CN115936294A - Equipment health management analysis method and terminal - Google Patents

Abstract

The invention discloses an equipment health management analysis method and a terminal, wherein observation values of various equipment in an operation system and equipment events which can affect equipment state change are collected, health indexes of each equipment are set, a structural analysis framework of an equipment health model is established based on the equipment observation values, the equipment events and the health indexes, equipment mechanisms are researched, the equipment mechanisms and data are fused, the health model is optimized, common rules and individual rules of the various equipment health models are discussed, and various equipment health generalized models are obtained; and researching the deviation degree of the specific equipment and the generalized model to acquire new phenomena and rules. By establishing a structural analysis framework of the equipment health model, the problems of how to carry out equipment health management and improve management work are solved, and a technology and a tool for equipment fault prediction and health management are provided; and (4) fusing mechanisms and data, establishing a plurality of mechanism models, and correcting and optimizing the data to obtain a reasonable health analysis model.

Description

Equipment health management analysis method and terminal

Technical Field

The invention relates to the technical field of equipment management, in particular to an equipment health management analysis method and a terminal.

Background

An operating system can not leave various equipment, and whether the working state of each equipment is good or not determines the result of the system to execute tasks, particularly in a ground service system of an aerospace launching site, hundreds of equipment facilities of dozens of different professional systems such as power supply, gas supply, air conditioning, fire fighting and the like are involved, and the equipment facilities must be ensured to be in a healthy working state to support every aerospace launching task.

In an operation system, various equipment is included and is used as an end user of the equipment instead of an equipment development unit, firstly, the understanding of internal mechanisms of different types of equipment is limited in a theoretical level, professional knowledge theories and test data of various specific equipment are lacked, the equipment is difficult to perform professional and systematic tests to obtain comprehensive test data in a practical level, and a general common law is difficult to master for guiding and analyzing whether each index data of a specific equipment is normal under an actual working condition; secondly, the equipment of the same type has few samples, and is lack of enough samples of specific equipment of a certain type, so that an effective reference group is difficult to establish, and common statistical type characteristic data is lacked; third, an extreme lack of equipment failure data.

Based on this, the equipment user is caused to be limited from general to specific deductive ability and from specific to general inductive ability. The space launching field can only operate and maintain the equipment according to the standard operation and maintenance guarantee scheme provided by the equipment manufacturer, and a specific equipment operation and maintenance and health management scheme is difficult to make according to the personalized working condition of the launching field equipment. Because the working condition of the launching task is different from the equipment working condition assumed by the standard operation and maintenance support scheme, the decision of the equipment operation and maintenance support only tends to be over maintenance strategy to ensure the success of the launching task in the actual operation and maintenance work, and the equipment support work under the strategy has considerable optimization and improvement space.

Therefore, how to build a set of equipment health management method system is a method which needs to be solved urgently at present.

Disclosure of Invention

The invention aims to provide an equipment health management analysis method and a terminal, which are based on equipment observation values, equipment events and health indexes, provide a structural analysis framework of health management, research equipment mechanisms, fuse the mechanisms and data, establish various health analysis models, correct and optimize the health analysis models by using real-time data, guide equipment operation and maintenance and health management by using the models, and realize the health management of the equipment.

In a first aspect, the above object of the present invention is achieved by the following technical solutions:

an equipment health management analysis method comprises the steps of collecting observation values of various equipment in an operation system and equipment events which can affect equipment state changes, setting health indexes of each equipment, establishing a structural analysis framework of an equipment health model based on the equipment observation values, the equipment events and the health indexes, researching equipment mechanisms, fusing the equipment mechanisms and data, optimizing a health model, discussing common rules and individual rules of the various equipment health models, and obtaining various equipment health generalized models; and researching the deviation degree of the specific equipment and the generalized model to acquire new phenomena and rules.

The invention is further configured to: a structural analysis framework of a health model is equipped, multiple observed values of similar equipment are taken as independent variables, clustering analysis is carried out, and a working condition discrimination model is established; taking the multiple observed values as independent variables and the health indexes as dependent variables, performing regression analysis, and establishing a health assessment model; taking multiple health indexes as independent variables, performing regression analysis, and establishing a life prediction model; taking the multiple observed values as independent variables and equipment events as dependent variables, carrying out causal analysis and establishing a fault diagnosis model; and (4) carrying out causal analysis by taking the equipment event as an independent variable and the health index as a dependent variable, and establishing a health management model.

