Equipment maintenance equipment demand prediction method and system based on digital twinning

Abstract

The invention discloses an equipment maintenance equipment demand prediction method and system based on digital twinning, and relates to the technical field of equipment maintenance equipment demand prediction, and the method comprises the steps: collecting the historical data and state data of target equipment, and obtaining the actual operation state of each piece of single equipment and each replaceable unit; synchronizing the actual operation state of each single device and each replaceable unit to a preset digital twin model as an initial state of simulation; determining task environment data; establishing an equipment state characteristic space-time relation model; inputting the equipment state characteristic space-time relation model and the task environment data into a preset digital twinborn model, calculating a sensitivity coefficient at the current moment, and constructing an equipment use failure rate model based on the sensitivity coefficient at the current moment; determining a simulation moment unit fault condition corresponding to the equipment use plan; and carrying out statistics on maintenance equipment requirements of the target equipment. According to the invention, the technical problem of large difference between the simulation result and the reality in the traditional equipment maintenance equipment demand prediction simulation is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment maintenance equipment demand prediction, and specifically to a method and system for equipment maintenance equipment demand prediction based on digital twins. Background Art

[0002] Maintenance equipment is a crucial material foundation for equipment maintenance support. Its level of support and replenishment strategies directly impact the combat effectiveness of troops and the efficiency of fund utilization. Insufficient maintenance equipment supply leads to a decrease in equipment availability, impacting combat effectiveness; excessive inventory backlogs increase inventory costs and affect fund efficiency. Therefore, accurately forecasting maintenance equipment demand is crucial for combat effectiveness and efficient fund utilization.

[0003] Currently, the two main methods for forecasting maintenance equipment demand are empirical and model-based. The empirical method primarily predicts maintenance equipment demand based on past equipment usage, while the model-based method primarily utilizes mathematical models. Existing methods have effectively addressed the problem of forecasting equipment maintenance equipment demand under different backgrounds and offer valuable insights for research on this issue. However, due to the highly complex nature of equipment, highly dispersed deployment, and varying mission intensities, traditional forecasting models are forced to make idealized assumptions to reduce computational complexity. These models fail to fully consider the impact of equipment usage and environmental factors on equipment failures, are incompatible with the complex nature of equipment use and support, and struggle to accurately forecast maintenance equipment demand and precisely ensure equipment support.

[0004] Although commonly used simulation methods comprehensively consider influencing factors, they often fail to consider the current technical status and maintenance records of the equipment, resulting in significant differences between simulation results and actual results. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for predicting equipment maintenance equipment demand based on digital twins in order to solve at least one of the above technical problems.

[0006] In the first aspect, an embodiment of the present invention provides a method for predicting equipment maintenance material demand based on digital twins, including: collecting historical data and status data of target equipment, and obtaining the actual operating status of each single equipment and each replaceable unit belonging to the target equipment; synchronizing the actual operating status of each single equipment and each replaceable unit to a preset digital twin model as the initial state of the simulation; determining task environment data based on the schedule of the task phase to be executed and the equipment usage plan; establishing an equipment status characteristic spatiotemporal relationship model based on the historical data of the target equipment and the inherent failure rate model of the replaceable unit belonging to the target equipment; inputting the equipment status characteristic spatiotemporal relationship model and the task environment data into the preset digital twin model, calculating the sensitivity coefficient at the current moment, and constructing an equipment usage failure rate model based on the sensitivity coefficient at the current moment; determining the unit failure situation at the simulation moment corresponding to the equipment usage plan based on the equipment usage failure rate model; and counting the maintenance material demand of the target equipment based on the unit failure situation at the simulation moment.

[0007] Furthermore, the historical data of the target equipment includes: equipment usage record data, equipment failure record data and equipment maintenance record data; the status data of the target equipment includes: equipment composition structure data and equipment status data.

[0008] Furthermore, the mission environment data includes: mission terrain environment data and mission meteorological environment data.

[0009] Furthermore, the inherent failure rate module includes: a constant failure rate model, a normal distribution model, and a Weibull distribution model.

[0010] Furthermore, the usage failure rate model includes: a usage failure rate model of a failure unit without state monitoring data and a usage failure rate model of a failure unit with state monitoring data.

[0011] Furthermore, the stateless monitoring data failure unit uses a failure rate model, including:

[0012] λ a =kλ

[0013]

[0014] Where λ a is the equipment failure rate, k is the sensitivity coefficient, k 10 is the task intensity coefficient, k 20 is the operating environment coefficient, k 30 is the personnel level coefficient, k 40 is the protection factor, and λ is the inherent failure rate of the equipment.

[0015] Furthermore, the state monitoring data fault unit uses a failure rate model, including: a data feature extraction model, a fault diagnosis model and a fault prediction model; wherein, the data feature extraction model is used to obtain dynamic data of the equipment to be diagnosed during operation, and preprocess, extract features and reduce the dimension of the dynamic data to obtain the frequency domain characteristics of the dynamic data; the fault diagnosis model includes a support vector machine model trained and constructed based on the dynamic data and the frequency domain characteristics of the dynamic data, and is used to diagnose faults of the equipment to be diagnosed; the fault prediction model includes a radial basis neural network model trained and constructed based on the dynamic data and the frequency domain characteristics of the dynamic data, and is used to predict faults of the equipment to be diagnosed.

