. The multiple information processing devices 6 have a hierarchical structure that transmits environmental information detected by the environmental sensor 5 and operational information generated by the controller 61 from the downstream side to the upstream side. Of the multiple information processing devices 6, the controller 61 to which the environmental sensor 5 is connected is the most downstream (lower layer, lower level) information processing device 6, the edge computer 62 is the next downstream information processing device 6 after the controller 61, the gateway 63 is the next downstream information processing device 6 after the edge computer 62, the guest computer 64 is the next downstream information processing device 6 after the gateway 63, and the host computer 65 is the most upstream (upper layer, higher level) information processing device 6.

[0052] The information processing devices 6 are connected to each other via a communication network. Within buildings and ships, a local area network is used as the communication network, while outside buildings and ships, commercial wide-area networks such as the internet and mobile networks are used. Satellite communication is used for communication between ships and land.

[0053] Environmental information is detected by the environmental sensor 5. The controller 61, which is a lower-level information processing device in the hierarchical structure, acquires real-time environmental information from the environmental sensor 5. The real-time environmental information detected by the environmental sensor 5 and collected by the controller 61 is transmitted from the controller 61 to the edge computer 62 via the communication network. Real-time operational information generated by the controller 61 is also transmitted from the controller 61 to the edge computer 62 via the communication network. The environmental information and operational information transmitted to the edge computer 62 are then transmitted from the edge computer 62 to the gateway 63 via the communication network. The environmental information and operational information transmitted to the gateway 63 are then transmitted from the gateway 63 to the guest computer 64 via the communication network. The environmental information and operational information transmitted to the guest computer 64 are then transmitted from the guest computer 64 to the host computer 65 via the communication network.

[0054] The controller 61 is a lower-level information processing unit in the hierarchical structure. The edge computer 62, gateway 63, and guest computer 64 are intermediate information processing units in the hierarchical structure. The host computer 65 is a higher-level information processing unit in the hierarchical structure.

[0055] The lower-level information processing unit functions as an IoT device to which the environmental sensor 5 is connected. The lower-level information processing unit can transmit various environmental and operational information to the intermediate information processing unit. The intermediate information processing unit functions as a relay between the lower-level information processing unit and the upper-level information processing unit. The intermediate information processing unit can receive various environmental and operational information from the lower-level information processing unit and transmit various environmental and operational information to the upper-level information processing unit. The upper-level information processing unit can receive various environmental and operational information from the intermediate information processing unit. The upper-level information processing unit has an information storage platform. The information storage platform is, for example, an open IoT operating system based on cloud computing, and is capable of systematic storage of information aggregates, etc., as described later. Some or all of the functions realized by the information storage platform may be incorporated into the intermediate information processing unit.

[0056] [4] Functions of information processing equipment As shown in Figure 2, the information processing device 6 includes a computer 10 having a processor 11. Figure 5 is a functional block diagram showing the information processing device 6 according to this embodiment. As shown in Figure 5, the information processing device 6 includes an information processing engine 31, a condition index adjustment engine 32, an inference value storage engine 33, a task generation engine 34, a knowledge information extraction engine 35, a knowledge information transmission engine 36, an information storage platform 37, and a database 38. The information processing device 6 implements each engine by executing a program. The database 38 may be a group of databases that handle a wide variety of information, or it may be a component of the information storage platform 37.

[0057] The information processing device 6 is communicatively connected to the information terminal 40. Examples of the information terminal 40 include a personal computer, a tablet device, and a smartphone. The information terminal 40 includes a display device such as a liquid crystal display or an organic EL display.

[0058] The multiple information processing devices 6 (61, 62, 63, 64, 65) each have, in one or more layers, an information processing engine 31, a condition index adjustment engine 32, an inference value storage engine 33, a task generation engine 34, a knowledge information extraction engine 35, a knowledge information transmission engine 36, an information storage platform 37, and a database 38. That is, each of the information processing engine 31, the condition index adjustment engine 32, the inference value storage engine 33, the task generation engine 34, the knowledge information extraction engine 35, the knowledge information transmission engine 36, the information storage platform 37, and the database 38 can be a functional unit in one or more layers of the multiple information processing devices 6 (61, 62, 63, 64, 65).

[0059] Furthermore, it is preferable that the information processing engine 31 be a functional unit of the edge computer 62 or gateway 63, while the condition index adjustment engine 32, inference value storage engine 33, task generation engine 34, knowledge information extraction engine 35, knowledge information transmission engine 36, information storage platform 37, and database 38 are functional units of the host computer 65.

[0060] <4-1> Information Processing Engine The information processing engine 31 performs predefined information processing on acquired environmental information and operational information. The information processing performed by the information processing engine 31 includes batch processing to adjust time-series environmental information and operational information to the required time granularity. The information processing performed by the information processing engine 31 also includes grouping processing to integrate multiple environmental information and operational information of different types into an information set linked to the hierarchical level of industrial activity.

[0061] (4-1-1) Generation process of real-time environmental information and real-time operational information The generation process for real-time environmental information and real-time operation information is performed in the information processing engine 31 of the lower-level information processing device. Alternatively, the generation process for real-time environmental information and real-time operation information may be performed in the information processing engine 31 of the intermediate information processing device or the information processing engine 31 of the higher-level information processing device.

[0062] The information processing engine 31 acquires and stores real-time environmental information from the environmental sensor 5 at predetermined sampling intervals. The information processing engine 31 has functions such as A / D conversion to convert analog signals from the environmental sensor 5 into digital signals, substitution of A / D values ​​with measured sample values, selection of measured sample values, moving average function of measured sample values, and period or frequency measurement function of pulse signals, and uses these functions to calculate confirmed measured values ​​in real time.

[0063] If the environmental sensor 5 is a temperature sensor, pressure sensor, or water level sensor, the information processing engine 31 reads the analog signal from the environmental sensor 5 at a predetermined sampling interval (approximately 10-25 ms), quantizes it, and converts it into a digital signal to calculate an A / D value. The information processing engine 31 processes the results of the most recent N samplings at each sampling timing, and takes the average of the A / D values ​​for M (=N-4) samples, excluding the first and second largest values ​​and the first and second smallest values, as the final A / D value. The information processing engine 31 generates temperature, pressure, or water level as real-time environmental information from the final A / D value by referring to a judgment table or using a calculation formula. The information processing engine 31 also determines whether the environmental sensor 5 is normal or abnormal according to the final A / D value.

[0064] If the environmental sensor 5 is a flow sensor, the information processing engine 31 measures the pulse width as the interval from one falling edge to the next falling edge of a pulse signal obtained by converting a sine wave signal to a square wave signal. The pulse width is counted with the period time of the count clock as 1 unit. The information processing engine 31 updates the pulse width at predetermined sampling intervals (approximately 100 ms) and calculates a sample value of the instantaneous flow rate by dividing the pulse constant (L / P: liters per pulse) by the pulse width. At each sampling timing, the information processing engine 31 processes the results of the most recent N samplings and generates the latest instantaneous flow rate as real-time environmental information from the average value of M (=N-2) samples excluding the maximum and minimum values. The information processing engine 31 detects the number of times the falling edge of the pulse signal is detected every second as the pulse count (P / s). At the pulse count update timing, the information processing engine 31 adds the flow rate obtained by multiplying the pulse count by the pulse constant to the previous cumulative flow rate to generate the latest cumulative flow rate as real-time environmental information.