The invention is further configured to: and establishing a model component based on the health model modeler for establishing a health model and providing support for subsequent upgrading and expanding.

The invention is further configured to: the model component comprises a health management front end, a health management rear end, a model component micro-service and a model modeler; the system comprises a model modeler and a related algorithm program, wherein the model modeler and the related algorithm program are used for providing model health state data, the model component micro-service is stored in a micro-service gateway and is used for providing the modeler and the related algorithm program, the health management front end is used for executing equipment operation and maintenance work on site, and the health management rear end is used for calling the model component micro-service when the health management front end executes the equipment operation and maintenance work.

The invention is further configured to: according to historical fault data, partitioning various equipment according to fault frequency and fault elimination cost, determining equipment object variables of the equipment, taking the equipment which belongs to a region with high fault cost and low fault frequency as a health management target object, carrying out mechanism research on the equipment, acquiring principle characteristics of the equipment, and establishing a model based on a data set; spare parts are reserved for equipment belonging to a region with low fault cost and frequent fault occurrence frequency; representing a good state for equipment belonging to a region with low fault cost and low fault frequency; and replacing the equipment belonging to the area with high fault cost and frequent fault occurrence frequency.

The invention is further configured to: performing mechanism research on the health management target object, searching data corresponding to the model according to the open data set, and verifying to obtain at least one candidate model; verifying the candidate model by using real test data of an equipment manufacturer to obtain an actual deployment model; according to the requirements of an actual deployment model, data acquisition and configuration of the Internet of things system are carried out, and data are collected to a data center in real time; preprocessing data of a data center to obtain subject data for at least one mechanism model; and (4) surrounding the theme data, combining the model calculation output and the actual acquired data, and optimizing the model parameters.

The invention is further configured to: and accumulating the real-time data along with time, continuously optimizing the model parameters, researching the change rule of the real-time data and the model parameters, and discovering a new rule to obtain new knowledge.

The invention is further configured to: obtaining principle characteristics based on public data, and performing mechanism research, including physical mathematical equation modeling, agent subject object modeling, simulation modeling and probability modeling; and (4) optimizing model parameters by combining a physical information neural network.

The invention is further configured to: the observation database comprises operation and maintenance data, operation data and basic archive data, and the equipment event data comprises alarm event data generated based on the real-time observation.

In a second aspect, the above object of the present invention is achieved by the following technical solutions:

an equipment health management analysis terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor implementing the method when executing the computer program.

Compared with the prior art, the beneficial technical effects of this application do:

1. the structural analysis framework of the health management solves the problems of how to carry out equipment health management and improve management work, and provides a technology and a tool for equipment fault prediction and health management;

2. furthermore, the equipment mechanism and the data are fused, multiple models are built according to the equipment mechanism, and a reasonable model is provided for health analysis by using the data to correct and optimize the mechanism model;

3. furthermore, the model provided by the application can define the interface format between the model and the front end, ensure that the data interface format is unchanged, and provide guarantee for upgrading the algorithm and replacing a new model component.

Drawings

FIG. 1 is a schematic structural diagram of a structural analysis framework of an embodiment of the present application;

FIG. 2 is a schematic diagram of a structural analysis framework according to an embodiment of the present application;

FIG. 3 is a schematic diagram of a health management engineering technique according to an embodiment of the present application;

FIG. 4 is a schematic structural diagram of a health model in accordance with an embodiment of the present application;

FIG. 5 is a schematic diagram of an algorithm flow of a health model according to an embodiment of the present application.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings.

Detailed description of the preferred embodiment

The equipment health management analysis method comprises the steps of establishing a structural analysis framework of an equipment health model, analyzing equipment mechanism, fusing the equipment mechanism with data, optimizing the health model in the framework, finding problems existing in equipment in the equipment operation process, knowing whether the equipment is normal and knowing which factors influence the normal work of the equipment, and predicting the normal work time of the equipment.

A structured analysis framework of equipment health models, as shown in FIG. 1, builds at least one health model based on observations, health indicators, and equipment events.

Observed values (MVs) of equipment include raw physical quantities directly collected by sensors or physical quantities calculated from raw physical quantities by a physical law equation.

Such as temperature, voltage, current, etc. are the original physical quantities or resistances calculated from the measured voltage and current using ohm's law.

The observation value is from the detection data of the equipment, including the data that thing networking management software gathered, the data that operation and maintenance management gathered include the data of manual work filling.

The health index (state of health, SOH) is an abstract percentage value used for representing the utility value of certain equipment, and the value range is 0-100%.

