Maintenance equipment information prediction method and device based on digital twinborn model

Abstract

The invention relates to a maintenance equipment information prediction method and device based on a digital twin model. The method comprises the following steps: acquiring record, detection and environment information of equipment; according to the data, the system determines a three-layer description model of digital twinning and generates physical twinning state information of the equipment through interactive description. And predicting faults based on the physical twin state information by using the equipment maintenance physical prediction model and the machine prediction model, and obtaining related maintenance equipment fault information. And carrying out isolation prediction on the fault information, and combining feature extraction and data integration to obtain prediction information of each maintenance device. And the optimal maintenance equipment is selected from the prediction information in combination with the application scene and the efficiency parameter, so that the selected equipment is ensured to efficiently and accurately meet the requirements in the specific scene, and finally the optimal maintenance equipment information is obtained. By adopting the method, the accuracy of maintenance material preparation can be effectively improved, so that the maintenance efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method and device for predicting maintenance equipment information based on a digital twin model. Background Art

[0002] With the development of computer technology, maintenance equipment prediction technology has emerged. This technology usually uses the operating status and maintenance records of the equipment to determine in advance which parts may have problems, so as to prepare the necessary equipment in advance. This can not only reduce the waiting time for maintenance, but also help companies better manage inventory and ensure that maintenance work can proceed smoothly.

[0003] In traditional technology, maintenance personnel roughly predict possible failures and required equipment based on the equipment's operating conditions, service life, and historical failure records, and further use the manually recorded data mentioned above. Therefore, they often rely on experience and fixed maintenance cycles to determine equipment requirements. The lack of real-time data support and difficulty in extracting features from recorded data result in inaccurate preparation of maintenance materials, which in turn affects maintenance efficiency. Summary of the invention

[0004] Based on this, it is necessary to provide a maintenance equipment information prediction method, device, computer equipment, computer-readable storage medium and computer program product based on a digital twin model, which can effectively improve the accuracy of maintenance material preparation and thus improve maintenance efficiency, in response to the above technical problems.

[0005] In a first aspect, the present application provides a maintenance equipment information prediction method based on a digital twin model, comprising: Obtain equipment record information, equipment detection information, and equipment environment information of the target equipment; Determine each digital twin description layer of the object equipment according to the equipment record information, the equipment detection information and the equipment environment information; According to the equipment record information, the equipment detection information and the equipment environment information, interactive description is performed at each digital twin description layer to obtain equipment physical twin state information; Using the equipment maintenance physical prediction model and the equipment maintenance machine prediction model respectively, according to the equipment physical twin state information, the fault information of the target equipment is predicted to obtain the fault information of each maintenance equipment; According to each of the maintenance equipment failure information, information isolation prediction is performed on the maintenance equipment information of the target equipment to obtain each maintenance equipment prediction information; According to the application scenario parameters and application efficiency parameters of the target equipment, optimal information is selected from each of the maintenance equipment prediction information to obtain optimal maintenance equipment information.

[0006] In the second aspect, the present application also provides a maintenance equipment information prediction device based on a digital twin model, comprising: An equipment data acquisition module is used to acquire equipment record information, equipment detection information and equipment environment information of the target equipment; A twin model determination module, used to determine each digital twin description layer of the target equipment according to the equipment record information, the equipment detection information and the equipment environment information; A twin model description module, used to interactively describe the equipment at each digital twin description layer according to the equipment record information, the equipment detection information and the equipment environment information, so as to obtain the equipment physical twin state information; A fault information prediction module is used to use the equipment maintenance physical prediction model and the equipment maintenance machine prediction model to predict the fault information of the target equipment according to the equipment physical twin state information, and obtain the fault information of each maintenance equipment; A maintenance equipment prediction module, used to perform information isolation prediction on the maintenance equipment information of the target equipment according to the failure information of each maintenance equipment, and obtain prediction information of each maintenance equipment; The maintenance equipment determination module is used to select the best information from each maintenance equipment prediction information according to the application scenario parameters and application efficiency parameters of the target equipment to obtain the best maintenance equipment information.

[0007] The above maintenance equipment information prediction method and device based on the digital twin model can accurately reflect the actual working status of the equipment under different environmental conditions by constructing the digital twin description layer of the equipment, and then monitor and diagnose the operation of the equipment in real time. Further with the help of the equipment maintenance physical prediction model and machine learning prediction model, it is possible to intelligently predict potential fault points based on the physical twin state information of the equipment, and optimize the selection of maintenance equipment through information isolation prediction, which can not only effectively improve the accuracy of maintenance material preparation, thereby improving maintenance efficiency, but also improve the accuracy of fault prediction, identify and isolate possible fault information in advance, thereby reducing the maintenance time and cost caused by sudden failures, and significantly improving the maintainability and operation safety of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0009] Figure 1This is an application environment diagram of a maintenance equipment information prediction method based on a digital twin model in an embodiment; Figure 2 It is a flowchart of a maintenance equipment information prediction method based on a digital twin model in one embodiment; Figure 3 It is a flowchart of a method for obtaining maintenance equipment fault information in one embodiment; Figure 4 A schematic diagram of a flow chart of a method for obtaining physical twin state information of equipment in one embodiment; Figure 5 A flowchart of a method for obtaining maintenance equipment prediction information in one embodiment; Figure 6 A flowchart of a method for obtaining optimal maintenance equipment information in one embodiment; Figure 7 It is a flowchart of a method for obtaining optimal maintenance equipment information in another embodiment. DETAILED DESCRIPTION

[0010] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0011] The maintenance equipment information prediction method based on the digital twin model provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. Among them, the server 104 can be implemented with an independent server or a server cluster composed of multiple servers.

[0012] In an exemplary embodiment, Figure 2 As shown in the figure, a maintenance equipment information prediction method based on a digital twin model is provided. Figure 1 The server in the example is used to illustrate, including the following steps 202 to 212. Among them: Step 202: Acquire equipment record information, equipment detection information, and equipment environment information of the target equipment.

[0013] The object equipment can be a specific mechanical device or system that needs to be monitored, maintained and optimized. It can be any equipment in the fields of industry, transportation, energy, etc., such as power equipment, manufacturing machinery, vehicles or aircraft.

[0014] The equipment record information may be historical data of the target equipment, including production information, maintenance history, fault records, component replacement cycles, maintenance plans, etc. of the equipment.

[0015] Among them, equipment detection information can be real-time data about the current operating status of the equipment obtained through sensors, detection equipment or manual inspections, including parameters such as temperature, pressure, vibration, flow, current, power, etc.

[0016] The equipment environment information may be the external environment conditions of the equipment, such as temperature, humidity, air pressure, vibration, pollution and other factors.

[0017] Specifically, various basic data of the target equipment are collected through sensors and automated detection systems, including equipment record information, equipment detection information, and equipment environment information. Equipment record information usually includes equipment composition structure data, equipment usage record data, equipment fault record data, and equipment maintenance record data of the target equipment, such as fault occurrence history, maintenance records, and component replacement cycles. Equipment detection information includes real-time data of the equipment during actual use, such as temperature, pressure, vibration, power consumption, etc. Equipment environment information involves various parameters of the working environment of the equipment, such as humidity, temperature, load conditions, and external pressure in the workplace. These external factors directly affect the operating performance and life of the equipment.

[0018] Step 204: Determine each digital twin description layer of the target equipment based on the equipment record information, equipment detection information, and equipment environment information.

[0019] Among them, the digital twin description layer can be a component of the digital twin model, which is used to express different dimensions or levels of equipment, including the digital twin equipment layer (physical structure and performance layer), the digital twin system layer (control and operation layer), and the digital twin environment layer (external environment impact layer).

