Predictive Maintenance System for Ultra-High Efficiency Explosion-Proof Motors Based on Digital Twin

Abstract

This invention relates to the field of industrial production and discloses a predictive maintenance system for ultra-high-efficiency explosion-proof motors based on digital twins. The system includes: determining whether the activation threshold of the digital twin model has been reached when key operating parameters of the motor deviate from the normal operating range; constructing a virtual mapping of the motor's operation based on historical operating data, real-time sensor data, and environmental factors, and generating a current state evolution trajectory to predict whether a fault exists in the operating trend; assessing the existence of potential fault paths by matching the state trajectory change pattern with a historical fault database; determining whether early intervention measures are needed based on the fault risk level and expected occurrence time, combined with the current production plan and equipment health tolerance; and generating an intervention plan based on the state control parameters fed back by the digital twin model and the specific operating conditions. This invention has the advantages of improving operational safety and maintenance efficiency.

Description

Technical Field

[0001] This invention relates to the field of industrial production, specifically to a highly efficient predictive maintenance system for explosion-proof motors based on digital twins. Background Technology

[0002] In industrial production, explosion-proof motors, as critical power equipment, are widely used in flammable and explosive environments such as chemical, coal mine, and petroleum industries. Their operating status directly affects the safety and stability of the production system. Currently, many explosion-proof motor maintenance systems still use alarm mechanisms based on fixed thresholds. This involves setting warning thresholds for parameters such as temperature, current, and vibration, and triggering an alarm once the actual monitored data exceeds the set range. While this method improves the efficiency of detecting motor anomalies to some extent, it still has significant shortcomings. Threshold strategies are usually based on experience or general standards, failing to fully consider the individual characteristics of motors under different operating environments and conditions, leading to frequent false alarms and missed alarms. This approach is essentially a "post-event response" mechanism, only issuing alarms after the equipment status has significantly deviated from the normal range, lacking the ability to identify and intervene in the early stages of faults. Therefore, it is essential to design an ultra-efficient predictive maintenance system for explosion-proof motors based on digital twins, capable of early detection and precise intervention of potential faults. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides an ultra-efficient predictive maintenance system for explosion-proof motors based on digital twins, which has the advantages of improving operational safety and maintenance efficiency, and solves the problems mentioned in the background technology.

[0004] To achieve the aforementioned goals of improving operational safety and maintenance efficiency, this invention provides the following technical solution: a highly efficient predictive maintenance system for explosion-proof motors based on digital twins, comprising:

[0005] Operating status identification module: When the key operating parameters of the motor deviate from the normal operating range, it determines whether the activation threshold of the digital twin model has been reached. If it has, it enters the status modeling module.

[0006] State modeling module: Based on historical operating data, real-time sensor data and environmental factors, it calls the digital twin model to build a virtual mapping of motor operation, generates the current state evolution trajectory, predicts whether there is a fault in the operating trend, and if so, enters the fault risk assessment module.

[0007] Fault Risk Assessment Module: By matching the state trajectory change pattern with the historical fault database, it assesses whether there are potential fault paths. If so, it proceeds to the intervention needs analysis module.

[0008] Intervention Needs Analysis Module: Based on the failure risk level and expected occurrence time, combined with the current production plan and equipment health tolerance, determine whether it is necessary to implement early intervention measures. If so, proceed to the intervention strategy generation module.

[0009] Intervention strategy generation module: Based on the state control parameters fed back from the digital twin model and combined with the specific working conditions, an intervention plan is generated.

[0010] Preferably, the process for determining whether the activation threshold of the digital twin model has been reached is as follows:

[0011] Set key operating parameters and calculate the degree of deviation of the current parameters from the normal operating range;

[0012] If the deviation is greater than 0, it means that the motor operating parameters have exceeded the warning value, and it is necessary to assess whether it is close to the activation threshold.

[0013] Preferably, assessing whether the activation threshold is approached includes:

[0014] Calculate the difference between the current parameter and the activation threshold to obtain the activation proximity value;

[0015] If the activation proximity value is less than the activation proximity threshold, then state prediction of the current parameter is required.

[0016] If the activation proximity value is greater than or equal to the activation proximity threshold, then there is no need to predict the state of the current parameter.