The health index is 100% when the equipment is capable of 100% of achieving its utility, less than 100% when the equipment is not sufficiently capable of achieving its utility, and 0% when the equipment is not capable of achieving its utility.

Equipment Events (EEs) describe external stimuli that may affect equipment state changes or the phenomenon that the equipment state changes itself can be observed by the outside world, the basic property of which is to be able to be manipulated or observed.

Such as handling or disassembling equipment, rusting or deformation of equipment, etc., are equipment events.

The health index is closely related to equipment and needs to be repeatedly confirmed in an actual engineering project and a project relation person. The equipment event is experience, belongs to shallow knowledge, and comprises the triggering of various set alarm criteria, such as the result after the temperature is higher than a set value, the actions of routing inspection, test run, production start-stop, maintenance and repair caused by maintenance regulations in days, weeks, months and seasons, and the like.

Establishing the following health models based on the observation values, the health indexes and the equipment events, wherein the health models comprise: taking the multiple observed values as independent variables, carrying out clustering analysis, and establishing a working condition discrimination model; taking the observed value as an independent variable and the health index as a dependent variable, performing regression analysis, and establishing a health evaluation model; taking multiple health indexes as independent variables, performing regression analysis, and establishing a life prediction model; carrying out causal analysis by taking the observed value as an independent variable and the equipment event as a dependent variable, and establishing a fault diagnosis model; and (4) carrying out causal analysis by taking the equipment event as an independent variable and the health index as a dependent variable, and establishing a health management model. A process simulation model is established based on the device events.

The types of the models are generally called health models, and the types of the health models are set according to requirements, or more than the types of the health models, or less than the types of the health models.

Based on the health model, the commonality and the individual rule of the health model are discussed, the model research of the cross-equipment model and the cross-equipment type is realized, the equipment health generalization model with wider adaptability is obtained, the universality theory of equipment health management can be refined, and the application range of the model is conveniently expanded; on the basis of the generalization model, the deviation degree of a specific device and the generalization model is researched, a new phenomenon and a new rule are discovered, and deeper cognition on the specific device is developed.

As shown in fig. 2, based on the health assessment models of a plurality of devices, a common law of the health assessment models is found from concrete to abstract across device models and device types, and a generalized model suitable for all devices is obtained.

And (3) based on the working condition discrimination model and the generalization model of each device, finding new phenomena from abstraction to concrete, finding personalized rules and establishing a deviation model.

In the operation process of the equipment, an operation instruction is executed, various monitoring indexes are fed back through a sensor, and the equipment is subjected to daily maintenance operation, so that operation and maintenance data, operation data and basic data of the equipment are collected to form an equipment observation value (MVS) database, the basic data comprises data for describing equipment functional structure decomposition information, and the operation data comes from various sensing measurements on the equipment.

Based on the real-time observed value data, the alarm events caused constitute important data sources of equipment events.

And according to different requirements, increasing or reducing the types of the models in the analysis framework.

As shown in fig. 3, theoretical research and test analysis are performed on equipment objects, a fault diagnosis model, a health assessment model and a health management model of different equipment are established,

and establishing a fault diagnosis model and a health evaluation model based on historical data in an equipment observation value (MVS) database and by combining real-time alarm data.

On the basis of a fault diagnosis model and a health assessment model, equipment Event (EEs) data are combined, a health index (SOH) data are combined on the basis of the health assessment model, a health management model is built, and equipment Health Events (HEs) having influences on equipment health are analyzed.

The establishment of the health model depends on program software, various health models are closely related to user requirements, and the modeling method and the input and output of the health models can continuously evolve along with the technical evolution, so that the expandability and the replaceability of the model need to be fully considered from the model operation environment and the external interface when a model component is established at the software architecture level, so as to provide support for the subsequent upgrading and expansion.

As shown in fig. 4, the software components of the present application include a health management front end, a health management back end, a model component microserver, a model modeler; the system comprises a model modeler and a related algorithm program, wherein the model modeler and the related algorithm program are used for providing model health state data, the model component micro-service is stored in a micro-service gateway and is used for providing the model modeler and the related algorithm program, the health management front end is used for executing equipment operation and maintenance work on site, and the health management rear end is used for calling the model component micro-service when the health management front end executes the equipment operation and maintenance work.