[0016] In a second aspect, an embodiment of the present invention further provides an equipment maintenance material demand forecasting system based on digital twins, comprising: an acquisition module, a synchronization module, a first determination module, a first establishment module, a second establishment module, a second determination module and a statistical module; wherein the acquisition module is used to collect historical data and status data of the target equipment, and obtain the actual operating status of each single equipment and each replaceable unit belonging to the target equipment; the synchronization module is used to synchronize the actual operating status of each single equipment and each replaceable unit to a preset digital twin model as the initial state of the simulation; the first determination module is used to determine the task environment data based on the schedule of the task stage to be executed and the equipment usage plan; The first establishment module is used to establish a spatiotemporal relationship model of equipment status characteristics based on the historical data of the target equipment and the inherent failure rate model of the replaceable unit of the target equipment; the second establishment module is used to input the spatiotemporal relationship model of equipment status characteristics and the task environment data into the preset digital twin model, calculate the sensitivity coefficient at the current moment, and construct an equipment usage failure rate model based on the sensitivity coefficient at the current moment; the second determination module is used to determine the unit failure situation at the simulation moment corresponding to the equipment usage plan based on the equipment usage failure rate model; the statistical module is used to count the maintenance equipment requirements of the target equipment based on the unit failure situation at the simulation moment.

[0017] In a third aspect, an embodiment of the present invention further provides an electronic device comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the method provided in the embodiment of the present invention when executing the computer program.

[0018] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method provided in the embodiment of the present invention is implemented.

[0019] The present invention provides a method and system for predicting equipment maintenance material demand based on digital twins. Based on the digital twin model, the system takes the current state of the equipment as the initial state, fully considers the equipment mission and environmental conditions, and uses the failure rate of use at each moment to simulate the equipment operation and maintenance process, accurately predicting the demand for maintenance material, thereby preventing the shortage of some maintenance material and the squeeze of some maintenance material caused by the original guarantee mode, and alleviating the technical problem of large difference between simulation results and actual results in traditional equipment maintenance material demand prediction simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the implementation methods of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0021] Figure 1 A flowchart of a method for predicting equipment maintenance equipment demand based on digital twins provided in an embodiment of the present invention;

[0022] Figure 2 A schematic diagram of a framework of a preset digital twin model provided in an embodiment of the present invention;

[0023] Figure 3 A schematic diagram of a simulation process of a preset digital twin model provided in an embodiment of the invention;

[0024] Figure 4 A hierarchical schematic diagram of the equipment structure provided by an embodiment of the present invention;

[0025] Figure 5 A schematic diagram of a sensitivity coefficient system provided in an embodiment of the present invention;

[0026] Figure 6 A two-fold line sensitivity coefficient function curve provided by an embodiment of the present invention;

[0027] Figure 7 A schematic diagram of vibration data collection for an anti-aircraft gun automatic machine provided by an embodiment of the present invention;

[0028] Figure 8 A schematic diagram of a radial basis function neural network structure provided by an embodiment of the present invention;

[0029] Figure 9 A demand forecast simulation flow chart provided by an embodiment of the present invention;

[0030] Figure 10A schematic diagram of an equipment maintenance equipment demand forecasting system based on digital twins provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0031] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0032] Example 1

[0033] Figure 1 This is a flowchart of a method for predicting equipment maintenance equipment demand based on digital twins according to an embodiment of the present invention. Figure 1 As shown, the method specifically includes the following steps:

[0034] Step S102: collecting historical data and status data of the target equipment, and obtaining the actual operating status of each individual equipment and each replaceable unit belonging to the target equipment.

[0035] Step S104: Synchronize the actual operating status of each single device and each replaceable unit to the preset digital twin model as the initial state of the simulation.

[0036] Step S106: Determine mission environment data based on the mission phase schedule and equipment usage plan.

[0037] Specifically, mission environment data includes: mission terrain environment data and mission meteorological environment data.

[0038] Step S108 : establishing a spatiotemporal relationship model of equipment status characteristics based on historical data of the target equipment and an inherent failure rate model of a replaceable unit of the target equipment.

[0039] In step S110 , the equipment status feature spatiotemporal relationship model and the task environment data are input into the preset digital twin model, the sensitivity coefficient at the current moment is calculated, and an equipment usage failure rate model is constructed based on the sensitivity coefficient at the current moment.

[0040] Step S112: determining the unit failure condition at the simulation time corresponding to the equipment usage plan based on the equipment usage failure rate model.

[0041] Step S114: Based on the unit failure conditions at the simulation time, statistics are collected on the maintenance equipment requirements of the target equipment.

[0042] An embodiment of the present invention provides a method for predicting equipment maintenance material demand based on digital twins, establishes a digital twin model of the equipment support system, and synchronizes the operation data of the physical equipment to the digital twin model of the equipment; models the key parts of the equipment, collects the status data of the key parts of the physical equipment under different environments and different usage conditions, extracts feature data, and establishes a correspondence model between the feature data and the status of the key parts of the equipment; predicts the status of the key parts of the equipment according to the equipment usage task plan, and predicts the time when a failure may occur, thereby realizing the prediction of maintenance material demand and guiding the formulation of the physical equipment support strategy; based on the digital twin model, takes the current state of the equipment as the initial state, fully considers the equipment mission and environmental conditions, and uses the usage failure rate at each moment to simulate the equipment operation and maintenance process, accurately predicts the maintenance material demand, thereby preventing the problems of shortage of some maintenance materials and backlog of some maintenance materials caused by the original support model, and achieving the goal of precise equipment support.