[0065] If the environmental sensor 5 is an EC sensor, the information processing engine 31 calculates a confirmed A / D value using the same processing as for the temperature sensor, and calculates the electrical conductivity value by applying a predetermined calculation formula according to the range of the confirmed A / D value. Furthermore, the information processing engine 31 applies a correction coefficient to compensate for variations in individual sensor differences and temperature data measured by the temperature sensor to generate electrical conductivity converted to 25°C as real-time environmental information.

[0066] If the environmental sensor 5 is a power sensor (power monitor), the information processing engine 31 acquires instantaneous power sample values ​​from the power sensor at a predetermined sampling period (approximately 80-100 ms), adds up (sums up) all the sample values ​​for each sampling interval Δt for one period, and averages them over that period T to obtain real-time environmental information of instantaneous power [W]. When the instantaneous power for each sampling period is accumulated, it becomes real-time environmental information of the accumulated power. If one period is 1 second, the unit of the accumulated value is [W·s], and if this is multiplied by 3600, the unit becomes [W·h].

[0067] The information processing engine 31 acquires and stores real-time operation information from the controller 61 at predetermined sampling intervals. For example, the controller 61 transmits an "operating" signal when the industrial equipment 2 is operating normally and transmits an "abnormal stop" signal when it has stopped due to an abnormality. The information processing engine 31 uses the "operating" signal received via the input / output interface 14 to measure the uptime (actual operating time information) of the industrial equipment 2 in real time. The information processing engine 31 also uses the abnormal stop signal received via the input / output interface 14 to measure the downtime (non-operating time information) of the industrial equipment 2 in real time.

[0068] The time granularity of real-time environmental information and real-time operational information depends on the sampling interval or recording interval, and is generally quite fine (for example, the latest values ​​are updated at intervals of 10ms to 1s). When multiple types of real-time environmental information are used to calculate other information (real-time values ​​such as boiler efficiency, differential pressure, and permeate flux) at the time of updating the real-time environmental information, the information processing engine 31 also treats the other information as real-time environmental information. The information processing engine 31 transmits the real-time environmental information to the higher-level information processing device 6 along with an identification number (measurement item ID, device ID, location ID, etc.) and the update time.

[0069] (4-1-2) Batch Processing Real-time environmental information and real-time operational information are time-series data that are generated sequentially over time by the information processing engine 31. Batch processing refers to the process of adjusting the time-series real-time environmental information and real-time operational information to the required time granularity. Time granularity is an indicator that represents the degree of fineness of time, and can be selected from, for example, seconds, minutes, hours, or days. In the following explanation, the environmental information and operational information that have undergone batch processing may be referred to as "batch environmental information" and "batch operational information."

[0070] When batch processing is performed in the information processing engine 31 of the lower-level information processing device, the controller 61 includes a microcomputer 61A and a programmable logic controller 61B, so it is preferable to perform batch processing that takes into account the difference in processing capabilities. For example, the microcomputer 61A has high processing capabilities, which makes it possible to make the time granularity of real-time environmental information and real-time operational information finer than that of the programmable logic controller 61B. Therefore, it is preferable for the microcomputer 61A to perform batch processing in accordance with the programmable logic controller 61B, which has coarser time granularity.

[0071] When batch processing is performed in the information processing engine 31 of the intermediate information processing device, the edge computer 62 acquires real-time environmental information from the controller 61 at predetermined intervals, stores it for a predetermined period, performs batch processing, and processes it into batch environmental information with adjusted time granularity. The intermediate information processing device continuously collects and accumulates real-time environmental information, for example, at 1-second intervals. If the real-time environmental information is temperature, pressure, or instantaneous flow rate, the information processing engine 31 of the intermediate information processing device calculates an average value by dividing the accumulated value of the real-time environmental information by the number of accumulations when a predetermined period is reached, and uses the calculated average value as the batch environmental information. If the real-time environmental information is the accumulated flow rate or power consumption per unit time period, the information processing engine 31 of the intermediate information processing device uses the accumulated value of the real-time environmental information as the batch environmental information. When the information processing engine 31 of the intermediate information processing device processes the accumulated real-time environmental information into batch environmental information, it also treats other information (COP, specific energy, etc.) calculated using multiple types of batch environmental information as batch environmental information.

[0072] Furthermore, when batch processing is performed in the information processing engine 31 of the higher-level information processing device instead of the intermediate information processing device, real-time environmental information transmitted from the information processing engine 31 of the lower-level information processing device to the information processing engine 31 of the intermediate information processing device is acquired at predetermined intervals, stored for a predetermined period, and then batch processed to create batch environmental information with adjusted time granularity. The content of the batch processing can be the same as when it is performed in the information processing engine 31 of the intermediate information processing device.

[0073] Batch operation information is, for example, real-time operation information that has been thresholded at the required time granularity. For example, this could involve thresholding the combustion stage of a steam boiler at predetermined intervals, or thresholding the rotational load rate of an air compressor at predetermined intervals.

[0074] The time granularity of batch environment information and batch operation information should preferably be at a level that can be used at least for calculating or evaluating operational performance values, as described later, in the information utilization process associated with providing services to customers. For example, it can be selected from 0.5 hours, 1 hour, 2 hours, 6 hours, 12 hours, or 24 hours. Batch environment information and batch operation information with this level of time granularity can also be used for optimization diagnostic simulations of industrial machinery or production equipment (such as reviewing control setpoints and control patterns, and predicting the effects of equipment modifications). In addition, the calculation or evaluation of environmental impacts (such as carbon dioxide emissions and hazardous substance emissions) may be included in the calculation or evaluation of energy performance values.

[0075] (4-1-3) Grouping process Grouping refers to the process of integrating multiple types of environmental and operational information (real-time environmental information, batch environmental information, real-time operational information, batch operational information) into an information set linked to the hierarchical level of industrial activity.

[0076] As shown in Figure 4, the hierarchical levels of industrial activity include one or more levels from among the following: individual machine level, machine group level, cell level, line level, building level, and establishment level.

[0077] The individual machine level refers to a level where operational performance values ​​are managed, for example, for individual industrial equipment such as steam boilers, heat pumps, and air compressors.

[0078] The machine group level is a level that manages operational performance values, for example, when multiple industrial equipment units such as steam boilers, heat pumps, and air compressors are installed.

[0079] The cell level is the level used to manage the operational performance of distributed industrial equipment 2 attached to one or more cells in a site employing a cell production system.

[0080] The line level is the level used to manage the operational performance of distributed industrial equipment 2 attached to one or more lines (such as a container molding line or beverage filling and sterilization line in a beverage manufacturing plant) in a workplace employing a line production system.

[0081] The building level is the level at which the operational performance values ​​of industrial equipment 2 are managed, targeting the individual building units (units of Factory 1 4A and Factory 2 4B) dispersed within the site of Business Establishment 3.

[0082] The establishment level is the level at which operational performance values ​​of industrial equipment 2 are managed, covering the entire site of establishment 3 (both Factory 1A and Factory 2B).

[0083] When grouping processing is performed in the information processing engine 31 of the higher-level information processing device (host computer 65), the information processing engine 31 of the higher-level information processing device picks up multiple batch environment information and batch operation information necessary for calculating or evaluating operational performance values ​​from the diverse batch environment information and batch operation information generated by the edge computer 62 and performs grouping processing. The information set that has undergone this grouping processing is stored in the information storage platform for the period necessary for history management (for example, the past 12 months). Alternatively, the information set may be one which has been picked up and grouped the multiple batch environment information and batch operation information necessary for demand management of utilities (hot water, steam, compressed air, treated water) of the demand equipment 8.