The model modeler is a WEB program based on a Single Page Application (SPA) technology, and based on the model modeler, background programmers write algorithm programs related to a model and publish the algorithm programs through a microservice gateway by using model component microservices. And (4) field equipment operation and maintenance personnel execute equipment operation and maintenance work through a health management application software WEB program at the health management front end, and the health management rear end calls a model component micro-service through a micro-service gateway to obtain health condition data of the model.

The health management software presentation form of the health management front end is determined, the data interface format between the health management front end page and the health model can be defined, and the internal upgrading algorithm or the internal upgrading algorithm can be replaced by a new model under the condition that the data interface format is not changed.

The health management front end is arranged on the virtual server, and the computers of the on-site operation and maintenance personnel are interacted with the virtual server to install the health management front end on the computers of the on-site operation and maintenance personnel.

The virtual server also stores a health management back end and an operating environment.

And establishing a health model based on equipment mechanism and data fusion. As shown in fig. 5.

The method comprises the steps of firstly analyzing historical data of equipment, and dividing various equipment data into regions according to the fault frequency and the fault elimination cost.

And the failure frequency is taken as a vertical axis, the failure elimination cost is taken as a horizontal axis, four quadrants are formed, and different equipment is divided into different regions and different quadrants.

For equipment belonging to an area with low fault cost and frequent fault occurrence frequency, a plurality of spare parts are reserved for replacement; the equipment belonging to the area with low fault cost and low fault frequency is represented to be in a good state; for equipment belonging to the area with high failure cost and frequent failure occurrence frequency, the equipment is indicated to be unavailable and should be replaced.

The equipment which belongs to the area with high fault cost and low fault frequency is used as a health management target object, the object variable of the equipment is determined, the mechanism research of the equipment is carried out, the principle characteristic of the equipment is obtained, and a model is established based on a data set.

For health management target objects, mechanism research is carried out, principle characteristics are obtained mainly through published documents, patents and the like, and a plurality of mechanism models are established, wherein the mechanism models comprise establishing models according to physical data equations such as material energy balance and the like, or establishing probability models, or establishing multi-Agent models (ABM), or Agent subject object modeling.

And realizing a mechanism model based on programming, searching data corresponding to the mechanism model through an open data set, verifying the mechanism model, and selecting a plurality of mechanism models with ideal verification results as candidate models according to the verification results. The open data set includes a Kaggle data set.

The method for acquiring the data in the open data set and the data details are known, and the real data of equipment manufacturers are acquired, wherein the real data comprise equipment data manuals such as datasheets, product schematic diagrams, product test reports and test data.

And (4) verifying the candidate model by using the real data, wherein the verification result is ideal and is used as an actual deployment model.

And according to the actual deployment model, data acquisition configuration of the Internet of things system is carried out, the data acquired in real time are imported into the data center, and the data range of the data center is expanded.

And carrying out data preprocessing on the data of the data center to obtain subject data for the verification service of at least one mechanism model.

And calculating an actual deployment model based on the theme data to obtain an output result, comparing actual acquired data with the output result, adjusting actual deployment model parameters, and performing model optimization.

In optimizing the model, a method including a Physical Information Neural Network (PINN) is used.

Further, as time is accumulated, data acquisition is continuously carried out, the model is optimized, model parameters are adjusted, the change rule of the real-time acquired data and the model parameters is researched, new knowledge is obtained by finding the new rule, and the equipment management analysis method is expanded.

Detailed description of the invention

An embodiment of the present invention provides a device health management analysis terminal device, where the terminal device includes: a processor, a memory, and a computer program, such as a mechanistic program, stored in the memory and executable on the processor, the processor implementing the physical programming of embodiment 1 when executing the computer program.

Illustratively, the computer program may be partitioned into one or more modules/units, stored in the memory and executed by the processor, to implement the invention. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used for describing the execution process of the computer program in the equipment health management analysis terminal device. For example, the computer program may be divided into a plurality of modules, each module having the following specific functions:

1. the modeling module is used for establishing various health models;

2. the acquisition module is used for acquiring real-time data of each device;

3. and the analysis module is used for analyzing the equipment state according to the model.

The equipment health management analysis terminal equipment can be computing equipment such as a desktop computer, a notebook computer, a palm computer and a cloud server. The equipment health management analysis terminal device can include, but is not limited to, a processor, and a memory. It will be understood by those skilled in the art that the above examples are merely examples of the equipment health management analysis terminal device, and do not constitute a limitation of the equipment health management analysis terminal device, and may include more or less components than those shown, or combine some components, or different components, for example, the equipment health management analysis terminal device may further include an input-output device, a network access device, a bus, etc.