[0043] Figure 2 : is a schematic diagram of a framework of a preset digital twin model provided according to an embodiment of the present invention. In an optional implementation provided by an embodiment of the present invention, Figure 2 As shown in the figure, a digital twin model of a typical equipment support system is established based on the theory of the "Five-Dimensional Digital Twin Model" (physical dimension, twin dimension, data dimension, application dimension, and connection dimension). During model operation, the physical dimension synchronizes real-time system operation data to the twin dimension, data dimension, and application dimension through data acquisition. The twin dimension uses entity data to simulate physical entities through simulation operations, providing model support to the application dimension. The data dimension is responsible for receiving, storing, and preprocessing various types of data and distributing them to various dimensions. The application dimension uses data provided by the data dimension and models provided by the twin dimension to provide equipment demand forecasting application services, assist in the formulation of support decisions, and act on the physical dimension. The connection dimension provides standards and methods for data transmission between dimensions within the model and between the model and external models, ensuring real-time and accurate data transmission.

[0044] Specifically, such as Figure 2 As shown, the five dimensions are interconnected and interoperate with each other, forming a true mapping of the virtual model to the physical system. This can accurately grasp the system status, predict system events and assist in decision-making, and achieve precise equipment protection.

[0045] (1) Physical dimension: complete entity data collection.

[0046] The physical dimension operates under the influence of various factors in the physical world and utilizes various data collection technologies to collect data reflecting the various states of physical entities and transmit it to the data dimension. This primarily relies on sensor and detection technologies to obtain information on the status of equipment entities. Furthermore, it relies on the equipment management and support information system to obtain information on the type, time of occurrence, and replacement of critical components of various equipment use and support events.

[0047] (2) Twin dimension: perform virtual system simulation.

[0048] The twin dimension is a digital twin that enables real-time, accurate description and virtual simulation of physical entities and is the source of twin data for digital twin models. A digital twin must be established for each piece of equipment under study, dynamically tracking the physical entity (typical equipment, replaceable units) under data-driven control and performing system simulation as needed.

[0049] (3) Data dimension: processing various types of model data.

[0050] The data dimension is a collection of various data processing of the digital twin model, including the reception, storage and preprocessing of various types of data, including the entity data collected by the physical dimension, the twin data collected by the twin dimension virtual simulation, the analytical data generated by various analytical processing in the application dimension, and the preprocessed data obtained after the data dimension itself preprocesses various data.

[0051] Entity data mainly includes task data that reflects the type and time of tasks that the system needs to perform; status monitoring data obtained through sensors, regular inspections and support information systems; environmental data such as temperature, humidity, wind and sand, and air pressure for system operation; support data obtained through support information systems, namely support event information; equipment parameters that reflect the composition, operating relationships, technical parameters, and restrictive conditions of each system; operating data that records the operating time, operating intensity, and key operating parameters of each element of the system; historical data for recording various system events, etc.

[0052] Twin data primarily includes simulation parameters that support the operation of the simulation system; the patterns of various events occurring within each device and unit of the system; the internal logic of how these events operate within the simulation system; and the simulation results of various methods. This data is calibrated in real time during system operation and simulation to achieve an accurate mapping of the physical entity.

[0053] Analytical data refers to the simulation results generated by simulating system operation, the analysis results generated by analyzing system data, and the evaluation results generated by evaluating system status, etc., which are generated by applying various data and model algorithms of the digital twin model.

[0054] (4) Application dimension: realizing various functions of the model.

[0055] The application dimension is a collection of algorithms and programs used in the digital twin model application process. It extracts features, performs simulation predictions, and optimizes decisions on the data dimension, implementing various decision-making support functions and applying the resulting decisions to the physical dimension. It is the interactive window of the digital twin model, providing the means for the digital twin model to output operational results and realize model functions. The model's various functions are presented to users through the application dimension and act on the physical dimension.

[0056] (5) Connection dimension: transmits information inside and outside the model.

[0057] The connection dimension is a collection of hardware, software, and protocols used for communication between the physical dimension, twin dimension, data dimension, and application dimension within the digital twin model, as well as between the model and external models. It establishes full connectivity between all dimensions of the model. Through the connection dimension, the dimensions are closely connected and mapped to each other. Models can also share data and provide a basis for each other.

[0058] Figure 3 Schematic diagram of a simulation process of a preset digital twin model provided according to an embodiment of the present invention. Figure 3 As shown, the basic idea of simulation prediction of a preset digital twin model provided by an embodiment of the present invention is as follows:

[0059] (1) Equipment status data collection. In the physical dimension, real-time equipment usage records, equipment failure records, equipment maintenance records, and equipment status data are collected to understand the actual operating status of each individual equipment and each replaceable unit in real time.

[0060] (2) Synchronization of virtual and real data. In the connection dimension, the actual operating status of each single equipment and each replaceable unit is synchronized to the twin model as the initial state of the simulation.

[0061] (2) Analyze equipment data. In the application dimension, obtain the inherent failure rate model of the equipment's replaceable units and establish a spatiotemporal relationship model of the equipment's status characteristics. By analyzing historical data such as equipment status, equipment usage environment, and equipment failure characteristics, a spatiotemporal relationship model of the equipment's status characteristics is established.

[0062] (3) Set up equipment usage plan. In the data dimension, establish the tasks that may be executed in the next cycle or the next stage, the operating conditions, and obtain the task stage schedule, so as to solve the equipment characteristic data changes during the simulation process.

[0063] (6) Calculate the unit failure rate. In the application dimension, the inherent failure rate and the equipment status characteristic spatiotemporal relationship model can be used to obtain the failure rate of each unit.

[0064] (7) Equipment operation simulation. In the twin dimension, the use failure rate is used during the simulation process to determine the unit failure situation at the simulation time corresponding to the equipment use plan.

[0065] (8) Maintenance equipment demand estimation: According to the equipment unit failure conditions generated during the simulation process, the maintenance equipment demand is calculated.

