+

CN113205191A - Intelligent decision making system and method for equipment replacement based on reinforcement learning - Google Patents

Intelligent decision making system and method for equipment replacement based on reinforcement learning Download PDF

Info

Publication number
CN113205191A
CN113205191A CN202110408296.0A CN202110408296A CN113205191A CN 113205191 A CN113205191 A CN 113205191A CN 202110408296 A CN202110408296 A CN 202110408296A CN 113205191 A CN113205191 A CN 113205191A
Authority
CN
China
Prior art keywords
data
time
series data
time series
module
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110408296.0A
Other languages
Chinese (zh)
Inventor
李启娟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Terminus Technology Group Co Ltd
Original Assignee
Terminus Technology Group Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Terminus Technology Group Co Ltd filed Critical Terminus Technology Group Co Ltd
Priority to CN202110408296.0A priority Critical patent/CN113205191A/en
Publication of CN113205191A publication Critical patent/CN113205191A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2474—Sequence data queries, e.g. querying versioned data
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/00—2D [Two Dimensional] image generation
    • G06T11/20—Drawing from basic elements, e.g. lines or circles
    • G06T11/203—Drawing of straight lines or curves

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Business, Economics & Management (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Fuzzy Systems (AREA)
  • Probability & Statistics with Applications (AREA)
  • Databases & Information Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

An intelligent decision-making system and method for equipment replacement based on reinforcement learning. The system comprises: the system comprises a first data input module, a first data processing module and a second data processing module, wherein the first data input module is configured to acquire first time series data of each device of the whole life of a plurality of devices, the first data processing module is configured to fit a first curve according to the first time series data of each device, and the first time series data are generated by sampling the first curves at equal intervals; the training module is configured to train the neural grid of the decision module according to the multiple columns of first standard time series data; a second data input module configured to acquire second time-series data of a current healthy device through the sensor; a second data processing module configured to fit a second curve according to the second time series data and to sample the second curve at equal intervals to generate second time-scale time series data; and a decision module configured to calculate a replacement timing of the current healthy equipment according to the second time series data. The system and the method provided by the invention determine the replacement time through the neural network, thereby avoiding loss and unnecessary waste caused by sudden damage.