[0017] Preferably, the process of generating the current state evolution trajectory is as follows:

[0018] Construct a multi-dimensional time-series curve of the motor state based on the current key operating parameters of the motor, with the X-axis representing time and the Y-axis representing each operating parameter;

[0019] Based on historical operating data, a state evolution trajectory curve of the motor under standard conditions is constructed as a reference curve;

[0020] On the current state timing curve of the motor and the reference state trajectory curve, select several sampling points with the same time point;

[0021] Obtain the operating parameter values ​​of the motor's current state curve at different sampling points and integrate them into the first state data group;

[0022] Obtain the values ​​of each running parameter at the same sampling point of the reference state trajectory curve, and integrate them into a second state data group;

[0023] Calculate the cosine similarity between the first state data set and the second state data set to obtain the trajectory similarity value.

[0024] Preferably, a curve similarity threshold is set:

[0025] If the trajectory similarity value is less than the curve similarity threshold, it indicates that the curve type is linear;

[0026] If the trajectory similarity value is greater than or equal to the curve similarity threshold, it indicates that the curve type is non-linear.

[0027] Preferably, the process of predicting whether there is a fault in the operating trend includes:

[0028] If the curve type is linear, the least squares method is used to fit the trend curve of the key operating parameters of the motor to obtain the functional equation of the fitted straight line.

[0029] By using the critical threshold of the operating parameters as input to the function equation, the time point at which the operating parameters reach the fault threshold, i.e. the critical time point, is predicted.

[0030] Preferably, the critical time point is compared with the end time point of the current maintenance cycle:

[0031] If the critical time point is less than the end time point of the maintenance cycle, it means that the motor is not in a safe operating state during the current cycle.

[0032] If the critical time point is greater than or equal to the end time point of the maintenance cycle, it indicates that the motor is operating safely within the current cycle.

[0033] Preferably, the process of predicting whether there is a fault in the operating trend also includes:

[0034] If the curve type is nonlinear, perform growth analysis on the first state data group to determine the fastest growth rate of the key operating parameters. The first state data group is the historical change data of the key parameters in the time dimension. By performing difference operations, calculate the change rate of the parameters at each moment and determine the maximum growth rate.

[0035] Calculate the difference between the current value of the key parameter and the fault threshold to obtain the parameter approximation value. Divide the parameter approximation value by the fastest growth rate to obtain the time it takes for the parameter to reach the fault threshold, i.e., the critical time. Add the critical time to the current time point to obtain the critical time point.

[0036] Preferably, the process for determining whether early intervention measures are needed is as follows:

[0037] Obtain current production plan information and the equipment health tolerance of the motor;

[0038] Combine equipment health status data with production planning data to construct equipment health data groups and production task data groups;

[0039] The correlation between the health data set and the production plan data set was calculated using the Pearson correlation coefficient.

[0040] Preferably, the process of generating an intervention plan is as follows:

[0041] Collect operating condition information related to equipment operation, search and match the intervention strategy library, filter out intervention measures that are suitable for the current scenario, and form an intervention plan based on the selected strategy, combined with resource scheduling and production tasks.

[0042] Compared with existing technologies, this invention provides an ultra-efficient predictive maintenance system for explosion-proof motors based on digital twins, which has the following advantages:

[0043] This invention achieves intelligent predictive maintenance across the entire process, from anomaly identification to intervention decision-making, through multi-module collaborative operation. It activates a digital twin model as soon as critical motor parameters show abnormalities, pre-constructing a virtual mapping of the operating state and predicting evolution trends. This effectively identifies potential fault risks, achieves precise fault risk assessment through trajectory pattern matching, and scientifically decides whether to intervene based on production rhythm and equipment tolerance. Furthermore, it dynamically generates the optimal intervention plan according to operating conditions. This not only improves the accuracy and efficiency of fault prediction but also significantly reduces unplanned downtime and maintenance costs, ensuring the safety and continuity of motor operation. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the structure of the present invention;

[0045] Figure 2 This is a schematic diagram of the method of the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Example 1

[0048] Please see Figure 1 As shown in the embodiment of the present invention, the ultra-efficient predictive maintenance system for explosion-proof motors based on digital twins includes:

[0049] Operational status identification module: When the key operating parameters of the motor deviate from the normal operating range, it determines whether the activation threshold of the digital twin model has been reached. If it has, it enters the status modeling module.