[0066] In an optional implementation provided by an embodiment of the present invention, the historical data of the target equipment includes: equipment usage record data, equipment failure record data and equipment maintenance record data; the status data of the target equipment includes: equipment composition structure data and equipment status data.

[0067] Optionally, Figure 4 FIG. 1 is a hierarchical diagram of an equipment structure provided according to an embodiment of the present invention. Figure 4 As shown, for the equipment participating in the formation combat mission, a four-level equipment structure is established, broken down into platform, subsystem, assembly, and component. Next, the equipment's maintenance intervals for medium and major overhauls are collected. The inherent reliability model of each replaceable unit is also collected. If it follows a Weibull distribution, its three parameters, m, η, and γ, are collected. If it follows a service index distribution, its parameter, λ, is collected, along with whether the component is required for medium or major overhauls.

[0068] In this embodiment of the present invention, the failure rate represents the probability of equipment failure occurring per unit time after time t, assuming the equipment has not failed at time t. The failure rate reflects the probability of equipment failure and is a fundamental parameter for equipment failure simulation and calculation of repair equipment requirements. It can be categorized as either inherent failure rate or operational failure rate.

[0069] The inherent failure rate refers to the failure rate under ideal operating and maintenance conditions. This rate is assigned during the equipment's design and manufacturing process. Once the specific equipment design and process are finalized, the inherent failure rate is fixed.

[0070] The operational failure rate refers to the failure rate of equipment during actual use. It is affected by a combination of factors including equipment design, installation, quality, environment, usage, and maintenance.

[0071] When the environmental factors, quality level, working status, etc. of the replaceable unit meet the specified standard conditions, the failure rate of the replaceable unit is the inherent failure rate of the replaceable unit, which is only related to the failure mechanism and failure mode of the replaceable unit.

[0072] A common method for determining equipment's inherent failure rate is to analyze equipment failure mechanisms and fit equipment test and operational data. Optionally, the inherent failure rate model includes a constant failure rate model, a normal distribution model, and a Weibull distribution model. Replaceable units (RUs) have different inherent failure rate models depending on their structural properties.

[0073] (1) Constant failure rate model:

[0074] The constant failure rate model is often applicable to electronic devices such as transistors, resistors, capacitors, integrated circuits, etc., which have a constant failure rate during use. The failure rate function is:

[0075]

[0076] Where: λ is the failure rate of the replaceable unit; θ is the average life of the unit.

[0077] (2) Normal distribution model:

[0078] The normal distribution model is often applicable to mechanical parts with wear as the main failure mode, such as gearboxes and reducers, whose life distribution follows the normal distribution. The failure rate function is:

[0079]

[0080] Where: f(t) is the failure density function of the unit; F(t) is the cumulative failure distribution function of the unit; μ is the mean of the unit life; σ is the standard deviation of the unit life; t is the cumulative working time of the unit.

[0081] (3) Weibull distribution model:

[0082] The Weibull distribution model is often applicable to electromechanical components, such as rolling bearings, relays, switches, circuit breakers, etc. Their life distribution follows the Weibull distribution. Its failure rate function is:

[0083]

[0084] Where β is the shape parameter of the Weibull distribution, η is the scale parameter of the Weibull distribution, γ is the location parameter of the Weibull distribution, and t is the cumulative working time of the unit.

[0085] The inherent failure rate model of the replaceable unit (RU) is determined through failure mechanism analysis during the design phase and test data during the testing phase. By determining the type of the RU's inherent failure rate model through failure mechanism analysis and fitting the parameters of the inherent failure rate function through equipment test data, the RU's inherent failure rate function can be obtained.

[0086] For equipment replaceable units that have not experienced any failures since they were put into service at the factory, an inherent failure rate model is used for simulation calculations.

[0087] In an optional implementation provided by an embodiment of the present invention, using a failure rate model includes: using a failure rate model for a failure unit without state monitoring data and using a failure rate model for a failure unit with state monitoring data.

[0088] (1) Failure rate model for fault units without state monitoring data:

[0089] For faulty units without stateful monitoring data, different replaceable units follow different failure rate patterns and exhibit different operational failure rates under the influence of missions and environments. By calculating the operational failure rates of replaceable units, simulation analysis of equipment failures can be performed.

[0090] To determine the equipment's operational failure rate, it's necessary to understand the degree to which the failure rate is affected by the mission and environment—that is, the failure rate's sensitivity to the mission and environment. By comprehensively analyzing the factors influencing the unit's operational failure rate and analyzing the sensitivity of each replaceable unit to each influencing factor, a sensitivity coefficient system can be established. This allows us to derive each unit's sensitivity to various environmental factors, thereby determining the unit's operational failure rate and enabling equipment operation simulation.

[0091] Optionally, factors influencing the failure rate include:

[0092] The inherent failure rate (IFR) is the failure rate of a replaceable unit (RU) under ideal operating and maintenance conditions. However, the actual operating environment of equipment often falls short of ideal requirements, and its operational failure rate varies depending on the equipment's mission and environment. In actual operation, the operational failure rate of a particular RU is not only related to its inherent failure rate but is also often influenced by multiple factors, including mission intensity, operating environment, operator skill, and maintenance compliance. The greater the mission intensity, the harsher the operating environment, the greater the risk of human error, and the less effective maintenance, the higher the unit's operational failure rate.

[0093] Mission intensity: Different units and training phases have different training missions, resulting in varying equipment operational intensities and conditions, leading to varying operational failure rates. Excessively high mission intensity, such as prolonged continuous use, can significantly impact operational failure rates. Factors influencing mission intensity primarily include mission type, continuous use duration, and cumulative usage time.