Description

Intelligent decision making system and method for equipment replacement based on reinforcement learning
Technical Field
The invention relates to an intelligent decision-making system and method for equipment replacement based on reinforcement learning, and belongs to the technical field of artificial intelligence.
Background
In real life, people often suffer from the fact that used equipment is damaged suddenly and is very annoying, and the economic loss caused by the equipment is immeasurable. The maintenance and replacement decision refers to estimating and predicting the state of the monitored equipment based on the state monitoring information and the analysis and explanation thereof, and recommending the current optimal maintenance and replacement strategy according to certain optimization targets, such as cost, safety, downtime, availability and the like. In the prior art, the remaining working time and the replacement time of the equipment in a healthy state are usually calculated according to the average service life and the used time of the equipment, and although the parts adopted during the manufacturing of each type of equipment are the same and the manufacturing process is the same, the working life is different due to different operations of operators and working environments. The method for calculating the working life of the equipment by adopting the average working time in the prior art can cause resource waste.
Disclosure of Invention
The invention provides an equipment replacement intelligent decision-making system and method based on reinforcement learning, which are used for determining the maintenance and replacement time of current healthy equipment according to the life cycle of the equipment similar to that of the equipment, so that economic waste is avoided.
In order to achieve the above object, the present invention provides an intelligent decision system for device replacement based on reinforcement learning, which is characterized by comprising: the system comprises a first data input module, a first data processing module and a second data processing module, wherein the first data input module is configured to acquire first time series data of each device of the whole life of a plurality of devices, the first data processing module is configured to fit a first curve according to the first time series data of each device, and the first time series data are generated by sampling the first curves at equal intervals; the training module is configured to train the neural grid of the decision module according to the multiple columns of first standard time series data; a second data input module configured to acquire second time-series data of a current healthy device through the sensor; a second data processing module configured to fit a second curve according to the second time series data and to sample the second curve at equal intervals to generate second time-scale time series data; and a decision module configured to calculate a replacement timing of the current healthy equipment according to the second time series data.
Preferably, the decision module comprises: the self-organizing competitive neural network comprises an input layer and a competitive layer, during a training stage, a plurality of rows of first time-scaling time sequence data are input into the input layer, the self-organizing competitive neural network learns a plurality of rows of first standard time sequence data, and the plurality of rows of first standard time sequence data and corresponding time form neurons of the competitive layer; in the decision stage, the second time series data are provided for the input layer, and the similarity of the second time series data, the competition layer and each neuron is calculated in sequence by the self-organizing competition neural network; and the judging unit judges the current residual working time and the current degradation working time of the healthy equipment according to the equipment full-life curve represented by the neuron sequence with the maximum similarity.
In order to achieve the purpose, the invention also provides an intelligent decision-making method for equipment replacement based on reinforcement learning, which comprises the following steps:
a first data acquisition step of acquiring first time series data of each of a plurality of devices over the life;
a first data processing step, configured to fit a first curve according to the first time series data of each device, and sample each first curve at equal intervals to generate a plurality of columns of first time-scale time series data;
a training step, training the decision module according to the multi-column first standard time series data;
a second data acquisition step of acquiring second time series data of the current healthy equipment through a sensor;
a second data processing step configured to fit a second curve to the second time-series data and to sample the second curve at equal intervals to generate second time-series data, the method further comprising:
and a decision step, namely calculating the current replacement time of the healthy equipment according to the second time sequence data by using a decision module.
Preferably, the decision module comprises: the self-organizing competitive neural network comprises an input layer and a competitive layer, during a training stage, a plurality of rows of first time-scale time sequence data are input into the input layer, the self-organizing competitive neural network learns a plurality of rows of first standard time sequence data, and the plurality of rows of first standard time sequence data and corresponding time form neurons of the competitive layer; in the decision stage, the second time series data are provided for the input layer, and the similarity of the second time series data, the competition layer and each neuron is calculated in sequence by the self-organizing competition neural network; and the judging unit judges the current residual working time and the current degradation working time of the healthy equipment according to the equipment full-life curve represented by the neuron sequence with the maximum similarity.
Compared with the prior art, the intelligent decision-making system and method for equipment replacement based on reinforcement learning provided by the invention can determine the maintenance replacement time of the current healthy equipment according to the equipment full-life time similar to that of the equipment, thereby avoiding economic waste.
Drawings
FIG. 1 is a block diagram of an intelligent decision system for device replacement based on reinforcement learning provided by the present invention;
FIG. 2 is a block diagram of the hardware components of the building management system provided by the present invention;
FIG. 3 is a block diagram of the hardware components of the campus health management system provided by the present invention;
FIG. 4 is a block diagram of the decision module provided by the present invention;
FIG. 5 is a schematic diagram of the components of the self-organizing competitive neural network model provided by the present invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In describing the present invention, it is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. The terms "comprises" and/or "comprising," when used in this specification, are intended to specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. The term "and/or" includes any and all combinations of one or more of the associated listed items. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