[0020] Specifically, since digital twins are technologies that digitally represent physical equipment and can simulate the behavior and status of equipment, they can be decomposed into multiple different digital twin description layers for description based on the complexity and diversity of equipment record information, equipment detection information, and equipment environment information. Further, based on the equipment record information, equipment detection information, and equipment environment information, a layer suitable for describing each of the above information is selected from the digital twin model created based on historical information. For example, the digital twin equipment layer, the digital twin system layer, and the digital twin environment layer can be selected from the digital twin model. For example, the digital twin equipment layer focuses on the physical characteristics of the equipment, monitors the mechanical components, structural changes, and performance data of the equipment in real time, such as temperature, pressure, vibration, etc., to reflect the current working status of the equipment. The digital twin system layer involves the control and information processing system inside the equipment, simulates the control logic, sensor feedback, and actuator operation of the equipment, and helps understand the operating behavior and intelligent decision-making process of the equipment at the system level. The digital twin environment layer takes into account the impact of external factors on the equipment, such as the temperature, humidity, and air pressure of the working environment, which directly or indirectly affect the operating efficiency and health status of the equipment.

[0021] Step 206: According to the equipment record information, equipment detection information and equipment environment information, interactive description is performed in each digital twin description layer to obtain the equipment physical twin state information.

[0022] Among them, the interactive description can be the information exchange and fusion between the various description layers of the digital twin model to construct a comprehensive physical twin state of the equipment.

[0023] Among them, the physical twin status information of the equipment can be the current operating status of the equipment obtained through interactive analysis of each description layer based on the digital twin model. This information includes the mechanical, control and environmental status of the equipment, which can reflect the health status of the equipment, potential problems and its operating performance.

[0024] Specifically, the equipment record information, equipment detection information and equipment environment information are input into the digital twin equipment layer, digital twin system layer and digital twin environment layer respectively. For the digital twin equipment layer, the physical structure and performance status data of the object equipment are used to monitor the operation status of the equipment in real time, such as the temperature, vibration, pressure and other information of the mechanical parts, and the health status of the equipment is evaluated through the data collected by sensors and real-time detection equipment; further analysis can reveal the wear, deviation or other potential faults of the equipment. The digital twin system layer integrates the control system data inside the equipment, analyzes the operation logic, actuator response, sensor feedback and execution of the control command of the equipment, and helps to determine whether the system operates according to the predetermined logic, as well as the control deviation or fault that may exist in the system. For example, when the control signal does not match the sensor feedback, it may mean that a module in the system has failed or a precursor to a failure has appeared. Finally, the digital twin environment layer provides relevant data on the external environment of the equipment, such as temperature, humidity, air pressure, etc. These external factors can have a direct impact on the performance of the equipment. For example, high temperature may cause overheating of components, and high humidity may cause electrical failures. By interactively integrating the data of the digital twin equipment layer, digital twin system layer, and digital twin environment layer, the operating status of the equipment can be comprehensively evaluated to form a complete physical twin status information of the equipment. The physical twin status information of the equipment can not only accurately reflect the current health status of the equipment, but also predict future fault trends based on historical data and real-time data, thereby providing data support for subsequent fault diagnosis and maintenance decisions.

[0025] Step 208 , respectively use the equipment maintenance physical prediction model and the equipment maintenance machine prediction model to predict the fault information of the target equipment according to the equipment physical twin state information, and obtain the fault information of each maintenance equipment.

[0026] Among them, the equipment maintenance physical prediction model can be a physical model used to predict equipment failure based on the physical characteristics, usage history and operating environment of the equipment.

[0027] Among them, the equipment maintenance machine prediction model can be a model based on a machine learning algorithm that uses historical failure data, equipment usage patterns and real-time detection information to predict equipment failures.

[0028] The fault information may be data related to any abnormal state or problem that may occur during the operation of the equipment or in the future.

[0029] The maintenance equipment failure information may be failure data of tools, components or equipment related to the equipment.

[0030] Specifically, the equipment maintenance physical prediction model and equipment maintenance machine prediction model are used respectively to predict the fault information of the target equipment according to the physical twin state information of the equipment. The equipment maintenance physical prediction model includes an inherent fault calculation model, a usage fault calculation model and an environmental stress fault calculation model, while the equipment maintenance machine prediction model includes a convolutional neural network (CNN), a long short-term memory network (LSTM), a deep belief network (DBN), etc. For the inherent fault calculation model, based on the design and manufacturing characteristics of the equipment, the equipment analyzes the possible faults that may occur due to its own structure or material defects under normal operation. This type of model usually focuses on the long-term wear, aging and material fatigue of the equipment, and predicts the inherent faults that may occur in the equipment without external interference. For the usage fault calculation model, it focuses on the failure mode of the equipment under specific working conditions, taking into account the influence of factors such as load, operation mode and frequent start and stop on the equipment in actual application, and predicts the possible faults that may occur during its use. For example, overload or frequent high-load operation may cause early damage to certain components. The environmental stress failure calculation model takes the external environmental conditions of the equipment into consideration, analyzes the impact of external stress factors such as high temperature, high humidity, and severe vibration on the equipment, and predicts failures caused by environmental factors. These models help identify equipment failures caused by changes in external conditions.

[0031] At the same time, using convolutional neural networks (CNN), long short-term memory networks (LSTM), deep belief networks (DBN), etc., complex equipment data can be deeply learned and pattern recognized to further improve the accuracy of fault prediction. For convolutional neural networks (CNN), by using convolutional layers to extract spatial features in equipment data, it is suitable for analyzing fault modes related to image or time series data. For example, CNN can be used to process vibration waveforms, heat maps or other time series data collected from equipment sensors to identify potential fault features. For long short-term memory networks (LSTM), it is good at processing time series data and can effectively capture long-term dependent time series information during equipment operation; through memory units, LSTM can analyze the historical status of equipment at different time points and identify potential problems caused by historical faults or abnormal behaviors, which is particularly suitable for predicting the timing of equipment failures and fault development trends. For deep belief networks (DBN), it is usually used to extract features and perform pattern recognition from a large amount of unlabeled data. It can identify complex correlation patterns in equipment status through multi-layer nonlinear transformations, which is particularly suitable for complex and multi-dimensional equipment data. It obtains high-level abstract features through layer-by-layer learning, thereby predicting potential fault information.

[0032] Step 210, based on the failure information of each maintenance equipment, information isolation prediction is performed on the maintenance equipment information of the target equipment to obtain prediction information of each maintenance equipment.

[0033] Among them, information isolation prediction can be a technical method for multiple fault information, which can be achieved by combining the characteristic data of different maintenance equipment or fault types with the original data for independent packaging, separate processing and analysis to avoid information interference or cross-influence.

[0034] The maintenance equipment prediction information may be forecast data of equipment demand related to equipment maintenance obtained through fault prediction model analysis.

[0035] Specifically, for each maintenance equipment fault information, the system will extract relevant fault features, such as the frequency of fault occurrence, historical maintenance records, environmental factors (such as temperature, humidity, etc.), and key parameters such as equipment usage; at the same time, the system will package the fault feature information corresponding to each maintenance equipment fault information, and each type of fault feature information corresponds to different maintenance equipment. For example, some faults may involve the maintenance of electronic control systems, while other faults may be related to the damage of mechanical components. Each packaged data packet can display the fault features corresponding to each maintenance equipment fault information, that is, each maintenance equipment fault information and the corresponding fault feature information will be isolated into independent data packets. The prediction model matched by each data packet will independently analyze and predict the occurrence trend and severity of the fault, determine whether maintenance is required, and evaluate the equipment required for maintenance. At this time, each prediction model works in parallel with each other, does not affect each other, and independently completes the fault prediction of different maintenance equipment. Through the parallel processing system, personalized prediction information can be provided for each maintenance equipment quickly and accurately, thereby avoiding redundant predictions or data interference, ensuring that the needs of each maintenance equipment are fully analyzed and optimized, and obtaining prediction information for each maintenance equipment.

[0036] Step 212, according to the application scenario parameters and application efficiency parameters of the target equipment, the optimal information is selected from the maintenance equipment prediction information to obtain the optimal maintenance equipment information.

[0037] Among them, the application scenario parameters can be the specific operating conditions of the equipment in a specific working environment, such as workload, operating frequency, ambient temperature, humidity, interaction, personnel protection, etc.