[0084] When grouping processing is performed in the information processing engine 31 of the intermediate information processing device (edge ​​computer 62), the information processing engine 31 of the edge computer 62 can perform the grouping processing while simultaneously performing the batch processing described above. The content of the grouping processing can be the same as that performed in the information processing engine 31 of the higher-level information processing device.

[0085] The time granularity of the batch environment information and batch operation information that constitute the information set is basically consistent, and each information set is accompanied by information on the storage period of the real-time environment information that is based on the batch environment information and the real-time operation information that is based on the batch operation information.

[0086] The following describes an example of grouping processing. The information processing engine 31 of the higher-level information processing device performs grouping processing of batch environment information for the purpose of energy management of the heat recovery air compressor, demand management of the heat recovery air compressor, energy management of the heat pump, demand management of the heat pump, energy management of the steam boiler, and demand management of the steam boiler. The batch environment information that has undergone grouping processing is stored in the information storage platform 37. The time granularity of the batch environment information subject to grouping processing is, for example, in units of 30 minutes or 1 hour.

[0087] (4-1-3-1) Energy management of heat recovery type air compressors A heat recovery air compressor is equipped with a heat recovery heat exchanger that recovers the heat of compression contained in compressed air and lubricating oil to generate hot water from the cooling water. The information processing engine 31 of the intermediate information processing device acquires real-time environmental information corresponding to each sensor from a lower-level information processing device to which temperature sensors, flow sensors, and power sensors are connected, and generates batch environmental information. The batch environmental information includes the discharge air volume [m³ 3 [kWh], power consumption [kWh], average inlet water temperature [°C], average outlet water temperature [°C], and cumulative hot water volume [m³] 3 This includes [ ]. The higher-level information processing device generates an information set by grouping these multiple types of batch environment information. The information processing engine 31 of the higher-level information processing device uses the batch environment information integrated into the information set to process the specific energy [kW / m 3 Calculate the energy recovery rate [%].

[0088] (4-1-3-2) Demand management of heat recovery air compressors The information processing engine 31 of the intermediate information processing device acquires real-time environmental information corresponding to the sensors, etc., from the lower-level information processing device to which the flow sensor and the contact signal output of the air supply valve are connected, and generates batch environmental information. The batch environmental information includes the discharge air volume [m³ 3 [h], air consumption [m³] 3 This includes the [h] batch environment information and the cumulative open time [h] of the air supply valve. The information processing engine 31 of the higher-level information processing device generates an information set by grouping these multiple types of batch environment information. The information processing engine 31 of the higher-level information processing device calculates the operating rate [%] of the demand equipment using the batch environment information integrated into the information set. The information processing engine 31 of the higher-level information processing device also diagnoses whether there is any leak loss in the air transport piping network by comparing the discharge air amount and consumption air amount integrated into the information set. The air supply valve is installed at the end of the air transport piping and is opened when compressed air is used by the load equipment.

[0089] (4-1-3-3) Energy management of heat pumps The machinery and equipment subject to management include air-source heat pumps and water-source heat pumps. The information processing engine 31 of the intermediate information processing device acquires real-time environmental information corresponding to each sensor from lower-level information processing devices to which temperature sensors, flow sensors, and power sensors are connected, and generates batch environmental information. The batch environmental information includes the average inlet water temperature [°C], the average outlet water temperature [°C], and the cumulative amount of hot water [m³]. 3 This includes [various types of batch environment information], power consumption [kWh], and average heat source temperature [°C]. The information processing engine 31 of the higher-level information processing device generates an information set by grouping these multiple types of batch environment information. The information processing engine 31 of the higher-level information processing device calculates the coefficient of performance (COP) and heat supply amount [W] using the batch environment information integrated into the information set.

[0090] (4-1-3-4) Demand management of heat pumps When using hot water generated by a heat pump at a demand facility, there are two modes of operation: sequentially supplying hot water to the demand facility and circulating hot water to the demand facility. In the former mode, the hot water itself is consumed, generating waste hot water. In the latter mode, only the thermal energy of the hot water is consumed, and the hot water used for heat utilization is recirculated. The information processing engine 31 of the intermediate information processing device acquires real-time environmental information corresponding to each sensor from a lower-level information processing device to which temperature sensors and flow rate sensors are connected, and generates batch environmental information. In the mode of hot water consumption at the demand facility, the batch environmental information includes the average supply temperature [°C], the average wastewater temperature [°C], and the average hot water supply flow rate [m³]. 3 This includes [ / h], and in the case of hot water consumption by the demand facility, the average supply temperature [°C], average return temperature [°C], and average circulation flow rate [m 3 This includes [h], etc. The information processing engine 31 of the higher-level information processing device generates an information set by grouping these multiple types of batch environment information. The information processing engine 31 of the higher-level information processing device calculates the heat consumption [W] using the batch environment information integrated into the information set.

[0091] (4-1-3-5) Energy Management of Steam Boiler The information processing engine 31 of the intermediate information processing device acquires real-time environment information corresponding to each sensor from a lower-level information processing device to which a temperature sensor, a pressure sensor, a flow sensor, etc. are connected, and generates batch environment information. The batch environment information includes the average header supply air pressure [MPa], the average feed water temperature [°C], the integrated steam supply amount [m 3 , the integrated fuel consumption amount [m 3 , the average boiler efficiency [%], the average blowdown rate [%], etc. The boiler efficiency and the blowdown rate are calculated based on multiple types of real-time environment information. The information processing engine 31 of the upper-level information processing device generates an information aggregate obtained by grouping these multiple types of batch environment information. The information processing engine 31 of the upper-level information processing device calculates the heat supply amount [W] and the energy efficiency [%] using the batch environment information integrated in the information aggregate. Note that since the heat gain of boiler feed water by drain recovery or heat pump heating and the heat loss by concentrated blowdown affect the fuel consumption amount and change the apparent boiler efficiency, the boiler efficiency and the blowdown rate are integrated into the information aggregate as relevant batch environment information.

[0092] (4-1-3-6) Demand Management of Steam Boiler The information processing engine 31 of the intermediate information processing device acquires real-time environment information corresponding to each sensor from a lower-level information processing device to which a temperature sensor, a pressure sensor, a flow sensor, etc. are connected, and generates batch environment information. The batch environment information includes the average terminal supply steam pressure [MPa], the average terminal exhaust steam pressure [MPa], and the integrated steam arrival amount [m 3 , etc. The information processing engine 31 of the upper-level information processing device calculates the heat consumption amount [W] and the heat arrival amount [W] using the batch environment information integrated in the information aggregate. Note that if the batch environment information related to each of them is integrated prior to the calculation of the heat supply amount, the heat consumption amount, and the heat delivery amount, the heat dissipation loss can be evaluated by calculating the difference between the heat supply amount and the heat delivery amount.

[0093] <4-2> Information Storage Platform The information storage platform 37 stores various types of information, such as numerical data, images, and documents, in the storage device 12, and also provides a foundational environment for operating software (engines, applications) and hardware. The information storage platform 37 is composed of, for example, an operating system and a database, and centrally manages a wide variety of big data, enabling smooth processing and utilization of information. The information storage platform 37 has the function of converting different data formats at each hierarchical level into a unified format.

[0094] The information storage platform 37 stores environmental information and operational information in the storage device 12 so that the environmental sensor 5, industrial equipment 2, detection date and time, etc. can be identified. The information storage platform 37 stores environmental information and operational information that have undergone prescribed information processing by the information processing engine 31, as well as various registration information registered via the input device, in the storage device 12. Furthermore, the information storage platform 37 may be configured to reproduce the operating state and operating environment of the industrial equipment 2 that exists in physical space as a digital twin in virtual space using the stored environmental information and operational information.