[0094] Operating Environment: Equipment's operating environment often differs from the standard environment. Changes in temperature, humidity, wind and sand, average slope, rainfall, snowfall, and other environmental conditions can affect equipment reliability and, consequently, the failure rate. The operating environment has a more sensitive and significant impact on the failure rate than other factors.

[0095] Personnel quality: During equipment operation, the structure and ability level of the operating personnel also have a significant impact on the failure rate. For example, a high number of unskilled personnel, a high number of unskilled operations, or a high number of operations due to a small number of personnel can easily lead to illegal operations, improper inspections, and insufficient maintenance, which will increase the equipment failure rate. Factors affecting personnel structure include educational background, years of military service, years of professional experience, and professional level.

[0096] Maintenance Implementation: Strictly implementing all equipment management regulations is an effective way to reduce equipment failures. The more effective the regulations and the stricter their implementation, the lower the failure rate. Factors influencing management regulations include maintenance intervals, maintenance support levels, and the degree of implementation of management regulations.

[0097] Specifically, the failure rate sensitivity coefficient system includes:

[0098] The sensitivity coefficient (k) of the failure rate of the replaceable unit to the external task and environment is the multiple effect of the external task and environment on the failure rate. a ) and the inherent failure rate (λ) are related as follows:

[0099] λ a =kλ

[0100] Assuming the sensitivity coefficient under standard environment is 1, the worse the environment, the higher the failure rate and the larger the coefficient. Conversely, the more conducive the environment is to equipment operation, the lower the failure rate and the smaller the coefficient.

[0101] Based on the factors affecting the failure rate in use, a sensitivity coefficient system of the failure rate of replaceable units in use to external tasks and environments can be established. Figure 5 : is a schematic diagram of a sensitivity coefficient system provided according to an embodiment of the present invention. Figure 5 As shown, corresponding to the influencing factors, the sensitivity coefficient system of the replaceable unit is composed of the task intensity coefficient k 10 , operating environment coefficient k 20 , user level coefficient k 30 , maintenance guarantee coefficient k 40 composition.

[0102] Specifically, the task intensity coefficient k 10 By the task type coefficient k 11 , continuous use time coefficient k 12 , cumulative usage time coefficient, etc.

[0103] Operating environment factor k 20 By the temperature coefficient k 21 , humidity coefficient k 22 , wind force coefficient k 23, blowing sand coefficient k 24 , average slope coefficient k 25 , rainfall coefficient k 26 , snowfall coefficient k 27 etc.

[0104] Personnel structure coefficient k 30 By the educational coefficient k 31 , enlistment years coefficient k 32 , professional years coefficient k 33 , professional level coefficient k 34 etc.

[0105] Maintenance implementation coefficient k 40 The maintenance timeliness coefficient k 41 , Maintenance completion coefficient k 42 etc.

[0106] In summary, the sensitivity coefficient k can be obtained by the following formula:

[0107]

[0108] Specifically, the method for determining the sensitivity coefficient includes:

[0109] The sensitivity coefficient of a replaceable unit is a function of the influencing factor value as the independent variable. It can be expressed in analytical form or tabular form. Analytical forms can be divided into linear, broken line, and curved forms. When the influencing factor is given in quantitative form, analytical form is generally used, while when the influencing factor is given in semi-quantitative or qualitative form, tabular form is generally used.

[0110] For example, the two-fold linear expression is:

[0111]

[0112] In the formula, k is the coefficient to be calculated, x is the quantitative value of the influencing factor, and a is the critical point. The function curve is as follows Figure 6 shown.

[0113] For example, the expression of sensitivity coefficient in tabular form is shown in Table 1:

[0114] Table 1 Expression of sensitivity coefficient in tabular form

[0115]

[0116] The sensitivity of the failure rate of each replaceable unit to the task and environment is analyzed. Through equipment reliability analysis, test data, historical data, similar comparison, expert scoring and other methods, the mapping relationship between the sensitivity coefficient of each replaceable unit and the value of the influencing factor can be determined.

[0117] During the operation of the equipment, the data actually displayed by the equipment is input into the preset digital twin model. The twin data is processed through the model application dimension, and the coefficients are gradually corrected using intelligent algorithms such as neural networks to obtain an accurate representation of the sensitive relationship.

[0118] In actual operation, the factors affecting the equipment are not static. They will change with different tasks, different mission locations, personnel changes, etc. At any time t, based on the impact of the replaceable unit on the task and environment at that moment, the sensitivity coefficient k(t) of the replaceable unit at that moment can be obtained. Multiplying it by the inherent failure rate λ(t) at that moment can give the service failure rate λ at that moment. a (t):

[0119] λ a (t) = k(t)λ(t)

[0120] The failure rate of replaceable units can be used to simulate failure conditions during equipment operation.

[0121] (2) The state monitoring data fault unit uses a failure rate model, including: data feature extraction model, fault diagnosis model and fault prediction model.

[0122] Specifically, the data feature extraction model is used to obtain dynamic data of the equipment to be diagnosed during operation, and to perform preprocessing, feature extraction, and feature dimension reduction on the dynamic data to obtain frequency domain features of the dynamic data;

[0123] Fault diagnosis models, including support vector machine models trained and constructed based on dynamic data and frequency domain features of dynamic data, used to diagnose faults in equipment to be diagnosed;

[0124] The fault prediction model includes a radial basis function neural network model trained and constructed based on dynamic data and frequency domain characteristics of the dynamic data, and is used to predict faults of equipment to be diagnosed.