[0038] Among them, the application efficiency parameter can be used to evaluate the effectiveness and efficiency of the maintenance process in a specific application scenario, including factors such as the time required for maintenance, cost-effectiveness, maintenance degree, and repair completeness.

[0039] The optimal maintenance equipment information may be the most suitable maintenance tools or parts selected from a plurality of optional maintenance equipment under a specific application scenario and efficiency requirements.

[0040] Specifically, the system obtains the application scenario parameters of the target equipment, where the application scenario parameters describe the specific working environment, operating conditions and interaction with the external environment of the equipment, such as the temperature, humidity, load, pressure, vibration, whether there is danger, whether it needs to be used with faults, etc. in the workplace. These parameters reflect the operating status of the equipment in different working environments and directly affect the selection of tools during the maintenance process. For example, in high temperature or extreme environments, specific protective tools or high temperature resistant materials may be required; in high-load operation scenarios, stronger and more wear-resistant maintenance equipment may be required. Therefore, the system will select the most suitable maintenance equipment based on these scenario characteristics to adapt to the current working conditions. At the same time, the system also obtains application efficiency parameters, that is, the maintenance efficiency requirements in specific working scenarios. These parameters include not only the time efficiency and cost-effectiveness of maintenance, but also the degree of maintenance and the completeness of repair in the current scenario. For example, in some scenarios, equipment failures may require simple local repairs based on the needs of the current scenario, while in other cases, they may require complete repairs or replacement of multiple parts based on the needs of other scenarios. The system will evaluate the effectiveness of different maintenance equipment in the repair process based on the complexity of the task and the urgency of the repair of the target equipment, and consider the completeness of the repair (such as whether the original performance of the equipment needs to be fully restored, or only temporary repairs are performed to ensure short-term operation). Based on the application scenario parameters and application efficiency parameters, the system will combine these factors to select the most suitable equipment for the current maintenance task from the maintenance equipment prediction information generated by each parallel prediction model to form the optimal maintenance equipment information.

[0041] In the above maintenance equipment information prediction method based on the digital twin model, by constructing the digital twin description layer of the equipment, the actual working status of the equipment under different environmental conditions can be accurately reflected, and the operation of the equipment can be monitored and diagnosed in real time. Further, with the help of the equipment maintenance physical prediction model and the machine learning prediction model, the potential fault points can be intelligently predicted according to the physical twin state information of the equipment, and the selection of maintenance equipment can be optimized through information isolation prediction, which can not only effectively improve the accuracy of maintenance material preparation, thereby improving maintenance efficiency, but also improve the accuracy of fault prediction, identify and isolate possible fault information in advance, thereby reducing the maintenance time and cost caused by sudden failures, and significantly improving the maintainability and operation safety of the equipment.

[0042] In an exemplary embodiment, Figure 3 As shown, each equipment maintenance physical prediction model and equipment maintenance machine prediction model are used respectively to predict the fault information of the target equipment according to the physical twin state information of the equipment, and the fault information of each maintenance equipment is obtained, including steps 302 to 306. Among them: Step 302 , using each equipment maintenance physical prediction model respectively, predict the maintenance equipment information of the target equipment according to the equipment physical twin state information, and obtain the equipment fault information of each physical model.

[0043] The physical model equipment fault information may be fault data obtained based on physical principles and the working environment of the equipment when the equipment fault is predicted by a physical modeling method.

[0044] Among them, the inherent fault calculation model can be a calculation model that focuses on the inherent defects in the design and manufacturing stages of the equipment, or the faults that are bound to occur after a certain period of use. These defects usually do not depend on the use of the equipment or external environmental factors, but are derived from the design defects of the equipment itself, material defects or problems that exist during long-term use.

[0045] The usage fault calculation model may be a calculation model that focuses on the impact of factors such as load, usage frequency, and operation mode of the equipment in actual operation on equipment failure.

[0046] Among them, the environmental stress failure calculation model can be a calculation model that focuses on the failure risk of equipment under specific environmental conditions, especially those caused by external environmental factors such as temperature, humidity, corrosive substances, vibration, etc.

[0047] Specifically, since the physical prediction models for equipment maintenance are respectively the inherent fault calculation model, the usage fault calculation model and the environmental stress fault calculation model, the physical twin state information of the equipment is respectively input into the inherent fault calculation model, the usage fault calculation model and the environmental stress fault calculation model for prediction.

[0048] The inherent fault calculation model also focuses on defects that may exist in the design and manufacturing stages of the equipment, especially failures caused by improper material properties, processing technology or design of components. These problems are usually hidden and difficult to show up in the early stages of the equipment, but they may gradually be exposed over time. For example, some mechanical parts may crack or suffer material fatigue in a high-stress environment due to improper material selection or minor defects in production. The inherent fault calculation model deeply analyzes the equipment design, the mechanical properties of the material and its manufacturing process, combined with the historical use of the equipment, to predict the failure of the equipment due to inherent factors in the absence of external excessive load or abnormal operation.

[0049] For the inherent fault calculation model, the calculation formula of the inherent fault calculation model is: Where N(t) is the remaining life at time t, A is the material fatigue characteristic constant, which is obtained by obtaining the basic fatigue characteristics of the material under different stresses through laboratory material testing, such as fatigue testing, E is the material elastic modulus, which is obtained by directly measuring the elastic modulus of the material through tensile testing or other mechanical testing methods, b and c are fitting coefficients, which are empirical coefficients obtained by regression analysis using historical failure data and experimental data, and F i It is an environmental factor, which is calculated by collecting environmental data in real time through environmental sensors (such as temperature, humidity, and corrosion sensors) installed on the equipment. is the sensitivity coefficient of the environmental factor. By controlling the environmental factor to conduct experiments and observing its effect on the material life, the sensitivity coefficient is fitted. m is the number of environmental factors and D(t) is the time. Cumulative fatigue damage is obtained by calculating the microscopic damage index of the material using a material damage model (such as a fatigue damage model). max The maximum damage resistance of the material is determined by the material limit fatigue test to determine the maximum damage value that the material can withstand. is the microstructure damage influence coefficient, which is obtained by fitting the life change under different microstructure damage degrees through fatigue test or stress test. For time The stress is obtained by FEA simulation of the stress distribution of the equipment under different working conditions. is the reference stress of the material, which is determined by experiments under standard conditions. k is the fatigue strength index, which is determined by fatigue test data using regression analysis or nonlinear fitting methods. The environmental impact index is estimated based on environmental data and damage data using multiple regression analysis. is the material microstructure change function, which is generated by simulating the microstructure change using material microstructure simulation software (such as MOOSE, Abaqus). is the inherent failure rate at time t, is the life span change rate.

[0050] The fault calculation model used is based on the actual operating conditions of the equipment, taking into account the impact of factors such as workload, frequency of use, and operating habits on the performance of the equipment during operation. The focus of this model is to evaluate the failures that may be caused by excessive use, incorrect operation, or high-load operation of the equipment in actual applications. For example, when the equipment is under high load or works continuously for a long time, some components (such as motors, transmission systems, bearings, etc.) will fail prematurely due to excessive wear or fatigue. The fault calculation model is used to monitor the real-time working status of the equipment, combined with parameters such as load, operating speed, and working cycle, to identify and predict possible failures of the equipment under specific conditions of use. This method helps to formulate maintenance plans based on actual usage and ensure that the equipment remains in good operating condition under high load or high frequency use.

[0051] For the fault calculation model, the calculation formula for the fault calculation model is: in, is the failure rate at time t, is the benchmark failure rate, which is determined by counting the failure rate under standard operating conditions from the historical failure records of the equipment. P(t) is the actual operating load at time t, which is measured in real time by the load sensor. max The maximum design load of the equipment is determined by referring to the design specifications or technical manual of the equipment. is the load sensitivity coefficient, which is determined by regression analysis using historical operating load and failure rate data. H(t) is the time The operation health index is calculated by combining multiple health indicators (such as temperature, vibration, wear degree, etc.) through weighting or machine learning methods. is the sensitivity coefficient of the health index to the failure rate, which is estimated and determined through regression analysis based on the historical health index and failure rate data. is the failure rate after the combined operation behavior at time t, is the jth operation behavior feature at time t, extracting operation behavior data from the operating system or control system, such as start frequency, number of overload operations, number of emergency stops, and other determined feature data, is the influence coefficient of the jth operation on the failure rate, which is determined by multiple regression analysis using historical operation behavior data and failure rate data. n is the number of operation behavior features. is the usage failure rate of the combined operating load, operating behavior and health status at time t.