FIG. 1 is a block diagram of an intelligent decision system for device replacement based on reinforcement learning provided by the present invention; as shown in fig. 1, the intelligent decision making system at least comprises a campus health management system 200 and N building management systems 100, where N is an integer greater than or equal to 1. Each building management system 100 includes a plurality of subsystem health management layers 101 and 102. Each building is provided with a power supply subsystem, a water supply subsystem, a heating subsystem, a cooling subsystem, a monitoring subsystem, and a fire fighting subsystem, and a plurality of sensors are arranged for each subsystem to measure health values thereof, thereby forming a sensor layer 300, which includes a plurality of sensors, such as a sensor 301, a sensor 302, a sensor 303, and the like, arranged in a tree at nodes of each subsystem, and each sensor is provided with a unique identification ID, and each sensor can be connected to the building health management layer system 100 through a wireless channel or a wired channel, and the sensors of each subsystem can constitute a sensor network. The sensor of each subsystem is timed to send the measurement data to the sensor of the previous stage and receives the response information of the wireless sensor of the previous stage, the sensor of the previous stage collects the measurement data of the sensor managed by the sensor of the previous stage and then sends the measurement data to the sensor of the previous stage, and so on until all the data of the subsystem is sent to the corresponding subsystem management layer of the building health management system, the corresponding subsystem management layer 101 at least comprises a data processing module 1011 and a decision module 1012, the data processing module 1011 is configured to process the data provided by the sensor layer 300 and then provide the data to the decision module 1012, and the decision module 1012 at least comprises a neural network model trained by the model training module 202 of the campus health management system 200 through big data. The health management layer 102 of each building, for example, includes a database 1021, an integration module 1022, and an input/output interface 1023, where the database 1021 is used for storing various trained neural network models, and data processed by the data processing module 1011 on the data measured by the sensor layer 300. The integration module 1022 integrates the data and decision results of the various subsystem layers and sends them to the campus health management system 200, which has a hardware configuration as shown in fig. 2.
The campus health management system 200 configures the devices of each building through the configuration module 400, and includes a data processing module 201, a model training module 202, a performance evaluation module 203, an input/output interface, a database 204, and the like, where the data processing module 201 is configured to receive the measurement data of each device of each subsystem transmitted by each building system, process the obtained measurement data, store the data in the database 204, call and process the data of the database by the model training module 202, train various models, and transmit the trained modules to the health management system 100 of each building. The performance assessment module 203 invokes data in the database to perform a performance assessment of the devices of the subsystems of the managed building.
Fig. 2 is a block diagram of the hardware components of the building health management system according to the present invention, as shown in fig. 2, the hardware of the building management system includes a processor 10 and a memory 11, the memory 11 includes a database for storing data, applications and trained decision models 1012, the processor 10 calls the trained decision models 1012 and calls curves and initial conditions stored in the database to predict the remaining life of the equipment based on the equipment data measured by the sensors, thereby determining the replacement timing. The application program comprises a subsystem health management layer and a building health management layer, wherein the subsystem health management layer comprises a data processing module 1011 and a decision module 1012, the data processing module 1011 comprises a data input module and a data processing submodule, and the data input module is configured to acquire time series data of current health equipment through a sensor; a data processing sub-module configured to fit a curve according to the time series data, sample the curve at equal intervals to generate time-series data and provide the time-series data to the decision module 1012, and the decision module 1012 is configured to calculate a replacement timing of the healthy equipment according to the time-series data.
In the present invention, the hardware of the building health management system 100 also includes a communication unit 13 configured to communicate with the campus health management system. The hardware of the building health management system also includes a display 12 for displaying the measurements of the sensors, the application programs and interfaces to the system programs, etc., the application programs including at least a data input module configured to obtain time series data for the current health devices via the sensors; the system comprises a data processing module and a decision-making module, wherein the data processing module is configured to fit a curve according to the time series data and sample the curve at equal intervals to generate standard time series data, and the decision-making module is configured to calculate the current replacement opportunity of the healthy equipment according to the standard time series data. The hardware of the building health management system also includes an input/output interface 14, which is an interface for a user to input data, for accessing a keyboard, a mouse, an optical drive, a usb disk, etc.
Fig. 3 is a block diagram of the hardware components of the campus health management system provided by the present invention, and as shown in fig. 3, the system includes a processor 20 and a memory 21 connected by a bus, where the memory 21 includes a database for storing data sent from each building management system, and storing decision modules of a power supply subsystem, a water supply subsystem, a heating subsystem, a cooling subsystem, a monitoring subsystem, and a fire protection subsystem. The processor 20 invokes a program stored in memory to implement the functions of the campus health management system, which includes a data processing module 201 and a model training module 202. The data processing module 201 processes data sent by each building system and stores the data in the database 204, and includes a data input module configured to obtain multiple rows of time series data of the full lives of multiple devices, where each row of time series data corresponds to one device, that is, the similar devices include different full-life curve sequences. The data processing sub-module is configured to fit a plurality of first curves according to the plurality of columns of first time series data, and sample the plurality of first curves at equal intervals to generate time-scale time series data; the model training module 202 calls the processed data and trains various models (e.g., device replacement determination models), preferably configured to train a neural network in the decision module 1012 according to a multi-column standard time series data to learn different life-cycle curve series data for the same class of devices. The campus health management system also includes an input-output interface 24 for data output and also for inputting instructions and the like. The hardware of the campus health management system further includes a communication unit 23 for communicating with each building system and service department (e.g., fire department, power supply department, water supply department, etc.) and transmitting at least the trained equipment replacement determination model to the health management system of each building. The hardware of the park health management system also includes a display 22, and the execution processes and final results of the model training module, the performance evaluation module, etc. can be displayed on the display 22 for the operator to observe.