[0095] Figure 6 shows an example of an information storage platform 37 according to the embodiment. In the example shown in Figure 6, the information storage platform 37 stores environmental information and operational information that have undergone specified information processing by the information processing engine 31, linking them together. The information storage platform 37 stores multiple types of environmental information and operational information (real-time environmental information, batch environmental information, real-time operational information, batch operational information) linked to the hierarchical level of industrial activity. In the example shown in Figure 6, the hierarchical levels of industrial activity include the machine level, machine group level, cell level, line level, building level, office level, etc.

[0096] <4-3> Condition Index Adjustment Engine Returning to Figure 5, the condition index adjustment engine 32 extracts or generates condition indexes for the industrial equipment 2 in a time series from at least one of the environmental information and operational information stored in the information storage platform 37. Condition indexes are pre-set indicators that can affect the operational performance values ​​of the industrial equipment 2. Operational performance values ​​include, for example, operational performance values, energy performance values, quality performance values, etc.

[0097] The operational performance value is an indicator value that shows the operational performance of industrial equipment 2. The higher the operational performance value, the less frequent the downtime due to malfunctions, and the more positively it impacts the user's main business (production of goods, provision of services) (increased production efficiency, increased service provision efficiency).

[0098] The operational performance value is determined by the ratio of actual operating time (the sum of operating time and standby time (output standby, warm-up time, etc.)) associated with the normal state of industrial equipment 2 (machinery and equipment) during the operating hours of factory 4, and non-operating time (the sum of abnormal stop time and recovery time) associated with abnormal conditions (downtime). If multiple units of the same type of machine or equipment are installed, there are cases where the operational performance value is calculated for each unit, and cases where the operational performance value is calculated for the entire group. In the latter case, if at least one unit is running, it is considered to be in operation, and the operational performance value is calculated accordingly. The operational performance value is calculated based on the following (Equation 1). The denominator, the sum of actual operating time and non-operating time, is also called load time. Operating performance [%] = Actual operating time / (Actual operating time + Non-operating time) × 100 …(Equation 1)

[0099] Energy performance refers to an index value that indicates the energy performance of industrial equipment. A higher energy performance value indicates less energy waste and a more positive impact on the user's main business (production of goods, provision of services) (reduction of energy costs, reduction of carbon dioxide emissions).

[0100] Energy performance values ​​are indicators showing the operational performance of industrial equipment 2 (utility conversion machinery and equipment), and, with some exceptions, are basically expressed as a ratio in which the input amount of the primary utility is the denominator and the output amount of the secondary utility is the numerator. Energy performance values ​​include energy conversion efficiency [%], energy recovery efficiency [%], and media generation efficiency [m 3 Examples include [J] and energy transport efficiency [%].

[0101] Quality performance refers to an index value that indicates the quality performance of industrial equipment 2. A higher quality performance value indicates better quality and a more positive impact on the user's main business (production of goods, provision of services) (continuity of business activities through the production of high-quality products, sterilized items, and cleaned items).

[0102] Examples of quality performance values ​​include the quality performance values ​​for steam supply pressure, hot or chilled water supply temperature, and compressed air supply pressure. The quality performance value [%] for steam supply pressure is calculated as A / (A+B)×100, where A is the boiler operating time when the supply pressure is greater than or equal to the required pressure, and B is the boiler operating time when the supply pressure is less than the required pressure, during the predetermined operating period of the steam boiler. The quality performance value [%] for hot water supply temperature is calculated as A / (A+B)×100, where A is the heat pump operating time when the supply temperature is greater than or equal to the required temperature, and B is the heat pump operating time when the supply temperature is less than the required temperature, during the predetermined operating period of the heat pump. The quality performance value [%] for chilled water supply temperature is calculated as A / (A+B)×100, where A is the chiller operating time when the supply temperature is less than or equal to the required temperature, and B is the chiller operating time when the supply temperature is greater than the required temperature, during the predetermined operating period of the chiller. The quality performance value [%] of the compressed air supply pressure is calculated as A / (A+B)×100, where A is the time the air compressor is operating when the supply pressure is greater than or equal to the required pressure, and B is the time the air compressor is operating when the supply pressure is less than the required pressure, during the air compressor's predetermined operating period.

[0103] (4-3-1) Condition Indicators The condition index for industrial equipment 2 is a measure of the performance of industrial equipment 2 that can be identified and expressed as a physical quantity or a coefficient based on a physical quantity. A physical quantity is, for example, a quantity of the state of a physical object (89 quantities) as defined by the Measurement Law. An index is a marker for judging or evaluating things, and includes, for example, observed values ​​of variables. The condition index can be expressed as a physical quantity or coefficient categorized as follows: [a] Physical quantities and coefficients that indicate current performance: Decay or regression from the initial performance value represents deterioration or alteration of the equipment. [b] Physical quantities and coefficients indicating the operating state: Deviations from reference values ​​or deviations from the reference range indicate non-conforming operation of the equipment. [c] Physical quantities and coefficients indicating operating conditions: The magnitude of the output value represents the high or low load state of the equipment (overload or light load). [d] Physical quantities and coefficients indicating the operating environment: High or low ambient temperature indicates the severity of the equipment's operating environment (overheating or overcooling).

[0104] <4-4> Inference Value Accumulation Engine The inference value storage engine 33 has a trained inference model 330 and, by inputting condition indicators, infers the probability of a malfunction event occurring and stores the obtained inferred probability value in the information storage platform 37. That is, the inference value storage engine 33 inputs condition indicators to the trained inference model 330 and stores the inferred probability value of the malfunction event occurring output by the trained inference model 330 in the information storage platform 37. In this embodiment, as shown in Figure 6, the inference value storage engine 33 links the condition indicators input to the trained inference model 330 with the inferred probability value of the malfunction event occurring and stores them in the information storage platform 37. The inference value storage engine 33 performs inference of the probability of occurrence using the condition indicators at that time, for example, every hour.

[0105] (4-4-1) Pre-trained inference models Returning to Figure 5, the trained inference model 330 can be generated using the logistic regression algorithm. The logistic regression algorithm is an algorithm that expresses binary data consisting of 0 and 1, or data such as probabilities consisting of values ​​between 0 and 1, using an equation with explanatory variables. In other words, the logistic regression algorithm can be used to predict the probability of a certain event occurring using explanatory variables. The explanatory variables, which are the inputs to the trained inference model 330, include, for example, pre-selected condition indicators that may affect operational performance. The dependent variable, which is the output of the trained inference model 330, is, for example, an inferred value of one of the following: the probability of abnormal shutdown, the probability of efficiency deterioration, or the probability of quality defects.

[0106] In this embodiment, the trained inference model 330 is described as being generated using a logistic regression algorithm, but it is not limited to this. The trained inference model 330 may also be generated using other algorithms, such as a neural network or linear regression.

[0107] The trained inference model 330 includes one or more of the first inference model 331, the second inference model 332, and the third inference model 333. In this embodiment, the case in which the trained inference model 330 includes the first inference model 331, the second inference model 332, and the third inference model 333 will be described.

[0108] Figure 7 shows an example of a trained inference model 330 according to the embodiment. As shown in Figure 7, the first inference model 331 is an inference model that infers the probability of an abnormal shutdown occurring, which is a state in which the operation of the industrial equipment 2 is stopped due to the occurrence of a malfunction. The second inference model 332 is an inference model that infers the probability of a deterioration in efficiency occurring, which is a state in which the industrial equipment 2 does not meet the required energy efficiency due to the occurrence of a malfunction. The third inference model 333 is an inference model that infers the probability of a quality defect occurring, which is a state in which the industrial equipment 2 does not meet the required quality for industrial activities due to the occurrence of a malfunction.