[0052] The environmental stress failure calculation model focuses on the stress that the external environment puts on the equipment, especially the challenges that the equipment may face in extreme weather or harsh working environments. Equipment may encounter various stress sources under different environmental conditions, such as temperature fluctuations, moisture, dust, chemical corrosion, etc., which will have a long-term impact on the physical properties of the equipment. For example, at high temperatures, electronic components inside the equipment may fail due to thermal expansion, or metal parts are prone to corrosion in a high humidity environment. The environmental stress failure calculation model predicts the failure mode of the equipment under specific environmental pressure by analyzing the characteristics of the environment in which the equipment is located, such as climate conditions, humidity, pollutants, etc. The model helps optimize the equipment's ability to resist environmental stress during the design phase, and guides how to take targeted protection measures to extend the service life of the equipment during the maintenance process.

[0053] For the environmental stress failure calculation model, the calculation formula of the environmental stress failure calculation model is: in, is the failure rate under the influence of environmental stress at time t, is the basic failure rate when the environment has no influence, which is measured by conducting experiments under controlled environmental conditions. k (t) is the kth environmental stress factor at time t, and the stress factor is monitored in real time by installing corresponding sensors (such as temperature sensors, humidity sensors, salt spray sensors, etc.), E max,k is the maximum design value of the kth environmental stress factor, obtained by referring to the design specifications or technical manuals of the equipment. is the sensitivity coefficient of the kth environmental stress factor, determined by regression analysis using historical environmental stress data and failure rate data, p is the number of environmental stress factors, D 环境 (t) is the accumulated damage of environmental stress at time t, E k (t') is the kth environmental stress factor at time t, and the stress factor is monitored in real time by installing corresponding sensors (such as temperature sensors, humidity sensors, salt spray sensors, etc.), is the damage attenuation coefficient of the kth environmental stress factor, which is determined by regression analysis using historical environmental stress data and damage data. is the failure rate after time t combined with environmental stress, is the cumulative impact coefficient of environmental stress on failure rate. It is determined by regression analysis using historical environmental stress cumulative damage data and failure rate data.

[0054] Step 304 , respectively use the maintenance machine prediction model of each equipment to predict the maintenance equipment information of the target equipment according to the physical twin state information of the equipment, and obtain the equipment fault information of each machine model.

[0055] Among them, machine model equipment failure information can be failure information obtained based on machine learning or artificial intelligence model analysis. Unlike physical models, machine learning models predict possible equipment failures by learning and training on a large amount of historical operating data.

[0056] Specifically, since the equipment maintenance machine prediction models are convolutional neural networks, long short-term memory networks, and deep belief networks, the physical twin state information of the equipment is respectively input into the convolutional neural network, long short-term memory network, and deep belief network for prediction.

[0057] Among them, convolutional neural networks (CNNs) are deep learning models suitable for processing signals with spatial structures and local correlations. In equipment failure prediction, CNNs are mainly used to automatically extract features from sensor data and identify fault patterns. Through convolution operations, CNNs can capture local patterns of signals during equipment operation, especially in signals such as vibration, sound, and temperature, which may reflect the slight differences between normal operation and failure of equipment. CNNs are good at identifying abnormal waveforms, frequency changes, and irregular patterns in signals, so when processing multidimensional signals or time series signals, they can effectively extract potential hidden fault modes of equipment. For example, CNNs can perform convolution processing on equipment vibration data to discover mutations or anomalies in its spectrum, which are often precursors to equipment failure.

[0058] Among them, the long short-term memory network (LSTM) is specifically used to process time series data and can capture long-term dependencies in the data. Through its unique gating mechanism, LSTM can effectively remember and transmit important information in the time series, avoiding the gradient vanishing problem encountered by traditional recurrent neural networks (RNN) in long sequences. In equipment failure prediction, LSTM can identify the trend changes of equipment in long-term operation by analyzing the historical operation data of the equipment, such as temperature changes, vibration intensity, pressure fluctuations, etc. For example, LSTM can identify the gradual increase in temperature or vibration over a long period of time, predict the time when the equipment may fail, and issue an alarm in advance. This model is very suitable for handling failure modes with time dependencies, especially when the equipment has progressive failures or cumulative damage, LSTM can issue early warnings through long-term data analysis before the failure appears.

[0059] Among them, the deep belief network (DBN) learns the neural network model of complex data representation through layers of nonlinear transformation. The structure of DBN enables it to mine deep potential features from large amounts of complex historical data, and is particularly suitable for processing multidimensional data. In equipment failure prediction, DBN can automatically extract and model the failure characteristics of different components during equipment operation through a multi-layer unsupervised training process. This enables DBN to effectively learn and identify potential failure signals even when the data is very diverse, there are multiple interference factors, or the failure mode is complex. For example, in the face of different types of equipment, different failure modes, and environmental interference, DBN can automatically identify the complex failure modes of equipment in a specific working environment through deep feature learning, and predict possible systemic failures in advance. Since DBN can extract complex nonlinear features from large-scale historical data, it can provide more accurate prediction and decision support when processing equipment failure prediction.

[0060] Step 306, using the equipment parts characteristic information of the target equipment as the guide information, the consistency of each physical model equipment fault information and each machine model equipment fault information is identified to obtain each maintenance equipment fault information.

[0061] The equipment accessory feature information may be detailed information about each accessory or component in the equipment, such as the model, material, purpose, working environment, service life, manufacturer, etc. of the accessory.

[0062] Among them, consistency identification can be a cross-comparison and analysis between the fault information generated by multiple prediction models to ensure the consistency and accuracy of the results output by different models in actual situations.

[0063] Specifically, since the fault information predicted by the physical model and the machine model may involve the same or similar maintenance equipment, the equipment accessory feature information of the object equipment is used as the guiding information for consistency identification, that is, the equipment accessory feature information of the object equipment is used as the error calculation item for limiting consistency identification. The output results of the physical model and the output results of the machine learning model are carefully compared according to the equipment accessory feature information of the accessory, and the actual situation of the maintenance equipment involved in each prediction result (such as the material, durability, working conditions, etc. of the accessory) is comprehensively considered. For example, if the physical model predicts that a certain accessory may fail due to excessive environmental stress, and the machine learning model predicts that the accessory has problems due to excessive wear, through the consistency matching of the accessory characteristics, the system can determine whether the two predictions are for the same failure mode or different failure risks caused by different reasons. If all are consistent, the fault information of each maintenance equipment is directly output. If not consistent, the fault information of each maintenance equipment that is temporarily consistent is output, and the physical model and neural network corresponding to the inconsistent information are used to re-predict until they are consistent, and the fault information of each maintenance equipment processed later is output.

[0064] In this embodiment, by combining the equipment maintenance physical prediction model and the machine prediction model, the fault information of the maintenance equipment is predicted. The physical model can identify potential faults from the operating status and environmental stress of the equipment, while the machine learning model can mine the complex patterns in the historical data. Through the combination of the two, the possible faults of the maintenance equipment can be predicted comprehensively and accurately. At the same time, through consistency identification, the system can eliminate conflicts or repeated predictions between different models, ensure the uniformity and efficiency of the maintenance equipment fault information, and provide more reliable data support for subsequent maintenance decisions and resource allocation, thereby improving the maintenance efficiency and reliability of the equipment.