The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. The general processor may be a microprocessor or the processor may be any conventional processor, etc., and the processor is a control center of the equipment health management analysis terminal device and connects various parts of the whole equipment health management analysis terminal device by using various interfaces and lines.

The memory can be used for storing the computer program and/or the module, and the processor can realize various functions of the equipment health management analysis terminal device by running or executing the computer program and/or the module stored in the memory and calling the data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data, a phonebook, etc.) created according to the use of the cellular phone, and the like. In addition, the memory may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.

The embodiments of the present invention are preferred embodiments of the present invention, and the scope of the present invention is not limited by these embodiments, so: equivalent changes made according to the structure, shape and principle of the invention shall be covered by the protection scope of the invention.

Claims (10)

1. The equipment health management analysis method is characterized by collecting observation values of various equipment in an operation system and equipment events which can affect equipment state change, setting health indexes of each equipment, establishing a structural analysis framework of an equipment health model based on the equipment observation values, the equipment events and the health indexes, researching equipment mechanisms, fusing the equipment mechanisms and data, optimizing the health model, discussing common rules and individual rules of the various equipment health models, and obtaining various equipment health generalization models; and researching the deviation degree of the specific equipment and the generalized model to acquire new phenomena and rules.

2. The equipment health management analysis method according to claim 1, wherein a structured analysis framework of the equipment health model takes multiple observed values of similar equipment as independent variables to perform cluster analysis and establish a working condition discrimination model; taking the multiple observed values as independent variables and the health indexes as dependent variables, performing regression analysis, and establishing a health evaluation model; taking multiple health indexes as independent variables, performing regression analysis, and establishing a life prediction model; taking the multiple observed values as independent variables and equipment events as dependent variables, carrying out causal analysis and establishing a fault diagnosis model; and (4) carrying out causal analysis by taking the equipment event as an independent variable and the health index as a dependent variable, and establishing a health management model.

3. The equipment health management analytics method of claim 2, wherein model components are created based on the health model modeler for building a health model to provide support for subsequent upgrade development.

4. The equipment health management analytics method of claim 3, wherein the model components comprise a health management front end, a health management back end, a model component microservice, a model modeler; the system comprises a model modeler and a related algorithm program, wherein the model modeler and the related algorithm program are used for providing model health state data, the model component micro-service is stored in a micro-service gateway and is used for providing the modeler and the related algorithm program, the health management front end is used for executing equipment operation and maintenance work on site, and the health management rear end is used for calling the model component micro-service when the health management front end executes the equipment operation and maintenance work.

5. The equipment health management analysis method according to claim 1, wherein according to historical fault data, with fault frequency and fault elimination cost, partitioning various types of equipment, taking equipment which has high fault cost but has low fault frequency as a health management target object, determining the object variables of the equipment, performing mechanism research on the equipment, acquiring the principle characteristics of the equipment, and establishing a model based on a data set; spare parts are reserved for equipment belonging to a region with low fault cost and frequent fault occurrence frequency; representing a good state for equipment belonging to a region with low fault cost and low fault frequency; and replacing the equipment belonging to the area with high fault cost and frequent fault occurrence frequency.

6. The equipment health management analysis method of claim 5, wherein the health management target object is subjected to mechanism research, and data corresponding to the model is searched for verification according to the open data set to obtain at least one candidate model; verifying the candidate model by using real test data of an equipment manufacturer to obtain an actual deployment model; according to the requirements of an actual deployment model, data acquisition and configuration of the Internet of things system are carried out, and data are collected to a data center in real time; preprocessing data of a data center to obtain subject data for at least one mechanism model; and (4) surrounding the theme data, combining the model calculation output and the actual collected data, and optimizing the model parameters.

7. The equipment health management analysis method of claim 6, wherein real-time data is accumulated over time, model parameters are continuously optimized, the change rules of the real-time data and the model parameters are studied, and new knowledge is obtained by discovering new rules.

8. The equipment health management analysis method of claim 6, wherein the principle features are obtained based on public data, and mechanism studies including physical mathematical equation modeling, agent subject object modeling, simulation modeling, and probability modeling are performed; and (4) optimizing model parameters by combining a physical information neural network.

9. The equipment health management analysis method of claim 1, wherein the observation database comprises operation and maintenance data, operational data, and base profile data, and the equipment event data comprises alarm event data generated based on real-time observations.

10. An equipment health management analysis terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: the processor, when executing the computer program, implements the method of any of claims 1-9.

Images