[0125] For faulty units with status monitoring data, the use failure rate can be calculated more accurately based on the status monitoring data. Data features are extracted from the status data, and the dimension of the features is reduced. The equipment status corresponding to the features is determined according to different fault diagnosis algorithms.

[0126] Taking the vibration data of XX antiaircraft gun automatic machine as an example, four vibration sensors are used to collect the automatic machine status data during operation. The sensors are installed as follows: Figure 7 shown.

[0127] (1) Data feature extraction model:

[0128] The collected dynamic data from the automated machine's operation is preprocessed and feature extracted. The preprocessing stage uses wavelet or ICEEMDAN algorithms to remove noise and interference from the data, resulting in smoothed data. The feature extraction stage uses methods such as entropy analysis to extract useful features from the preprocessed data. These features are then used in subsequent fault diagnosis and prediction model construction. Furthermore, since the features may be high-dimensional, dimensionality reduction algorithms should be considered to reduce the dimensionality of the features and improve data processing efficiency.

[0129] ① Data preprocessing:

[0130] Data preprocessing is the first step in the feature extraction module. It smoothes the collected dynamic data before feature extraction to remove noise and interference to obtain relatively clean data. In this module, the present invention uses wavelet transform or independent component analysis (ICEEMDAN) algorithm to implement data preprocessing.

[0131] ② Feature extraction:

[0132] After data preprocessing, the present invention extracts meaningful features from the smoothed data, which will be used for subsequent fault diagnosis and fault prediction model construction. In the automaton dynamic data feature extraction module, the present invention uses entropy method, frequency domain feature and other methods to extract features.

[0133] ③ Feature dimensionality reduction:

[0134] Since features may be high-dimensional, this increases the complexity of data processing and may cause problems such as overfitting. In order to improve the efficiency of data processing and reduce the dimension of features, the present invention uses a dimensionality reduction algorithm such as principal component analysis (PCA) or linear discriminant analysis (LDA).

[0135] ④Data feature update:

[0136] Since dynamic data is continuous during the operation of the automaton, new data may be continuously collected. Therefore, to ensure the accuracy and timeliness of features, it is considered to implement a data feature update function. When new data is collected, it is preprocessed and feature extracted. The resulting features are then merged with existing features to ensure comprehensiveness.

[0137] (2) Fault diagnosis model: The collected dynamic data and features are used to train and establish a model to build an automatic fault diagnosis model. Algorithms such as Mustang are used to optimize the support vector machine parameters to improve the model's diagnostic accuracy and generalization ability. Based on the established fault diagnosis model, fault diagnosis is performed on the newly collected dynamic data, and the diagnostic results are output.

[0138] ①Model establishment:

[0139] In the automaton fault diagnosis module, the present invention adopts support vector machine (SVM) as the basic model for fault diagnosis. SVM is a nonlinear classifier that classifies samples by finding an optimal hyperplane.

[0140] In the automatic machine fault diagnosis, the present invention uses the characteristics of dynamic data as the characteristic vector of the sample and the fault category as the label of the sample to establish a fault diagnosis model.

[0141] Specifically, the process of establishing the SVM model is as follows:

[0142] Dataset partitioning: The collected dynamic data and features are divided into training sets and test sets. The training set is used for model training, and the test set is used to evaluate the model's performance.

[0143] Feature standardization: Standardize the features in the training set and test set to ensure that each feature has the same scale and avoid some features from having too much influence on the model.

[0144] Model training: Use the training set to train the SVM model to find the optimal hyperplane and classify the samples.

[0145] Model evaluation: Use the test set to evaluate the trained model, calculate the model's accuracy, recall rate, F1 value and other indicators, and evaluate the model's performance.

[0146] ② Parameter optimization: The parameters in the support vector machine model have a great influence on the performance and generalization ability of the model. Therefore, in the automatic machine fault diagnosis module, the present invention uses algorithms such as Mustang to optimize the support vector machine parameters.

[0147] The Wild Horse Optimizer (WHO) is a new intelligent optimization algorithm proposed in 2021. This algorithm primarily simulates the process of wild horses and other animals searching for food. It has strong optimization capabilities and fast convergence speed.

[0148] ③ Fault diagnosis: After the model is established and the parameters are optimized, the automatic machine fault diagnosis module can perform fault diagnosis on the newly collected dynamic data. The process includes: extracting features from the newly collected dynamic data to obtain feature vectors; using the established support vector machine model to classify the feature vectors to obtain the fault category; outputting the fault diagnosis results, displaying the fault category and confidence level.

[0149] (3) Fault prediction model: Based on the collected dynamic data and features, select an appropriate neural network structure and build the model. Divide the model into training and test sets, use the training set to train the model, and use the test set to evaluate the model. Predict the newly collected dynamic data and output the fault prediction results.

[0150] ① Model selection: The present invention adopts RBF neural network to predict automaton faults.

[0151] Specifically, the radial basis neural network (RBFNN) is based on a multidimensional space interpolation function, also known as a radial basis function. Each iteration of training only requires modifying fewer parameters, so the training efficiency is high and the local approximation effect can be achieved. The RBF network model is usually a three-layer network structure including an input layer, a hidden layer, and an output layer. Its structure is as follows: Figure 8 shown.

[0152] Figure 8 Where x is the input data, C i is the center of the i-th node in the hidden layer, and y is the final output of the model. The activation functions selected for the hidden layer nodes of the RBF neural network have various forms, all of which have local response characteristics.

[0153] RBF networks can approximate arbitrary nonlinear functions, can handle difficult-to-analyze regularities within the system, have good generalization capabilities, and have a fast learning convergence speed. They have been successfully applied to nonlinear function approximation, time series analysis, data classification, pattern recognition, information processing, image processing, system modeling, control, and fault diagnosis.