[0065] In an exemplary embodiment, Figure 4 As shown, according to the equipment record information, equipment detection information and equipment environment information, interactive description is performed at each digital twin description layer to obtain the equipment physical twin state information, including steps 402 to 410. Among them: In step 402, the equipment record information, equipment detection information and equipment environment information are respectively input into the digital twin equipment layer, the digital twin system layer and the digital twin environment layer to obtain the real-time operation status data of the equipment, the overall stability data of the system operation and the environmental impact simulation data.

[0066] Among them, the digital twin equipment layer can be a digital twin working layer that is digitally modeled and driven by real-time data to virtually reproduce the working status and operation process of the actual equipment.

[0067] Among them, the digital twin system layer can be a digital reproduction of the entire system, including multiple devices, subsystems, and their interrelationships and collaborative work, a digital twin working layer.

[0068] Among them, the digital twin environment layer can be a digital twin working layer that simulates and analyzes the impact of external environmental factors on equipment and systems.

[0069] Among them, the real-time operating status data of the equipment can be the real-time monitoring data obtained after the digital twin equipment layer simulates the equipment, including various sensor data of the equipment (such as temperature, pressure, vibration, etc.), which is used to reflect the health status and operating performance of the equipment at the current moment.

[0070] Among them, the overall stability data of system operation can be obtained by analyzing and integrating the operating status of multiple devices and system components at the digital twin system level to derive the overall stability and performance status of the system.

[0071] Among them, environmental impact simulation data can be data generated by simulating the impact of external environmental factors (such as temperature, humidity, pressure, etc.) on equipment and systems through the digital twin environment layer.

[0072] Specifically, the equipment record information, equipment detection information and equipment environment information will be input into three different levels of digital twin equipment layer, digital twin system layer and digital twin environment layer respectively. At the digital twin equipment layer, the real-time operating status of the equipment will be processed and simulated to generate real-time operating status data of the equipment including equipment health, performance, workload and other data; at the digital twin system layer, the system will conduct an overall performance evaluation, taking into account the interaction of various equipment and components, and generate overall stability data of the system operation to reflect the operating stability of the entire system; and at the digital twin environment layer, the system will predict the potential impact of the external environment on the equipment by simulating environmental changes, generate environmental impact simulation data, and help understand the interference or assistance to equipment operation under different environmental conditions.

[0073] Step 404: Apply the real-time operating status data of the equipment to the digital twin system layer and the digital twin environment layer to obtain equipment system interaction data and equipment environment status data.

[0074] Among them, equipment system interaction data can be the data generated by the interaction between the real-time operating status data of the equipment and other system components. These data reflect how the equipment affects the overall performance and stability of the system in the process of working in collaboration with other equipment or subsystems.

[0075] The equipment environment status data may be data generated by combining the real-time operation status data of the equipment with the external environment data, and reflects how the external environment affects the operation of the equipment.

[0076] Specifically, the real-time operating status data of the equipment are input into the digital twin system layer and the digital twin environment layer respectively. At the digital twin system layer, by combining these real-time data with the operating status of other components and equipment in the system, the system can analyze the interaction between the equipment and other system components, identify potential influencing factors, and generate equipment system interaction data, which includes how the equipment affects the stability and efficiency of the system, whether there are problems such as load overload and equipment coordination; at the same time, the real-time operating data is also applied to the digital twin environment layer to obtain equipment environmental status data. These data reflect the working performance and health status of the equipment in the actual environment, and can explain the direct impact of the external environment (such as temperature, humidity, climate and other factors) on equipment performance and efficiency, thereby providing data support for the adaptability evaluation of the equipment.

[0077] Step 406: Apply the overall stability data of the system operation to the digital twin equipment layer and the digital twin environment layer to obtain the system equipment command data and the system environment status data.

[0078] Among them, the system equipment command data can be the overall stability data of the system operation and the data about the system operation command generated in the digital twin system layer. These data guide how the equipment should operate and adjust according to the overall needs and goals of the system.

[0079] The system environment status data may be a combination of the overall stability data of the system operation and external environmental conditions, and may be generated to reflect the operating status of the system under different environmental influences.

[0080] Specifically, the overall stability data of the system operation are input into the digital twin equipment layer and the digital twin environment layer respectively. At the digital twin equipment layer, these system-layer stability data are applied to the analysis of specific equipment to generate system equipment command data, which reflects how the overall operating status of the system affects the operation of each individual device, helps to judge the response of the equipment to system load fluctuations, and whether the equipment needs to be adjusted to keep the system running smoothly; at the digital twin environment layer, the stability of the system is combined with external environmental factors to generate system environmental status data, which reflects the mutual influence between external environmental factors and the overall operation of the system. For example, in extreme weather or high load conditions, whether the equipment can continue to operate stably under the existing environmental conditions, or whether the system will be affected by environmental changes and cause reduced efficiency or failures.

[0081] Step 408: Apply the environmental impact simulation data to the digital twin equipment layer and the digital twin system layer to obtain environmental equipment impact data and environmental system impact data.

[0082] Among them, the environmental equipment impact data can be data generated by combining environmental impact simulation data with equipment operating status, reflecting the impact of the external environment on the operation of a single device.

[0083] Among them, environmental system impact data can be data generated by the impact of environmental impact simulation data on the operation of the entire system. These data reveal the impact of environmental changes (such as extreme climate, temperature and humidity changes, etc.) on the overall operation of the system, and help evaluate the potential risks of environmental factors to system stability, efficiency and life.

[0084] Specifically, the environmental impact simulation data is input into the digital twin equipment layer and the digital twin system layer. In the digital twin equipment layer, these environmental data help simulate the performance of the equipment under specific environmental conditions and generate environmental equipment impact data. These data describe how environmental conditions (such as temperature, humidity, pollution, etc.) affect the operating efficiency, health status and failure risk of the equipment. For example, extreme temperatures may cause the cooling system to fail or the performance of electronic components to deteriorate; in the digital twin system layer, environmental data will be used to analyze the impact of the environment on the entire system and generate environmental system impact data. These data reflect the impact of environmental factors on the overall load and performance of the system. For example, high humidity may cause electrical equipment failures, or high temperature environments may increase the burden on the cooling system and affect the stability of the entire system.

[0085] Step 410, the real-time operation status data of the equipment, the overall stability data of the system operation, the environmental impact simulation data, the equipment system interaction data, the equipment environment status data, the system equipment command data, the system environment status data, the environmental equipment impact data and the environmental system impact data are integrated at the decision level to obtain the physical twin state information of the equipment.

[0086] Among them, decision-level fusion can be a comprehensive analysis of multi-dimensional data from different levels (equipment, system, environment), eliminating redundancy and conflict, and obtaining a decision data set through weighted processing through algorithms and models.

[0087] Specifically, the system will conduct decision-level fusion of all collected data. All data sources, including equipment real-time operation status data, system operation overall stability data, environmental impact simulation data, equipment system interaction data, equipment environment status data, system equipment command data, system environment status data, environmental equipment impact data, and environmental system impact data will be integrated and analyzed through advanced data fusion algorithms. Specifically, data from different levels and dimensions (such as equipment real-time operation status, system stability, environmental impact, equipment adjustment data, etc.) will be comprehensively processed and weighted analyzed, and various types of data will be integrated with algorithms to eliminate redundant information, and the mutual influence between data will be modeled to generate equipment physical twin state information.

[0088] In this embodiment, by inputting equipment recording information, detection information and environmental information into each level of the digital twin (equipment level, system level, and environmental level), and performing data interaction and comprehensive analysis at these levels, comprehensive monitoring and simulation of equipment operation status, system stability and environmental impact can be achieved. This multi-level and multi-dimensional information fusion not only enhances the control of the real-time operation status of the equipment, but also reveals the comprehensive impact of different factors (such as environmental changes, system load, cooling adjustment, etc.) on equipment performance. Through the decision-level fusion of various types of data, the system can generate accurate physical twin state information, providing a more accurate basis for equipment maintenance and optimization, thereby improving the equipment's operating efficiency, prediction capabilities, and fault warning capabilities, and ultimately ensuring efficient and safe operation of the equipment.