[0109] (4-4-2) First Inference Model The first inference model 331 is a logistic regression algorithm and is an inference model generated by machine learning using training data that uses the condition indicators of industrial equipment 2 as features and the presence or absence of abnormal stoppages as training labels. The operational information includes actual operating time information linked to the normal state of industrial equipment 2 and non-operating time information linked to the abnormal state. In this case, the probability of abnormal stoppages occurring is an influencing factor for the operational performance value based on the actual operating time information and non-operating time information.

[0110] The first inference model 331 is an inference model generated by machine learning using the following training data. The training data uses information on the silica concentration of boiler water in a steam boiler (estimated from the amount of chemical injected and the equivalent high-burning time) as a feature, and the presence or absence of pitting corrosion punctures as the training label. • Training data using permeation flux information (water permeation coefficient) from an RO membrane system as a feature, and the presence or absence of water quality abnormalities as the training label. The training data uses internal temperature information (such as the surface temperature of the heat sink or power module) of inverters in air compressors and heat pumps as features, and the presence or absence of overheating abnormal shutdown is used as the training label.

[0111] (4-4-3) Second Inference Model The second inference model 332 is a logistic regression algorithm and is an inference model generated by machine learning using training data that uses the condition indicators of industrial equipment 2 as features and the presence or absence of efficiency deterioration as the training label. The environmental information includes input energy information and output energy information for industrial equipment 2. In this case, the probability of efficiency deterioration occurring is an influencing factor for the energy performance value based on the input energy information and output energy information.

[0112] The second inference model 332 is an inference model generated by machine learning using the following training data. • Training data using heat drop information in a heat pump or wet-bulb temperature information of the outside air as features, with the presence or absence of COP deterioration as the training label. • Training data using temperature information from a scale monitor in a steam boiler as a feature, and the presence or absence of boiler efficiency degradation as the training label. • Training data using steam trap failure rate information in steam transport piping as a feature, and the presence or absence of increased energy loss as the training label. The training data uses pressure loss information of filters (suction filter, oil separator element, air filter) in an air compressor as a feature, and the presence or absence of specific energy degradation as the training label.

[0113] (4-4-4) Third Inference Model The third inference model 333 is a logistic regression algorithm and is an inference model generated by machine learning using training data that uses the condition indicators of industrial equipment as features and the presence or absence of quality defects as training labels. The environmental information includes actual quality level information of the media or goods produced by industrial equipment 2. In this case, the probability of quality defects occurring is an influencing factor for the quality performance value based on actual quality level information and required quality level information.

[0114] The third inference model 333 is an inference model generated by machine learning using the following training data. • Training data using ambient temperature information in an air compressor as a feature, and whether or not a supply pressure shortage occurs as the training label. • Refrigerant leak information in the heat pump is used as a feature, and the presence or absence of insufficient hot water temperature is used as training data for the label. The training data uses differential pressure information (degree of blockage recovery) immediately after membrane cleaning in a membrane filtration system as a feature, and the presence or absence of insufficient processing flow rate as the training label.

[0115] (4-4-5) Probability of occurrence The type of probability of occurrence can be selected depending on the condition indicator and the type of malfunction that may occur. When adding explanatory variables such as duration, cumulative duration, number of occurrences, and frequency of occurrence to the inference model, these are also added to the features of the training data to obtain a multi-axis regression equation.

[0116] (4-4-5-1) Duration type Duration-based probability is the probability that an abnormal shutdown, efficiency degradation, or quality defect will occur if a certain condition indicator persists for a predetermined period of time. Duration is added as an explanatory variable to the inference model.

[0117] (4-4-5-2) Cumulative time type The cumulative time-type probability of occurrence is the probability that an abnormal shutdown, efficiency deterioration, or quality defect will occur when a certain condition indicator accumulates for a predetermined time, and the cumulative time is added as an explanatory variable to input into the inference model.

[0118] (4-4-5-3) Occurrence frequency type The occurrence count type of probability is the probability that an abnormal shutdown, efficiency deterioration, or quality defect will occur if a certain condition indicator occurs a predetermined number of times. The number of occurrences is added as an explanatory variable to input into the inference model.

[0119] (4-4-5-4) Frequency type The frequency-based probability of occurrence is the probability that an abnormal shutdown, efficiency degradation, or quality defect will result if a certain condition indicator occurs a predetermined number of times within a predetermined period. The frequency of occurrence is added as an explanatory variable to input into the inference model.

[0120] (4-4-6) Others It is preferable to prepare the same inference model 330 if the industrial equipment 2 is from the same manufacturer and belongs to the same category. However, even if the equipment is from the same manufacturer and belongs to the same category, if the probability of failure occurring for the same condition indicator changes due to differences in design philosophy, multiple inference models may be prepared. Similarly, if the industrial equipment 2 is from a different manufacturer, it is likely that the design philosophy will be different, so a different inference model may be prepared.

[0121] The inference value storage engine 33 basically applies the first inference model 331, the second inference model 332, and the third inference model 333 to each piece of equipment in the same category of industrial equipment 2 to infer the probability of abnormal shutdown, the probability of efficiency deterioration, and the probability of quality defects. Note that, in principle, only one type of condition indicator is input to the trained inference model 330, so as the number of condition indicators to be monitored increases, the number of inference models to be applied also increases.

[0122] In the inference value storage engine 33, the number of trained inference models 330 applied remains the same regardless of whether one or multiple units of the same category of equipment are installed in the industrial equipment 2. Since the trained inference models 330 are regression models, high-speed processing is possible, and in the case of multiple units, they can be processed sequentially starting from unit 1 at a predetermined time.

[0123] The inference value storage engine 33 does not necessarily require a pre-trained inference model 330 for each factory or customer business unit. The inference value storage engine 33 can process the data sequentially even if the number of monitored industrial equipment 2 increases. However, if the number of monitored industrial equipment 2 increases beyond a certain point, the inference value storage engine 33 may be configured to add the same inference model and perform parallel processing.

[0124] <4-5> Task generation engine The task generation engine 34 generates maintenance recommendation tasks based on the results of data analysis using time-series data of inferred probability values ​​stored in the information storage platform 37. The task generation engine 34 generates maintenance recommendation tasks, for example, by turning on a flag that is controlled on / off. The following are examples of data analysis methods, including threshold judgment, first trend judgment, and second trend judgment.

[0125] (4-5-1) Threshold determination The task generation engine 34 generates a maintenance recommendation task immediately or at a specified time when the latest occurrence probability exceeds a threshold. For example, if the latest occurrence probability of instantaneous value, moving average, or moving trimmed average, harmonic average, or geometric average exceeds a threshold of 75-80%, the task generation engine 34 generates a maintenance recommendation task immediately or at a specified time such as midnight. Note that if the threshold is too high, there may not be enough time to complete the action.

[0126] (4-5-2) First trend judgment The task generation engine 34 calculates the rate of increase from the probability a on day 1 to the probability b on day 5 if the probability of occurrence is on an upward trend, and generates a maintenance recommendation task immediately or at a specified time if the rate of increase exceeds a threshold. For example, if the probability exceeds the previous day's probability for five consecutive days, the task generation engine 34 calculates the rate of increase (=(ba) / a × 100), and generates a maintenance recommendation task immediately or at a specified time such as midnight if the rate of increase exceeds a threshold of 5-10%. The task generation engine 34 determines whether the probability of occurrence is on an upward trend by using the maximum value for each day as a representative value. The probability a and b used when calculating the rate of increase are the maximum values ​​for each day as representative values.