In the present invention, the one or more processors as hardware may be one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, logic circuits, and/or any devices that manipulate signals based on operational instructions. The processor is configured to retrieve and execute computer readable instructions stored in the memory. The software system may be implemented in various computing systems, such as a laptop, a notebook, a handheld device, a workstation, a mainframe, a server, a network cloud, and so forth. Input output (I/O) interfaces may include various software and hardware interfaces, for example, to which a printer, keyboard, usb disk, network, cable, mouse, etc. may be connected. The communication unit is configured to communicate with other devices over a wireless network, such as a WLAN, cellular, or satellite. The display is used for visual interaction with a user.
In the present invention, the campus health management system 200 and the building health management system 100 are both installed with an equipment replacement intelligent decision system application program based on reinforcement learning, and their constituent modules are the same and have the composition shown in fig. 4.
Fig. 4 is a block diagram of a decision module provided in the present invention, and as shown in fig. 4, the decision module 1012 includes: the self-organizing competitive neural network 331, the judging unit 332 and the output unit 333, wherein the self-organizing competitive neural network 331 includes an input layer and a competitive layer, neurons of the competitive layer form a two-dimensional structure, as shown in fig. 5, an X direction represents time, a Y direction represents a measurement value, an input node of the input layer and an output node of the competitive layer use a variable weight uijFully connecting, wherein i is 1,2, and m is the number of input nodes; j is 1,2, …, n, n is the number of output nodes. In the training stage, multiple columns of standard time sequence data are processed 201 by the data processing module and input into the input layer, and neurons in the competition layer learn the multiple columns of standard time sequence data to form different life-cycle curve sequences of the same type of equipment, such as f1 and f 2. The specific process of the training phase is as follows:
s01: an initialization step: initializing the neural network to initialize the weights, determining an initial learning value alpha0And total number of learning times K;
s02: distance calculation step: calculating an input standard time series vector X ═ X1,…,xi,…xm]Distance d between each input node in the system and each neuron in the output layerj：
dj＝|xi-fj|＝|xi-uij(0)xi|＝|xi(1-uij(0))|
In the formula uij(0) Is an initial weight;
s03: and (3) selecting neurons: will be input to node xiNeurons of the output layer with the smallest distance
Figure BDA0003021179040000081
Making initial neurons matched with the initial neurons;
s04: weight adjusting step: modulating neurons by
Figure BDA0003021179040000082
Node (neuron) weight coefficients contained in the life-cycle curve sequence (e.g. f1) of the device:
uij(k)＝uij(k-1)+αk(xi-uij(k-1))
Figure BDA0003021179040000083
in the formula uij(k) Weight, u, output for the current time kij(K-1) is the weight of the previous output, K is 1, 2.. K;
s05: judging whether the learning frequency K is reached or not, and if not, repeating the steps S02-S04; if the optimal weight coefficient u is reached, outputting the optimal weight coefficient uij。
In the decision-making stage, the device time sequence data obtained by processing of the data processing module 1011 are provided for an input layer, and the similarity between the time sequence data and each neuron of a competition layer is calculated in sequence by the self-organizing competition neural network. The judging unit judges the remaining working time and the degradation working time of the current healthy equipment according to the equipment full-life curve characterized by the neuron sequence with the maximum similarity, for example, the measured time sequence data of the current healthy equipment is most similar to the front section of the curve sequence f1, the full life of the equipment is judged to be the same as the full life of the equipment characterized by the curve sequence f1, and the remaining working time Ru and the performance degradation working time Ts of the equipment are judged according to the curve sequence f1, wherein the remaining working time Ru is the working time of the life:
Ru＝r-T-TS(hour)
In the formula, r is the total life time of the equipment, and T is the working time; and Ts is performance decay working time Ts which is the time from the degradation of the equipment performance to the complete failure of the equipment performance, and can also be the time from (0.5-0.7) when the equipment performance is degraded to the normal performance to the complete failure of the equipment performance.
According to an embodiment of the present invention, there is also provided an intelligent decision method for device replacement based on reinforcement learning, including the following steps: a first data acquisition step, namely acquiring multi-row first time sequence data of the whole life of multi-row equipment; a first data processing step configured to fit a plurality of first curves according to the plurality of columns of first time series data, and equally-spaced sampling the plurality of first curves to generate a plurality of columns of first time-scale time series data; a training step, training the decision module according to the first standard time sequence data; a second data acquisition step of acquiring second time series data of the current healthy equipment through a sensor; a second data processing step configured to fit a second curve according to the second time series data, sample the second curve at equal intervals to generate second time-scale time series data, and a decision step of calculating a maintenance opportunity currently being the healthy equipment according to the second time series data.
There is also provided, in accordance with an embodiment of the present invention, a storage medium for storing program code in a computer language for programming the above-described method to be executable by a processor, the program code for implementing one or more of the steps of the method. A readable medium can be any means that can store, communicate, or transport the computer program for use by the instruction execution system, apparatus, or device.
The invention determines the performance degradation time of each device through the decision module with the neural network, so that the device can exert the maximum efficiency and the resources are saved.
It should be understood that the above examples are only for clarity of illustration and are not intended to limit the embodiments. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. And obvious variations or modifications therefrom are within the scope of the invention.