[0089] In an exemplary embodiment, Figure 5 As shown, according to the failure information of each maintenance equipment, the maintenance equipment information of the target equipment is isolated and predicted to obtain the prediction information of each maintenance equipment, including steps 502 to 508. Among them: Step 502: extract the features of each maintenance equipment fault information to obtain the mark feature data of each maintenance equipment.

[0090] The maintenance equipment identification feature data may be key data points extracted from the failure information of the maintenance equipment, and used to describe and distinguish possible failure features of different maintenance equipment.

[0091] Specifically, feature extraction is performed on the fault information of each maintenance equipment, mainly analyzing important data such as possible fault mode, fault frequency, fault type, etc. The feature extraction process includes preprocessing the fault information, removing irrelevant data, and identifying important features related to the fault through algorithms (such as statistical analysis, signal processing or machine learning).

[0092] Step 504 , performing packet processing on each maintenance equipment failure information and the maintenance equipment mark characteristic data corresponding to each maintenance equipment failure information to obtain each packet maintenance equipment failure information.

[0093] Among them, packet processing can be a process of packaging different types of data (such as fault information and characteristic data) into an overall data unit.

[0094] The packet maintenance equipment failure information may be a data unit after packet processing, which includes the failure information of the maintenance equipment and related mark characteristic data.

[0095] Specifically, the fault information of each maintenance equipment and its corresponding maintenance equipment mark characteristic data are matched, and after the matching is completed, each group of data is packaged separately, that is, the fault information of each maintenance equipment and its mark characteristic data are packaged together to form "packaged maintenance equipment fault information", ensuring that each fault information and its characteristic data can be closely associated and not interfered with by other data. The packaged maintenance equipment fault information will be displayed and displayed with the corresponding maintenance equipment mark characteristic data.

[0096] Step 506 : Selecting a target maintenance equipment prediction model corresponding to each maintenance equipment mark feature data from the preset maintenance equipment prediction model set.

[0097] The preset maintenance equipment prediction model set may be a set consisting of a plurality of trained maintenance equipment prediction models, and is used to process different types of maintenance equipment fault prediction tasks.

[0098] The target maintenance equipment prediction model may be a final model selected through feature matching in a set of preset maintenance equipment prediction models.

[0099] Specifically, since the preset maintenance equipment prediction model set includes multiple maintenance equipment prediction models of different types, these models are trained based on different fault characteristics and data processing methods. By matching the maintenance equipment mark feature data corresponding to each package of maintenance equipment fault information and the maintenance equipment prediction model, the system can accurately select the maintenance equipment prediction model most relevant to each package of maintenance equipment fault information as the most matched target maintenance equipment prediction model, that is, according to the pairing of the maintenance equipment mark feature data and the maintenance equipment prediction model, the maintenance equipment prediction model corresponding to each package of maintenance equipment fault information is determined.

[0100] Step 508, using each maintenance equipment mark feature data as guide information, input each package maintenance equipment fault information into the corresponding target maintenance equipment prediction model to obtain each maintenance equipment prediction information.

[0101] Specifically, each maintenance equipment mark feature data is used as the guide information, that is, the pairing information between the maintenance equipment mark feature data and the target maintenance equipment prediction model is used to input each package maintenance equipment fault information into the corresponding target maintenance equipment prediction model. Each target maintenance equipment prediction model uses the corresponding package maintenance equipment fault information to perform maintenance equipment demand prediction analysis. Each target maintenance equipment prediction model will calculate the input data according to the trained algorithm and output the corresponding maintenance equipment prediction information. These prediction information usually include the probability of failure, possible failure type, maintenance suggestions, etc., to help maintenance personnel make effective decisions in practical applications.

[0102] In this embodiment, by extracting features and packaging the maintenance equipment fault information, the system is helped to extract key features from complex fault data. Packaging ensures the structuring and consistency of the information, which is convenient for subsequent processing and analysis. By selecting the most appropriate target model from the preset maintenance equipment prediction model set and using the extracted feature data as a guide input, the system can accurately predict the fault information of various types of maintenance equipment. This method not only optimizes the prediction process, but also selects the most appropriate prediction model according to the characteristics of different maintenance equipment, improves the efficiency of fault diagnosis, ensures the accuracy of maintenance decisions, and ultimately improves the maintenance efficiency and reliability of equipment.

[0103] In an exemplary embodiment, Figure 6 As shown, according to the application scenario parameters and application efficiency parameters of the target equipment, the optimal information is selected from the maintenance equipment prediction information to obtain the optimal maintenance equipment information, including steps 602 to 606. Among them: Step 602: Determine the equipment information optimization target of the target equipment according to the application scenario parameters and application efficiency parameters of the target equipment.

[0104] Among them, the equipment information optimization goal can be the optimization goal determined based on the application scenario and application efficiency parameters of the equipment during the equipment maintenance process. These goals reflect how to choose the most suitable maintenance equipment under specific environments and needs.

[0105] Specifically, the system needs to comprehensively consider the application scenario parameters of the target equipment (such as equipment working conditions, ambient temperature and humidity, load conditions, etc.) and the application efficiency parameters in the current scenario (such as maintenance time, repair completeness, resource consumption, etc.) to determine the optimization target of the maintenance equipment. For example, if the equipment works in a high temperature environment, it may be necessary to give priority to high temperature resistant maintenance equipment; if the equipment needs to be restored to operation quickly, the repair speed will become the optimization target. Therefore, the system will establish the optimization target by analyzing the current working scenario and the required maintenance efficiency. It usually uses an optimization algorithm to weight and integrate multiple targets to form the equipment information optimization target of the target equipment.

[0106] Step 604, according to the equipment information optimization target, using each equipment information optimization algorithm, selects the best information from each maintenance equipment prediction information, and obtains each equipment information optimization data.

[0107] Among them, the equipment information optimization algorithm can be a mathematical method and calculation process for processing and analyzing maintenance equipment prediction information. It selects the most suitable maintenance equipment by simulating the impact of different options on the maintenance effect based on the set equipment information optimization goal.

[0108] The equipment information optimization data can be a set of maintenance equipment information data obtained after being processed by an optimization algorithm. Each piece of data represents the characteristics, fault prediction results, and maintenance adaptability of a maintenance equipment, and is evaluated by an optimization algorithm to meet the requirements of a specific optimization goal.

[0109] Specifically, according to the previously set equipment information optimization goal, various information optimization algorithms are used to process the prediction information of maintenance equipment. These algorithms can include genetic algorithms, particle swarm optimization algorithms, simulated annealing algorithms, etc. The purpose is to optimize the optimal solution that meets the needs from part or all of the maintenance equipment of many possible maintenance equipment prediction information options by simulating the impact of different options on the maintenance effect. Each optimization algorithm will be evaluated based on multiple dimensions such as maintenance efficiency, cost, equipment adaptability, etc., and ultimately each optimization algorithm will output the corresponding equipment information optimization data.

[0110] Step 606, selecting a data set with the smallest difference from the equipment information optimization target from among various equipment information optimization data, and obtaining the optimal maintenance equipment information.

[0111] Specifically, the system will select the maintenance equipment information with the smallest difference from the optimization target from the optimization data of each equipment information based on the target optimization standard. Specifically, the system will calculate the difference between each equipment information optimization data and the target optimization target (such as maintenance time, cost, equipment matching, etc.), and determine the optimal maintenance equipment information by screening those data with the smallest difference from the target. The key in the calculation process is to calculate the "fitness" of each maintenance equipment, that is, the matching degree between it and the preset optimization target, and finally select the optimal maintenance equipment solution as the optimal maintenance equipment information to ensure the optimal allocation of resources and maximize efficiency during the maintenance process.

[0112] In this embodiment, by combining the application scenario parameters and application efficiency parameters of the equipment, the system can accurately determine the optimization target of the maintenance equipment information, so as to select the most suitable maintenance equipment according to different working environments and efficiency requirements. Using various equipment information optimization algorithms, the system can filter out the optimal data that meets the needs from a large amount of maintenance equipment prediction information, ensuring that the selected equipment can meet the performance requirements in specific application scenarios to the greatest extent. By further selecting the data set with the smallest difference from the optimization target from multiple optimization results, the system can provide the most suitable maintenance equipment information. The accuracy and efficiency of maintenance equipment selection are significantly improved, ensuring that maintenance work can be carried out under optimal conditions, reducing unnecessary waste of resources, and improving the reliability and service life of equipment.