[0127] (4-5-3) Second trend judgment The task generation engine 34, when the probability of occurrence is on an upward trend, predicts the deadline date when the probability of occurrence will reach an acceptable limit using a linear regression model, and generates a maintenance recommendation task at a specified time on the day obtained by subtracting a predetermined number of grace days from the deadline date. If the probability of occurrence is on an upward trend, exceeding the previous day's probability for five consecutive days, the task generation engine 34 predicts the deadline date when the probability of occurrence will reach an acceptable limit of, for example, 95-100%, using a linear regression model. Then, the task generation engine 34 generates a maintenance recommendation task at a specified time, such as midnight, on the day obtained by subtracting, for example, a 30-day period during which treatment can be provided from the deadline date. The task generation engine 34 determines whether or not the probability of occurrence is on an upward trend by using the maximum value for each day as a representative value. The linear regression model is created using the representative value of the probability of occurrence for the five days used to determine the upward trend.

[0128] (4-5-4) Others The task generation engine 34 performs data analysis on the probability of failure inferred for each condition indicator as the number of trained inference models 330 applied to the inference value storage engine 33 increases. This is because each condition indicator has a different degree of impact on operational performance.

[0129] Data analysis (symptom diagnosis) by the task generation engine 34 should be performed on a per-device basis when multiple industrial equipment 2 units are installed. This is because a quick response is possible if even one task is generated. If the maximum value of the occurrence probability among multiple industrial equipment 2 units is extracted and data analysis is performed only on the maximum value, there is a risk of missing symptoms in the second or third highest probability units. Also, if data analysis is performed on the average or median of the occurrence probability among multiple industrial equipment 2 units, there is a risk of delayed detection of symptoms if there is a bias in the occurrence probability.

[0130] <4-6> Database Figure 8 shows an example of a database 38 according to the embodiment. As shown in Figure 8, the database 38 stores knowledge information 380 regarding an appropriate range of condition indicators that can suppress the probability of malfunction events to a predetermined probability or less, and maintenance activities that are effective in bringing condition indicators that have fallen outside the appropriate range back into the appropriate range.

[0131] The increased probability of a malfunction event resulting in the generation of a maintenance recommendation task is due to the condition indicator falling outside the appropriate range. Therefore, Knowledge Information 380 includes information that verbalizes the procedures for effective corrective actions to bring the condition indicator back within the appropriate range. The message text of Knowledge Information 380 transmitted to information terminals, etc., may be standardized according to the corrective action. The corrective actions may include those performed voluntarily by service personnel under a service contract, and those performed after being proposed to and approved by the equipment user.

[0132] (4-6-1) Details of repair / modification procedures Knowledge information 380 regarding the details of repair procedures contains text information such as the following example messages A to D.

[0133] Knowledge Information 380 includes standard phrases as Message Example A, such as, "Please refer to [Document or Video Procedure Name] and conduct or propose a damage investigation of [Piping Equipment Name]." Message Example A is used, for example, when instructing tasks such as searching for steam trap failures in steam transport piping or searching for air leaks in pneumatic transport piping.

[0134] Knowledge Information 380 includes standard phrases as Message Example B, such as, "Refer to [Document or Video Procedure Name] and replace or repair [Functional Part Name]." Message Example B is used when instructing tasks such as replacing the membrane element of an RO membrane system, injecting a descaling agent into the boiler feedwater, or repairing a leak in the heat pump refrigerant circuit.

[0135] Knowledge Information 380 includes standard phrases as Message Example C, such as, "Refer to [Document or Video Procedure Name] and replace or clean [Auxiliary Part Name]." Message Example C is used when instructing tasks such as cleaning the inverter heat dissipation fins of an air compressor, cleaning the condenser of a heat pump (removing dirt from the heat transfer surface and blockages in the flow path), and replacing the pre-filter or post-filter of a membrane filtration system.

[0136] Knowledge Information 380 includes standard phrases as Message Example D, such as, "Please replace or replenish [Consumable Name] by referring to [Document or Video Procedure Name]." Message Example D is used when instructing tasks such as changing the lubricating oil in an air compressor, replenishing regenerating agents in an ion exchange unit, or replenishing water treatment chemicals in a boiler chemical injection unit.

[0137] (4-6-2) Actions taken regarding operational changes Knowledge information 380 regarding operational change procedures contains text information such as the following example message E.

[0138] Knowledge Information 380 includes standard phrases as Message Example E, such as, "Please refer to [Document or Video Procedure Name] and implement or propose a review of the [Setting Value for XX]." Message Example E is applied when instructing tasks such as reviewing the hot water outlet temperature setting of a heat pump, reviewing the discharge pressure setting of an air compressor, reviewing the injection amount setting of water treatment chemicals to a steam boiler, and reviewing the priority setting of equipment number control (such as prioritizing the operation of high-efficiency machines).

[0139] <4-7> Knowledge Information Extraction Engine When a maintenance recommendation task is generated by the task generation engine 34, the knowledge information extraction engine 35 identifies the condition indicator at that time and extracts knowledge information 380 corresponding to the identified condition indicator from the database 38. The knowledge information extraction engine 35 stores the extracted knowledge information 380 in the storage device 12, linking it to the industrial equipment 2.

[0140] <4-8> Knowledge Information Transmission Engine The knowledge information transmission engine 36 converts the extracted knowledge information 380 into a message and transmits it to an information terminal usable by maintenance workers or equipment users in the format of the message examples A to E described above. The knowledge information transmission engine 36 transmits a message containing the extracted knowledge information 380 to the information terminal 40 via the communication interface 13. The information terminal includes, for example, a smartphone, tablet terminal, or personal computer. The information processing device 6 also provides a user interface (UI) for displaying the knowledge information 380 on the information terminal 40.

[0141] [5] Maintenance based on detection of abnormal signs in industrial equipment Maintenance methods for industrial equipment 2 include after-maintenance, which is performed after an abnormality occurs in industrial equipment 2, and before-maintenance, which is performed before an abnormality occurs in industrial equipment 2. The industrial equipment operation support system 1 detects signs of abnormality in industrial equipment 2 and enables before-maintenance. Maintenance is one example of activities for industrial equipment 2.

[0142] Figure 9 is a flowchart illustrating the maintenance method of the industrial equipment operation support system 1 according to this embodiment. As shown in Figure 9, the industrial equipment operation support system 1 uses an information processing engine 31 to perform predetermined information processing on environmental information and operational information (step S11). The information storage platform 37 then stores the environmental information and operational information that have undergone the predetermined information processing by the information processing engine 31.

[0143] The industrial equipment operation support system 1 extracts or generates condition indicators for the industrial equipment 2 in a time series using the condition indicator adjustment engine 32 (step S12). For example, the condition indicator adjustment engine 32 extracts or generates condition indicators for the industrial equipment 2 in a time series from at least one of the environmental information and operational information stored in the information storage platform 37. In detail, the condition indicator adjustment engine 32 extracts or generates condition indicators that are expressed in physical quantities or coefficients that indicate, for example, current performance, operating state, operating conditions, operating environment, etc., and provides the condition indicators to the inference value storage engine 33.