[0154] ② Data preparation: Before model training, the collected dynamic data needs to be preprocessed and feature extracted to facilitate the input of the neural network. Specifically, data preparation includes the following steps:

[0155] Data preprocessing: Preprocess the collected dynamic data to remove noise and interference to obtain smooth data.

[0156] Feature extraction: Extract meaningful features from the smoothed data, such as entropy value, frequency domain features, etc., as input to the neural network.

[0157] Data labeling: Label dynamic data and use fault categories as sample labels for training and evaluating prediction models.

[0158] Data partitioning: Divide the data into training sets and test sets. The training set is used to train the model, and the test set is used to evaluate the performance of the model.

[0159] ③ Model training: After data preparation is complete, model training begins. The goal of model training is to find the optimal network parameters so that the model's predictions on the training set are as close as possible to the true labels. Optionally, training methods used in the present invention include gradient descent and backpropagation algorithms.

[0160] ④ Model prediction: After model training is completed, the trained neural network model is used to predict the newly collected dynamic data. The prediction results will be output as the fault prediction results.

[0161] In this embodiment of the present invention, a pre-set digital twin model is used to simulate equipment operation, simulating both normal and faulty equipment processes, and forecasting equipment maintenance material needs. Within the pre-set digital twin simulation model, changes in equipment status are accounted for based on an equipment failure rate model. Simulating failure time using Monte Carlo simulation methods can yield relatively accurate forecasts.

[0162] Specifically, during the simulation process, it is necessary to set up the equipment usage mission plan, including:

[0163] (1) Terrain environment setting: Terrain environment setting is to set the terrain during the mission planning period. The terrain data such as elevation data, surface data, geological data, etc. are loaded through 3D GIS. The terrain of the mission area is set using polygonal frame method, and the calculation of sensitivity coefficients such as slope is supported.

[0164] (2) Meteorological environment settings: Meteorological environment settings are used to set the meteorological conditions during the mission plan. This can be done by importing historical meteorological data from previous years. At the same time, the weather can be customized for a specific period by entering information such as longitude, latitude, length, width, rainfall, start time, and end time. The system supports the calculation of sensitivity coefficients such as temperature, humidity, wind and sand, rainfall, and snowfall.

[0165] (3) Equipment use plan setting: The equipment use plan includes the mission name, deployment terrain, start and end time, upper equipment use start time, upper equipment use end time, chassis use start time, and chassis use end time.

[0166] In an embodiment of the present invention, simulation is performed using the Monte Carlo method. The Monte Carlo method, also known as the random sampling method, uses random numbers to solve computational problems, associates the problem to be solved with a certain probability model, and uses a sampling survey to obtain statistical values to infer unknown characteristic quantities. During computer simulation, random numbers uniformly distributed between (0, 1) are generated to simulate sampling and the occurrence of events. By averaging multiple simulations, simulation results close to reality can be achieved.

[0167] The occurrence of a replaceable unit failure is a probabilistic event. At any given moment, the probability of a unit failing is the unit's operational failure rate. Monte Carlo simulation can be used to simulate equipment failures during operation.

[0168] Figure 9 This is a demand forecast simulation flow chart provided according to an embodiment of the present invention. Figure 9 As shown in the figure, according to the basic idea of demand forecasting, the simulation process is refined to predict the demand for maintenance equipment.

[0169] (a) Read the mission phase plan and parse it to obtain information such as the mission organization, equipment, location, and mission start and end times.

[0170] (b) Load the equipment model according to the organization and equipment agreed in the mission plan, and use the currently collected data as the initial state of the equipment model.

[0171] (c) Load the environmental data for each stage of the mission, including location, weather conditions, and other information, according to the environment agreed upon in the mission plan.

[0172] (d) Entering simulation time T0, which is the starting time of the task plan.

[0173] (e) Based on the sensitivity coefficient system of each unit, combined with the current simulation task, composition and environmental data, calculate the sensitivity coefficient of each unit at that moment.

[0174] (f) Differentiate the categories (units with no failures, units with no state monitoring data failures, and units with state monitoring data failures) and calculate the usage failure rate of each unit at that moment.

[0175] (g) Using the Monte Carlo simulation method in combination with the failure rate to generate random numbers, the unit failure status at that moment is determined.

[0176] (h) Display the simulated situation, including the failure events of each equipment unit and the required maintenance equipment.

[0177] (i) Simulation time + 1, repeat (e) to (h) until the entire mission plan simulation is completed.

[0178] (j) Count the equipment requirements for this simulation maintenance and write them into the database.

[0179] (k) Repeat steps (a) to (h) to obtain multiple simulation results until the required number of simulations is completed.

[0180] (l) Analyze the simulation results into reports.

[0181] (m) Export simulation results.

[0182] From the above description, it can be seen that the present invention provides a method for predicting equipment maintenance equipment demand based on digital twins. Based on the digital twin model, it takes the current state of the equipment as the initial state, fully considers the equipment mission and environmental conditions, and uses the failure rate of use at each moment to simulate the equipment operation and maintenance process, and accurately predicts the demand for maintenance equipment, thereby preventing the shortage of some maintenance equipment and the squeeze of some maintenance equipment caused by the original guarantee model, and alleviates the technical problem of traditional equipment maintenance equipment demand prediction simulation that the simulation results are large different from the actual results.