[0113] In an exemplary embodiment, Figure 7As shown, the data set with the smallest difference from the equipment information optimization target is selected from each equipment information optimization data to obtain the optimal maintenance equipment information, including steps 702 to 708. Among them: Step 702, taking the equipment information optimization target as a constraint condition, using various Pareto optimization algorithms, respectively searching for the Pareto frontier in each equipment information optimization data, and obtaining the Pareto frontier data of each equipment.

[0114] Among them, the Pareto optimization algorithm can be an algorithm for multi-objective optimization problems. Its core idea is to find a set of "Pareto optimal solutions", that is, solutions in which no single objective can be further optimized without sacrificing other objectives among multiple optimization objectives. Pareto optimization algorithms include non-dominated sorting genetic algorithm (NSGA-II), multi-objective differential evolution algorithm (MODE), multi-objective ant colony optimization algorithm (MOACO), and multi-objective gradient descent method (MOGD).

[0115] Among them, the equipment Pareto frontier data can be a set of optimal solutions obtained by the Pareto optimization algorithm, which finds the best balance between multiple optimization objectives (such as maintenance time, cost, repair quality, etc.).

[0116] Specifically, all optimization sub-goals of the equipment information optimization goal are still used as the constraints (goal conditions) of Pareto optimization, and different Pareto optimization algorithms are used to screen out the optimal solution set from the equipment information optimization data. Among them, different Pareto optimization algorithms analyze the trade-off relationship between all optimization sub-goals of the equipment information optimization goal and identify the "Pareto frontier" - that is, the data set with no better solution between multiple goals. In this process, the system will perform multi-objective optimization on each piece of equipment information optimization data, find out those data that can no longer be optimized on certain goals without affecting other goals, and combine the application of the Pareto optimal principle to obtain a set of optimal solutions that best meet all optimization sub-goals of the equipment information optimization goal, called "equipment Pareto frontier data."

[0117] Step 704: adjust the Pareto frontier data of each device according to the application scenario parameter and the application efficiency parameter to obtain each adjusted Pareto frontier data.

[0118] The adjusted Pareto frontier data may be obtained by further optimizing and adjusting the frontier data according to application scenario parameters and application efficiency parameters after obtaining preliminary Pareto frontier data.

[0119] Specifically, the previously obtained equipment Pareto frontier data is adjusted and optimized according to the application scenario parameters and application efficiency parameters. The application scenario parameters usually include the specific working environment of the equipment, such as temperature, humidity, pressure, load, etc. These factors determine the working characteristics of the equipment in different environments. For example, some maintenance equipment may perform better when working in a high temperature environment, while some equipment is more adaptable in low temperature or extreme humidity. The application efficiency parameters involve efficiency indicators in the maintenance process, such as maintenance speed, maintenance time limit, repair quality requirements, etc., which determine the optimal maintenance equipment to be selected under specific conditions.

[0120] The system will make weighted adjustments to each Pareto frontier data based on these factors to ensure that the weight of each data in the objective function reflects the actual working environment and efficiency requirements. The adjustment process includes re-evaluating the adaptability of maintenance equipment according to the operating conditions of the equipment (such as operating load, environmental changes, etc.), and dynamically adjusting the performance of maintenance equipment in actual operation. For example, under high load conditions, the system may tend to choose equipment with higher load capacity or stronger durability, and vice versa, it may tend to choose light and efficient equipment. The adjusted data is called adjusted Pareto frontier data.

[0121] Step 706 , using the frontier data simulated annealing algorithm, locally optimize each adjusted Pareto frontier data to obtain each locally optimized Pareto frontier data.

[0122] The locally optimized Pareto front data may be data obtained by refining the adjusted Pareto front data through further optimization means (such as a simulated annealing algorithm, etc.).

[0123] Specifically, the system will use the frontier data simulated annealing algorithm to perform local optimization on each adjusted Pareto frontier data. The frontier data simulated annealing algorithm is an optimization algorithm used to find an approximate global optimal solution, inspired by the annealing process in physics. In this process, the system will simulate a gradual decrease in the "temperature" parameter, so that the system gradually tends to a low temperature state from the initial higher temperature state, thereby avoiding falling into a local optimal solution.

[0124] Specifically, the frontier data simulated annealing algorithm introduces randomness, which allows the frontier data simulated annealing algorithm to jump out of the local optimum during the search process and explore a wider solution space. At each "temperature" state, the frontier data simulated annealing algorithm will evaluate the quality of the current solution and accept a new solution based on a certain probability, even if this solution may be worse than the current solution in some aspects. As the temperature decreases, the probability of accepting a poor solution gradually decreases, and the system will eventually converge to a better solution. In the optimization of each adjusted Pareto frontier data, the frontier data simulated annealing algorithm randomly perturbs each adjusted Pareto frontier data and evaluates the quality of the results after the perturbation, gradually refining various indicators to find the best maintenance equipment selection plan. Each data set after the frontier data simulated annealing algorithm corresponds to each local optimized Pareto frontier data.

[0125] Step 708: cluster the local optimized Pareto front data to obtain the optimal maintenance equipment information.

[0126] Specifically, the system will further perform clustering processing on each local optimized Pareto front data, where the purpose of clustering processing is to divide the maintenance equipment information with similar characteristics in the optimized data set into different groups, so as to select the most representative maintenance equipment from them. The purpose of clustering is to reduce redundant information, simplify the selection range, and ensure that the most suitable maintenance equipment is selected from multiple optimization solutions. The clustering algorithms used in clustering processing of each local optimized Pareto front data include K-means, DBSCAN, etc. These algorithms classify data points into the same class based on the similarity between the data. For example, when using K-means clustering, the system will select a suitable number of clusters K, and then adjust the center point of each cluster in an iterative manner, and finally assign the data to the cluster with the closest center point. Each cluster represents maintenance equipment information with similar characteristics, and the clustering results help to effectively summarize the multi-dimensional attributes of different equipment information. Clustering each local optimized Pareto front data is to screen out the maintenance equipment information that is most prominent in all optimized data and meets the requirements of the application scenario. The clustered data set will screen out the optimal maintenance equipment based on standards such as minimizing distance, maximizing resource utilization, and optimizing repair effects, ensuring that the equipment can receive the best maintenance support, and the clustered information will be used as the optimal maintenance equipment information.

[0127] In this embodiment, by taking the equipment information optimization target as a constraint condition and using the Pareto optimization algorithm to find the Pareto frontier in each equipment information data, the optimal balance point between multiple targets can be effectively identified to provide the optimal solution. Further adjusting the Pareto frontier data according to the application scenario and application efficiency parameters can optimize the solution according to actual needs to ensure that the selected maintenance equipment achieves the best performance under different usage environments. The adjusted frontier data is locally optimized using a simulated annealing algorithm to further improve the accuracy of the solution and avoid the trap of local optimal solutions. Finally, by clustering the locally optimized frontier data, a set of optimal maintenance equipment information with high consistency can be obtained, thereby providing an accurate and effective maintenance equipment selection solution. This series of steps ensures that the selection of maintenance equipment not only meets actual needs, but also maximizes efficiency and reduces costs, ultimately improving the reliability and maintenance effectiveness of the equipment.

[0128] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows; moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but may be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least a portion of the steps or stages in other steps.

[0129] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0130] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0131] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps in the above-mentioned method embodiments.