[0144] The industrial equipment operation support system 1 uses an inference value storage engine 33 to infer the probability of a malfunction occurring by inputting condition indicators, and stores the obtained inferred probability values ​​in the information storage platform 27 (step S13). Specifically, the inference value storage engine 33 inputs the condition indicators as explanatory variables into a trained inference model 330, and the trained inference model 330 outputs an inferred value as the target variable for one of the following: the probability of abnormal shutdown, the probability of efficiency deterioration, or the probability of quality defects. The inference value storage engine 33 links the condition indicators input into the trained inference model 330 with the inferred probability values ​​of the malfunction occurrence and stores them in the information storage platform 37.

[0145] The industrial equipment operation support system 1 generates maintenance recommendation tasks (step S14) based on the results of data analysis using time-series data of inferred occurrence probabilities stored in the information storage platform 37, via the task generation engine 34. For example, the task generation engine 34 generates a maintenance recommendation task immediately or at a specified time if the latest occurrence probability exceeds a threshold. For example, if the occurrence probability is on an upward trend, the task generation engine 34 calculates the rate of increase of the occurrence probability after a predetermined number of days relative to the occurrence probability on the first day, and generates a maintenance recommendation task immediately or at a specified time if the rate of increase exceeds a threshold. For example, if the occurrence probability is on an upward trend, the task generation engine 34 predicts the deadline date on which the occurrence probability will reach an acceptable limit using a linear regression model, and generates a maintenance recommendation task at a specified time on the day obtained by subtracting a predetermined grace period from the deadline date.

[0146] The industrial equipment operation support system 1 uses a knowledge information extraction engine 35 to identify the condition indicator at the time the maintenance recommendation task was generated, and extracts knowledge information 380 corresponding to the identified condition indicator from the database 38 (step S15). For example, the knowledge information extraction engine 35 identifies the condition indicator at the time the maintenance recommendation task was generated by the task generation engine 34 from the information storage platform 37, and extracts knowledge information 380 corresponding to that condition indicator from the database 38.

[0147] The industrial equipment operation support system 1 uses a knowledge information transmission engine 36 to create a message from the extracted knowledge information 380 and send it to an information terminal available to maintenance workers or equipment users (step S16). For example, the knowledge information transmission engine 36 sends a message containing the extracted knowledge information 380 to the information terminal via the communication interface 13. As a result, the industrial equipment operation support system 1 can enable maintenance of the industrial equipment 2 by detecting signs of abnormality, by displaying the knowledge information 380 received by the information terminal to the maintenance worker or equipment user.

[0148] [6] Operation of the industrial equipment operation support system Figure 10 is a schematic diagram showing an example of the operation of the industrial equipment operation support system 1 according to the embodiment.

[0149] As shown in Figure 10, the industrial equipment 2, which is subject to maintenance by the maintenance company, is installed at the business premises 3. The maintenance company provides various maintenance services to the industrial equipment 2 through the industrial equipment operation support system 1.

[0150] The information processing device 6 performs information processing to support maintenance operations. The information processing device 6 manages the workers 60 (service technicians) who perform maintenance. The information processing device 6 provides maintenance-related knowledge information 380 to the information terminal 40 held by the workers 60 via a communication line. The information terminal 40 may also be able to access the information processing device 6 via a communication line.

[0151] The information processing device 6 manages items 610 used for maintenance. Items 610 include parts for industrial equipment 2 and tools used during operation. The items 610 are managed by the warehouse where the parts and tools are stored.

[0152] The information processing device 6 manages the content 620 used for maintenance. Examples of content 620 include document-based manuals (procedures, instructions, etc.), video-based manuals, and past maintenance reports.

[0153] The information processing device 6 performs predefined information processing on environmental information and operational information acquired from multiple environmental sensors 5 and stores it in the information storage platform 37. The information processing device 6 extracts or generates condition indicators for the industrial equipment 2 in a time series from the environmental information and operational information stored in the information storage platform 37. The information processing device 6 inputs the condition indicators into a trained inference model 330 to infer the probability of malfunction events occurring and stores the obtained inferred values ​​of occurrence probabilities in the information storage platform 37. Based on the results of data analysis using the time series data of inferred values ​​of occurrence probabilities stored in the information storage platform 37, the information processing device 6 generates maintenance recommendation tasks.

[0154] The database 38 stores knowledge information 380 regarding the appropriate range of condition indicators that can suppress the probability of malfunction events to a predetermined probability or less, and maintenance activities that are effective in bringing condition indicators that have fallen outside the appropriate range back within the appropriate range. The knowledge information 380 includes, for example, information that can identify items 610 to be used, content 620 that workers 60 refer to, etc.

[0155] When a maintenance recommendation task is generated, the information processing device 6 identifies the condition indicator at that time and extracts knowledge information 380 corresponding to the identified condition indicator from the database 38. The information processing device 6 then converts the extracted knowledge information 380 into a message and sends it to the information terminal 40. As a result, the information terminal 40 can display the knowledge information on its display unit, enabling pre-maintenance by detecting signs of malfunctions in the industrial equipment 2.

[0156] [7] Effects As described above, in this embodiment, the industrial equipment operation support system 1 comprises a plurality of environmental sensors 5 arranged on the industrial equipment 2, a plurality of controllers 61 equipped on the industrial equipment 2, and a plurality of information processing devices 6 configured to acquire and store environmental information detected by the environmental sensors 5 and operational information generated by the controllers 61, and having a hierarchical structure for transmitting environmental information and operational information from downstream to upstream. The plurality of information processing devices 6 each have, in one or more layers, an information storage platform 37 for storing acquired environmental information and operational information, a condition index adjustment engine 32 for extracting or generating condition indexes for the industrial equipment 2 in a time series from at least one of the environmental information and operational information stored on the information storage platform 37, an inference value storage engine 33 having a trained inference model 330 that infers the probability of a malfunction event occurring by inputting condition indexes and stores the obtained inferred value of the occurrence probability on the information storage platform 37, and a task generation engine 34 for generating maintenance recommendation tasks based on the results of data analysis using the time series data of the inferred value of the occurrence probability stored on the information storage platform 37.

[0157] According to the embodiment, the industrial equipment operation support system 1 can quickly detect signs of malfunctions in the industrial equipment 2 by analyzing data using time-series data of inferred probabilities of malfunctions occurring in the industrial equipment 2. As a result, the industrial equipment operation support system 1 can generate maintenance recommendation tasks and instruct equipment maintenance activities according to the situation. Consequently, maintenance operators can provide timely and efficient equipment maintenance activities for a wide variety of industrial equipment 2. Furthermore, the industrial equipment operation support system 1 can contribute to reducing manpower in equipment maintenance activities by eliminating daily or periodic inspections of the industrial equipment 2.

[0158] In the industrial equipment operation support system 1, multiple information processing devices 6 each have, at one or more levels, a database 38 that stores knowledge information 380 regarding the appropriate range of condition indicators that can suppress the probability of malfunction events to a predetermined probability or less, and maintenance activities that are effective in bringing condition indicators outside the appropriate range back within the appropriate range; a knowledge information extraction engine 35 that, when a maintenance recommendation task is generated, identifies the condition indicators at that time and extracts knowledge information 380 corresponding to the identified condition indicators from the database 38; and a knowledge information transmission engine 36 that converts the extracted knowledge information 380 into a message and sends it to an information terminal 40 that can be used by maintenance workers or equipment users. As a result, the industrial equipment operation support system 1 can instruct workers 60 on appropriate work content and shorten work man-hours by transmitting knowledge information 380 corresponding to condition indicators to the information terminal 40.