[0183] Example 2

[0184] Figure 10 Schematic diagram of a system for forecasting equipment maintenance needs based on digital twins according to an embodiment of the present invention. Figure 10 As shown, the system includes: a collection module 10 , a synchronization module 20 , a first determination module 30 , a first establishment module 40 , a second establishment module 50 , a second determination module 60 and a statistics module 70 .

[0185] Specifically, the acquisition module 10 is used to collect historical data and status data of the target equipment, and obtain the actual operating status of each single device and each replaceable unit belonging to the target equipment;

[0186] The synchronization module 20 is used to synchronize the actual operating status of each single device and each replaceable unit to the preset digital twin model as the initial state of the simulation;

[0187] A first determination module 30 is configured to determine mission environment data based on a timetable of mission phases to be performed and an equipment usage plan;

[0188] A first establishing module 40 is configured to establish a spatiotemporal relationship model of equipment status characteristics based on historical data of the target equipment and an inherent failure rate model of a replaceable unit of the target equipment;

[0189] The second establishment module 50 is used to input the equipment status feature spatiotemporal relationship model and mission environment data into the preset digital twin model, calculate the sensitivity coefficient at the current moment, and build an equipment use failure rate model based on the sensitivity coefficient at the current moment;

[0190] The second determining module 60 is used to determine the unit failure status at the simulation time corresponding to the equipment use plan based on the equipment use failure rate model;

[0191] The statistics module 70 is used to collect statistics on the maintenance equipment requirements of the target equipment based on the unit failure conditions at the simulation time.

[0192] The present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method provided in the embodiment of the present invention when executing the computer program.

[0193] The present invention also provides a computer-readable storage medium, which stores computer instructions. When the computer instructions are executed by a processor, the method provided in the embodiment of the present invention is implemented.

[0194] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

[0195] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A method for predicting equipment maintenance equipment demand based on digital twins, characterized in that: include: Collect historical data and status data of the target equipment, and obtain the actual operating status of each individual equipment and each replaceable unit belonging to the target equipment; Synchronize the actual operating status of each single device and each replaceable unit to the preset digital twin model as the initial state of the simulation; Determine mission environment data based on the mission phase schedule and equipment utilization plan to be performed; Establishing a spatiotemporal relationship model of equipment status characteristics based on historical data of the target equipment and an inherent failure rate model of a replaceable unit of the target equipment; Inputting the equipment status feature spatiotemporal relationship model and the task environment data into the preset digital twin model, calculating the current moment sensitivity coefficient, and constructing an equipment use failure rate model based on the current moment sensitivity coefficient; Determining the unit failure status at the simulation time corresponding to the equipment usage plan based on the equipment usage failure rate model; Based on the unit failure conditions at the simulation time, maintenance equipment requirements of the target equipment are counted.

2. The method according to claim 1, wherein: The historical data of the target equipment includes: equipment usage record data, equipment failure record data and equipment maintenance record data; the status data of the target equipment includes: equipment composition structure data and equipment status data.

3. The method according to claim 1, wherein: The mission environment data includes: mission terrain environment data and mission meteorological environment data.

4. The method according to claim 1, wherein: The inherent failure rate module includes: a constant failure rate model, a normal distribution model, and a Weibull distribution model.

5. The method according to claim 1, wherein: The usage failure rate model includes: a usage failure rate model of a failure unit without state monitoring data and a usage failure rate model of a failure unit with state monitoring data.

6. The method according to claim 5, characterized in that: The stateless monitoring data failure unit uses a failure rate model, including: l a =kλ Where λ a is the equipment failure rate, k is the sensitivity coefficient, k 10 is the task intensity coefficient, k 20 is the operating environment coefficient, k 30 is the personnel level coefficient, k 40 is the protection factor, and λ is the inherent failure rate of the equipment.

7. The method according to claim 5, characterized in that: The stateful monitoring data fault unit uses a failure rate model, including: a data feature extraction model, a fault diagnosis model and a fault prediction model; wherein, The data feature extraction model is used to obtain dynamic data of the equipment to be diagnosed during operation, and to perform preprocessing, feature extraction, and feature dimension reduction on the dynamic data to obtain frequency domain features of the dynamic data; The fault diagnosis model includes a support vector machine model trained and constructed based on the dynamic data and the frequency domain characteristics of the dynamic data, and is used to perform fault diagnosis on the equipment to be diagnosed; The fault prediction model includes a radial basis function neural network model trained and constructed based on the dynamic data and the frequency domain characteristics of the dynamic data, and is used to predict faults of the equipment to be diagnosed.

8. A digital twin-based equipment maintenance equipment demand forecasting system, characterized by: include: Acquisition module, synchronization module, first determination module, first establishment module, second establishment module, second determination module and statistics module; wherein, The acquisition module is used to collect historical data and status data of the target equipment, and obtain the actual operating status of each single device and each replaceable unit belonging to the target equipment; The synchronization module is used to synchronize the actual operating status of each single device and each replaceable unit to the preset digital twin model as the initial state of the simulation; The first determination module is configured to determine mission environment data based on a timetable of a mission phase to be performed and an equipment usage plan; The first establishing module is used to establish a spatiotemporal relationship model of equipment status characteristics based on historical data of the target equipment and an inherent failure rate model of a replaceable unit to which the target equipment belongs; The second establishment module is used to input the equipment status feature spatiotemporal relationship model and the task environment data into the preset digital twin model, calculate the current moment sensitivity coefficient, and construct an equipment use failure rate model based on the current moment sensitivity coefficient; The second determining module is used to determine the unit failure status at the simulation time corresponding to the equipment usage plan based on the equipment usage failure rate model; The statistical module is used to collect statistics on the maintenance equipment requirements of the target equipment based on the unit failure status at the simulation time.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images