Claims (4)

1. An intelligent decision-making system for equipment replacement based on reinforcement learning, comprising: the system comprises a first data input module, a first data processing module and a second data processing module, wherein the first data input module is configured to acquire first time series data of each device of the whole life of a plurality of devices, the first data processing module is configured to fit a first curve according to the first time series data of each device, and the first time series data are generated by sampling the first curves at equal intervals; the training module is configured to train the neural grid of the decision module according to the multiple columns of first standard time series data; a second data input module configured to acquire second time-series data of a current healthy device through the sensor; a second data processing module configured to fit a second curve according to the second time series data and to sample the second curve at equal intervals to generate second time-scale time series data; and a decision module configured to calculate a replacement timing of the current healthy equipment according to the second time series data.
2. The reinforcement learning-based equipment change intelligent decision system according to claim 1, wherein the decision module comprises: the self-organizing competitive neural network comprises an input layer and a competitive layer, during a training stage, a plurality of rows of first time-scaling time sequence data are input into the input layer, the self-organizing competitive neural network learns a plurality of rows of first standard time sequence data, and the plurality of rows of first standard time sequence data and corresponding time form neurons of the competitive layer; in the decision stage, the second time series data are provided for the input layer, and the similarity of the second time series data, the competition layer and each neuron is calculated in sequence by the self-organizing competition neural network; and the judging unit judges the current residual working time and the current degradation working time of the healthy equipment according to the equipment full-life curve represented by the neuron sequence with the maximum similarity.
3. An intelligent decision-making method for equipment replacement based on reinforcement learning comprises the following steps:
a first data acquisition step of acquiring first time series data of each of a plurality of devices over the life;
a first data processing step, configured to fit a first curve according to the first time series data of each device, and sample each first curve at equal intervals to generate a plurality of columns of first time-scale time series data;
a training step, training the decision module according to the multi-column first standard time series data;
a second data acquisition step of acquiring second time series data of the current healthy equipment through a sensor;
a second data processing step configured to fit a second curve to the second time-series data and to sample the second curve at equal intervals to generate second time-series data, the method further comprising:
and a decision step, namely calculating the current replacement time of the healthy equipment according to the second time sequence data by using a decision module.
4. The reinforcement learning-based equipment replacement intelligent decision method according to claim 3, wherein the decision module comprises: the self-organizing competitive neural network comprises an input layer and a competitive layer, during a training stage, a plurality of rows of first time-scale time sequence data are input into the input layer, the self-organizing competitive neural network learns a plurality of rows of first standard time sequence data, and the plurality of rows of first standard time sequence data and corresponding time form neurons of the competitive layer; in the decision stage, the second time series data are provided for the input layer, and the similarity of the second time series data, the competition layer and each neuron is calculated in sequence by the self-organizing competition neural network; and the judging unit judges the current residual working time and the current degradation working time of the healthy equipment according to the equipment full-life curve represented by the neuron sequence with the maximum similarity.
CN202110408296.0A 2021-04-15 2021-04-15 Intelligent decision making system and method for equipment replacement based on reinforcement learning Pending CN113205191A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110408296.0A CN113205191A (en) 2021-04-15 2021-04-15 Intelligent decision making system and method for equipment replacement based on reinforcement learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110408296.0A CN113205191A (en) 2021-04-15 2021-04-15 Intelligent decision making system and method for equipment replacement based on reinforcement learning