[0159] In the industrial equipment operation support system 1, the trained inference model 330 includes one or more of the following: a first inference model 331 that infers the probability of abnormal shutdown, where the industrial equipment 2 is stopped due to a malfunction; a second inference model 332 that infers the probability of efficiency degradation, where the industrial equipment 2 does not meet the required energy efficiency due to a malfunction; and a third inference model 333 that infers the probability of quality defects, where the industrial equipment 2 does not meet the required quality for industrial activities due to a malfunction. As a result, the industrial equipment operation support system 1 can infer any of the probabilities of abnormal shutdown, efficiency degradation, or quality defects in the industrial equipment 2, thereby generating maintenance recommendation tasks according to the type of malfunction that may occur in the industrial equipment 2 in the future, and avoiding a decline in equipment operation performance.

[0160] In the industrial equipment operation support system 1, the first inference model 331 consists of a logistic regression algorithm and is an inference model generated by machine learning using training data that uses the condition indicators of the industrial equipment 2 as features and the presence or absence of abnormal shutdowns as training labels. The operational information includes actual operating time information linked to the normal state of the industrial equipment 2 and non-operating time information linked to the abnormal state, and the probability of abnormal shutdowns is an influencing factor for the operational performance value based on the actual operating time information and non-operating time information. As a result, the industrial equipment operation support system 1 can generate maintenance recommendation tasks to prevent abnormal shutdowns of the industrial equipment 2.

[0161] In the industrial equipment operation support system 1, the second inference model 332 consists of a logistic regression algorithm and is an inference model generated by machine learning using training data that uses the condition indicators of the industrial equipment 2 as features and the presence or absence of efficiency deterioration as the training label. The environmental information includes input energy information and output energy information of the industrial equipment 2, and the probability of efficiency deterioration occurring is an influencing factor for the energy performance value based on the input energy information and output energy information. As a result, the industrial equipment operation support system 1 can generate maintenance recommendation tasks to prevent efficiency deterioration of the industrial equipment 2.

[0162] In the industrial equipment operation support system 1, the third inference model 333 consists of a logistic regression algorithm and is an inference model generated by machine learning using training data that uses the condition indicators of the industrial equipment 2 as features and the presence or absence of quality defects as training labels. The environmental information includes actual quality level information of the media or goods produced by the industrial equipment 2, and the probability of quality defects occurring is an influencing factor for the quality performance value based on the actual quality level information and the required quality level information. As a result, the industrial equipment operation support system 1 can generate maintenance recommendation tasks to prevent industrial equipment 2 from experiencing quality defects.

[0163] [8] Contribution to the United Nations-led Sustainable Development Goals (SDGs) The industrial equipment operation support system disclosed herein can collect a wide range of environmental and operational information from industrial and demand-related equipment, which can be used for operational management, energy management, quality control, and other purposes. As a result, it can improve the energy efficiency and productivity of business establishments, including factories, and contribute to achieving Sustainable Development Goal 7, "Affordable and Clean Energy." In addition, by reducing carbon dioxide emissions in conjunction with improved energy efficiency, it can contribute to achieving Goal 13, "Take urgent action to combat climate change and its impacts." [Explanation of Symbols]

[0164] 1…Industrial equipment operation support system, 2…Industrial equipment, 2A…First industrial equipment, 2B…Second industrial equipment, 3…Business office, 3A…First business office, 3B…Second business office, 3C…Third business office, 4…Factory, 4A…First factory, 4B…Second factory, 5…Environmental sensor, 6…Information processing device, 7…Data acquisition terminal, 8…Demand equipment, 10…Computer, 11…Processor, 12…Storage device, 13…Communication interface, 14…Input / output interface, 21…Supply route, 22…Transportation route, 23…Discharge route, 31…Information processing engine, 32…Condition index adjustment engine, 33…Inference value storage engine 34...Task generation engine, 35...Knowledge information extraction engine, 36...Knowledge information transmission engine, 37...Information storage platform, 38...Database, 40...Information terminal, 61...Controller, 61A...Microcomputer, 61B...Programmable logic controller, 62...Edge computer, 63...Gateway, 64...Guest computer, 65...Host computer, 330...Trained inference model, 331...First inference model, 332...Second inference model, 333...Third inference model, 380...Knowledge information, 610...Item, 620...Content

Claims

1. Multiple environmental sensors placed in industrial equipment, Multiple controllers equipped in the aforementioned industrial equipment, The system comprises a plurality of information processing devices configured to acquire and store environmental information detected by the environmental sensor and operational information generated by the controller, and having a hierarchical structure for transmitting the environmental information and operational information from downstream to upstream. Multiple information processing devices, in any one of the layers, An information storage platform for storing the acquired environmental information and operational information, A condition indicator preparation engine that extracts or generates condition indicators for industrial equipment in a time series from at least one of the environmental information and the operational information stored in the information storage platform, An inference value storage engine having a trained inference model, which, by inputting the condition index, infers the probability of a malfunction event occurring and stores the obtained inferred value of the occurrence probability in the information storage platform, A task generation engine that generates maintenance recommendation tasks based on the results of data analysis using time-series data of the inferred probability of occurrence stored in the information storage platform, Industrial equipment operation support system.

2. Multiple information processing devices, in any one of the layers, A database accumulating knowledge information regarding the appropriate range of the condition indicator that can suppress the probability of the occurrence of the aforementioned malfunction event to a predetermined probability or less, and maintenance activities that are effective in returning the condition indicator that has fallen outside the appropriate range back to the appropriate range, When the aforementioned maintenance recommendation task is generated, a knowledge information extraction engine identifies the condition indicator at that time and extracts the knowledge information corresponding to the identified condition indicator from the database. The system includes a knowledge information transmission engine that messages the extracted knowledge information and sends it to an information terminal usable by maintenance workers or equipment users. The industrial equipment operation support system according to claim 1.

3. The aforementioned trained inference model is A first inference model for inferring the probability of an abnormal shutdown occurring, which is a state in which the operation of the industrial equipment is stopped due to a malfunction. A second inference model for inferring the probability of efficiency degradation occurring, in which the industrial equipment fails to meet the required energy efficiency due to a malfunction, and It comprises one or more third inference models that infer the probability of a quality defect occurring, in which the industrial equipment fails to meet the required quality for industrial activities due to the occurrence of a malfunction. The industrial equipment operation support system according to claim 1 or claim 2.

4. The first inference model consists of a logistic regression algorithm and is an inference model generated by machine learning using training data in which the condition indicator of the industrial equipment is used as a feature and the presence or absence of abnormal stoppage is used as a training label. The aforementioned operational information includes actual operating time information linked to the normal state of the industrial equipment and non-operating time information linked to the abnormal state. The probability of abnormal shutdowns is a factor that influences the operational performance values ​​based on the actual operating time information and the non-operating time information. The industrial equipment operation support system according to claim 3.

5. The second inference model consists of a logistic regression algorithm and is an inference model generated by machine learning using training data in which the condition indicators of the industrial equipment are used as features and the presence or absence of efficiency deterioration is used as the training label. The aforementioned environmental information includes input energy information and output energy information for the industrial equipment. The probability of efficiency degradation occurring is an influencing factor for the energy performance value based on the input energy information and the output energy information. The industrial equipment operation support system according to claim 3.

6. The third inference model consists of a logistic regression algorithm and is an inference model generated by machine learning using training data in which the condition indicators of the industrial equipment are used as features and the presence or absence of quality defects is used as the training label. The aforementioned environmental information includes information on the actual quality level of the media or articles produced by the industrial equipment. The probability of quality defects occurring is a factor that influences quality performance values ​​based on actual quality level information and required quality level information. The industrial equipment operation support system according to claim 3.

Images