[0132] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0133] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A maintenance equipment information prediction method based on a digital twin model, characterized in that: The method comprises: Obtain equipment record information, equipment detection information, and equipment environment information of the target equipment; Determine each digital twin description layer of the target equipment according to the equipment record information, the equipment detection information and the equipment environment information; According to the equipment record information, the equipment detection information and the equipment environment information, interactive description is performed at each digital twin description layer to obtain equipment physical twin state information; Using the equipment maintenance physical prediction model and the equipment maintenance machine prediction model respectively, according to the equipment physical twin state information, the fault information of the target equipment is predicted to obtain the fault information of each maintenance equipment; According to each of the maintenance equipment failure information, information isolation prediction is performed on the maintenance equipment information of the target equipment to obtain each maintenance equipment prediction information; According to the application scenario parameters and application efficiency parameters of the target equipment, optimal information is selected from each of the maintenance equipment prediction information to obtain optimal maintenance equipment information.

2. The method according to claim 1, characterized in that: Using each equipment maintenance physical prediction model and equipment maintenance machine prediction model respectively, according to the equipment physical twin state information, the fault information of the target equipment is predicted to obtain the fault information of each maintenance equipment, including: Using each of the equipment maintenance physical prediction models respectively, according to the equipment physical twin state information, the maintenance equipment information of the target equipment is predicted to obtain the equipment fault information of each physical model; Using the equipment maintenance machine prediction model of each equipment respectively, according to the physical twin state information of the equipment, the maintenance equipment information of the target equipment is predicted to obtain the equipment fault information of each machine model; The equipment accessories feature information of the target equipment is used as the guide information to perform consistency identification on each of the physical model equipment fault information and each of the machine model equipment fault information to obtain each of the maintenance equipment fault information.

3. The method according to claim 2, characterized in that The equipment maintenance physical prediction model includes an inherent fault calculation model, and the calculation formula of the inherent fault calculation model is: Where N(t) is the remaining life at time t, A is the material fatigue characteristic constant, E is the material elastic modulus, b and c are fitting coefficients, and F i For environmental factors, is the sensitivity coefficient of the environmental factor, m is the number of environmental factors, D(t) is the time Cumulative fatigue damage, D max is the maximum damage resistance of the material, is the microstructure damage influence coefficient, For time The stress, is the reference stress of the material, k is the fatigue strength index, is the environmental impact index, is the material microstructure change function, is the inherent failure rate at time t, is the life span change rate.

4. The method according to claim 2, characterized in that: The equipment maintenance physical prediction model includes a fault calculation model, and the calculation formula of the fault calculation model is: in, is the failure rate at time t, is the benchmark failure rate, P(t) is the actual operating load at time t, and P max is the maximum design load of the equipment, is the load sensitivity coefficient, H(t) is the time The operational health index, is the sensitivity coefficient of the health index to the failure rate, is the failure rate after the combined operation behavior at time t, is the j-th operation behavior characteristic at time t, is the influence coefficient of the jth operation on the failure rate, n is the number of operation behavior characteristics, is the usage failure rate of the combined operating load, operating behavior and health status at time t.

5. The method according to claim 2, characterized in that: The equipment maintenance physical prediction model includes an environmental stress failure calculation model, and the calculation formula of the environmental stress failure calculation model is: in, is the failure rate under the influence of environmental stress at time t, is the basic failure rate when the environment has no impact, E k (t) is the kth environmental stress factor at time t, E max,k is the maximum design value of the kth environmental stress factor, is the sensitivity coefficient of the kth environmental stress factor, p is the number of environmental stress factors, D 环境 (t) is the accumulated damage of environmental stress at time t, E k (t') is the kth environmental stress factor at time t, is the damage attenuation coefficient of the kth environmental stress factor, is the failure rate after time t combined with environmental stress, is the cumulative impact coefficient of environmental stress on failure rate.

6. The method according to claim 1, characterized in that The digital twin description layer includes a digital twin equipment layer, a digital twin system layer and a digital twin environment layer; the interactive description is performed in each digital twin description layer according to the equipment record information, the equipment detection information and the equipment environment information to obtain the equipment physical twin state information, including: Input the equipment record information, the equipment detection information and the equipment environment information into the digital twin equipment layer, the digital twin system layer and the digital twin environment layer respectively, to obtain equipment real-time operation status data, system operation overall stability data and environmental impact simulation data; Applying the real-time operation status data of the equipment to the digital twin system layer and the digital twin environment layer to obtain equipment system interaction data and equipment environment status data; Applying the overall stability data of the system operation to the digital twin equipment layer and the digital twin environment layer to obtain system equipment command data and system environment status data; Applying the environmental impact simulation data to the digital twin equipment layer and the digital twin system layer to obtain environmental equipment impact data and environmental system impact data; The real-time operation status data of the equipment, the overall stability data of the system operation, the environmental impact simulation data, the equipment system interaction data, the equipment environment status data, the system equipment command data, the system environment status data, the environment equipment impact data and the environment system impact data are fused at the decision level to obtain the physical twin state information of the equipment.

7. The method according to claim 1, characterized in that The method of performing information isolation prediction on the maintenance equipment information of the target equipment according to the maintenance equipment failure information to obtain maintenance equipment prediction information includes: Extracting features of the fault information of each maintenance equipment respectively to obtain feature data of each maintenance equipment mark; Packaging the maintenance equipment failure information and the maintenance equipment mark characteristic data corresponding to the maintenance equipment failure information to obtain each package of maintenance equipment failure information; each package of maintenance equipment failure information displays the corresponding maintenance equipment mark characteristic data; Selecting a target maintenance equipment prediction model corresponding to each maintenance equipment mark feature data from a preset maintenance equipment prediction model set; Using the characteristic data of each maintenance equipment mark as guide information, each package maintenance equipment fault information is input into the corresponding target maintenance equipment prediction model to obtain each maintenance equipment prediction information.

8. The method according to any one of claims 1 to 7, characterized in that: The step of selecting optimal information from each maintenance equipment prediction information according to the application scenario parameter and the application efficiency parameter of the target equipment to obtain optimal maintenance equipment information includes: Determining an equipment information optimization target for the object equipment according to an application scenario parameter and an application efficiency parameter of the object equipment; According to the equipment information optimization target, using each equipment information optimization algorithm, selecting the best information from each maintenance equipment prediction information, and obtaining each equipment information optimization data; A data set with the smallest difference from the equipment information optimization target is selected from each of the equipment information optimization data to obtain the optimal maintenance equipment information.

9. The method according to claim 8, characterized in that The step of selecting a data set having the smallest difference from the equipment information optimization target from each of the equipment information optimization data to obtain the optimal maintenance equipment information comprises: Taking the equipment information optimization target as a constraint condition, using various Pareto optimization algorithms, respectively searching for the Pareto frontier in each of the equipment information optimization data, and obtaining the Pareto frontier data of each equipment; According to the application scenario parameters and the application efficiency parameters, the Pareto frontier data of each of the equipment is adjusted to obtain each adjusted Pareto frontier data; Using a frontier data simulated annealing algorithm, locally optimizing each of the adjusted Pareto frontier data, respectively, to obtain each locally optimized Pareto frontier data; Clustering is performed on each of the local optimized Pareto front data to obtain the optimal maintenance equipment information.

10. A maintenance equipment information prediction device based on a digital twin model, characterized in that: The device comprises: An equipment data acquisition module is used to acquire equipment record information, equipment detection information and equipment environment information of the target equipment; A twin model determination module, used to determine each digital twin description layer of the target equipment according to the equipment record information, the equipment detection information and the equipment environment information; A twin model description module, used to interactively describe the equipment at each digital twin description layer according to the equipment record information, the equipment detection information and the equipment environment information, so as to obtain the equipment physical twin state information; A fault information prediction module is used to use the equipment maintenance physical prediction model and the equipment maintenance machine prediction model to predict the fault information of the target equipment according to the equipment physical twin state information, and obtain the fault information of each maintenance equipment; A maintenance equipment prediction module, used to perform information isolation prediction on the maintenance equipment information of the target equipment according to the maintenance equipment failure information, and obtain the maintenance equipment prediction information; The maintenance equipment determination module is used to select the best information from each maintenance equipment prediction information according to the application scenario parameters and application efficiency parameters of the target equipment to obtain the best maintenance equipment information.

Images