Publications (1)

Publication Number Publication Date
CN113205191A true CN113205191A (en) 2021-08-03

Family

ID=77027094

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110408296.0A Pending CN113205191A (en) 2021-04-15 2021-04-15 Intelligent decision making system and method for equipment replacement based on reinforcement learning

Country Status (1)

Country Link
CN (1) CN113205191A (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20020032711A (en) * 2000-10-26 2002-05-04 이구택 Selection method of sensor position in a rotary machine and Diagnosis method of the rotary machine vibration using self-organized map neural network
CN102313577A (en) * 2011-06-24 2012-01-11 上海大学 Equipment health state evaluation and recession prediction method based on multi-channel sensing signals
CN108154190A (en) * 2018-01-12 2018-06-12 上海海事大学 A kind of gantry crane motor status clustering method based on SOM neural networks
CN109753872A (en) * 2018-11-22 2019-05-14 四川大学 Reinforcement learning unit matching recurrent neural network system and its training and prediction method
CN111274737A (en) * 2020-02-25 2020-06-12 山东大学 Method and system for predicting remaining service life of mechanical equipment

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20020032711A (en) * 2000-10-26 2002-05-04 이구택 Selection method of sensor position in a rotary machine and Diagnosis method of the rotary machine vibration using self-organized map neural network
CN102313577A (en) * 2011-06-24 2012-01-11 上海大学 Equipment health state evaluation and recession prediction method based on multi-channel sensing signals
CN108154190A (en) * 2018-01-12 2018-06-12 上海海事大学 A kind of gantry crane motor status clustering method based on SOM neural networks
CN109753872A (en) * 2018-11-22 2019-05-14 四川大学 Reinforcement learning unit matching recurrent neural network system and its training and prediction method
CN111274737A (en) * 2020-02-25 2020-06-12 山东大学 Method and system for predicting remaining service life of mechanical equipment

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
康重庆等: "《电力系统负荷预测》", 31 July 2007 *

Similar Documents

Publication Publication Date Title
CN111124840B (en) Method and device for predicting alarm in business operation and maintenance and electronic equipment
CN118568471B (en) Intelligent power distribution station operation fault prediction method and system
CN112904220B (en) UPS (uninterrupted Power supply) health prediction method and system based on digital twinning and machine learning, electronic equipment and storable medium
US11092460B2 (en) Sensor control support apparatus, sensor control support method and non-transitory computer readable medium
KR20090078501A (en) Prediction method and apparatus for influent flow rate and influent component concentration of sewage treatment plant by artificial neural network
CN115034315B (en) Service processing method and device based on artificial intelligence, computer equipment and medium
CN118229036B (en) Operation and maintenance management system and method for smart city
CN112203311B (en) Network element abnormity diagnosis method, device, equipment and computer storage medium
CN113825165A (en) 5G slice network congestion early warning method and device based on time chart network
CN120011874A (en) Power system fault diagnosis and early warning system based on AI
CN118765399A (en) Automatically quantify the uncertainty of predictions provided by trained regression models
CN119537844A (en) A transformer fault diagnosis method, system, storage medium and electronic device based on sparrow algorithm
CN114091750A (en) Power grid load abnormity prediction method, system and storage medium
CN113205191A (en) Intelligent decision making system and method for equipment replacement based on reinforcement learning
CN116662904B (en) Data type variation detection method, device, computer equipment and medium
CN112785074B (en) Equipment maintenance and repair opportunity determination system and method based on deep learning
RU2559401C1 (en) Diagnostic of aircraft service systems operating conditions
CN116911160A (en) Dynamic increment-based residual life prediction method, device and medium
CN113377595B (en) Fault diagnosis method, device, electronic device and storage medium
CN112818598A (en) Simulation demonstration system and method for vehicle-mounted health management system
CN119206625B (en) New energy source motive equipment remote monitoring management system based on Internet of things technology
CN116347466A (en) Base station out of service alarm prediction method and device
CN112884210B (en) Vehicle health management system based on fuzzy clustering
CN110689062A (en) System and method for integrated health management of electromechanical equipment using multiple agents
US20250013519A1 (en) Medical imaging device fault handling

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20210803

点击 这是indexloc提供的php浏览器服务，不要输入任何密码和下载