[0050] The process by which the operation status identification module determines whether the activation threshold of the digital twin model has been reached when key motor operating parameters deviate from the normal operating range is as follows:

[0051] Set key operating parameters, including real-time parameter value P. real Operating condition warning value P alert With the activation threshold P of the digital twin model max The formula for calculating the deviation of the current parameter from the normal operating range is as follows:

[0052] ΔP alert =P real -P alert

[0053] It should be noted that the operating condition warning value P alert <Activation threshold P of digital twin model max Operating condition warning value P alert The parameters are set by those skilled in the art based on the motor structure, usage scenario, historical data, etc., with the aim of preventing real-time parameter values ​​P from being set. aaal Exceeding the activation threshold P of the digital twin model max The set warning value;

[0054] If ΔP aleat >0 indicates that the motor operating parameters have exceeded the warning value and need to be further evaluated to see if they are close to the activation threshold;

[0055] Calculate the difference between the current parameter and the activation threshold to obtain the activation proximity value ΔP. margin :

[0056] ΔP aaaaaa =P max -P real

[0057] If the activation proximity value is less than the activation proximity threshold P thresh Then it is necessary to predict the state of the current parameters;

[0058] If the activation proximity value is greater than or equal to the activation proximity threshold P thresh Then there is no need to predict the state of the current parameters;

[0059] It should be noted that the real-time key motor parameter values ​​P are obtained through sensors such as current, voltage, temperature, vibration, and speed. real The tolerance limit of the parameter, i.e., the activation threshold P, is determined through experimental or empirical data. max Then, the activation proximity value ΔP is calculated. margin This represents the safety margin between the current state and the activation threshold, where the proximity threshold P is set. thresh This is a predefined tolerance range used to determine whether a digital twin model needs to be invoked for subsequent predictive analysis.

[0060] Understandably, the purpose of determining whether the activation threshold of the digital twin model has been reached is:

[0061] Function 1: By setting an activation threshold, the digital twin model is only triggered when the motor's operating state approaches danger or abnormality, thus avoiding over-response to slight fluctuations or changes within the normal range.

[0062] Function 2: Digital twin models have high computational costs and large data requirements. By using a threshold judgment mechanism, they can be activated on demand, avoiding unnecessary long-term operation of the model.

[0063] Understandably, by setting activation thresholds, accurate identification and graded response to abnormal motor conditions can be achieved. Digital twin modeling analysis is only initiated when the potential risk of fault is high, avoiding ineffective consumption of system resources, improving the response efficiency and practical value of predictive maintenance, effectively reducing false alarm rate and missed alarm rate, making fault warning more timely and accurate, and enhancing the system's intelligent and refined operation and maintenance capabilities.

[0064] State modeling module: Based on historical operating data, real-time sensor data and environmental factors, it calls the digital twin model to build a virtual mapping of motor operation, generates the current state evolution trajectory, predicts whether there is a fault in the operating trend, and if so, enters the fault risk assessment module.

[0065] The process in the state modeling module of calling the digital twin model to construct a virtual mapping of motor operation and generating the current state evolution trajectory is as follows:

[0066] Construct a multi-dimensional time-series curve of the motor state based on the current key operating parameters of the motor, with the X-axis representing time and the Y-axis representing each operating parameter;

[0067] Based on historical operating data, a state evolution trajectory curve of the motor under standard conditions is constructed as a reference curve;

[0068] On the current state timing curve of the motor and the reference state trajectory curve, select several sampling points with the same time point;

[0069] Obtain the operating parameter values ​​of the motor's current state curve at different sampling points and integrate them into the first state data group;

[0070] Obtain the values ​​of each running parameter at the same sampling point of the reference state trajectory curve, and integrate them into a second state data group;

[0071] Calculate the cosine similarity between the first state data set and the second state data set to obtain the trajectory similarity value;

[0072] For example, suppose the first state data set (parameters on the current state curve of the motor) is: A = {A1, A2, ..., A...} nThe second state data set (parameters on the reference state trajectory curve) is: B = {B1, B2, ..., B} n};

[0073] Calculate the cosine similarity between the first-state data set and the second-state data set to obtain the trajectory similarity value. The specific calculation formula is as follows:

[0074]

[0075] Set curve similarity threshold D thresh :

[0076] If the trajectory similarity value is less than the threshold D thresh This indicates that the curve type is linear;

[0077] If the trajectory similarity value is greater than or equal to the threshold D thresh This indicates that the curve type is non-linear.

[0078] The state modeling module, by combining curve type analysis results to predict the subsequent motor operating state and whether a fault exists in the operating trend, includes the following process:

[0079] If the curve type is linear, the least squares method is used to fit the trend curve of the key operating parameters of the motor to obtain the function equation of the fitted line, y = k*t + b, where y represents the value of the key operating parameter, t represents time, k represents the slope of the fitted line, and b represents the intercept.

[0080] The critical threshold of the operating parameters is used as the input of the function equation to predict the time point at which the operating parameters reach the fault threshold, i.e. the critical time point.

[0081] Compare the critical time point with the end time of the current maintenance cycle:

[0082] If the critical time point is less than the end time point of the maintenance cycle, it means that the motor is not in a safe operating state during the current cycle.

[0083] If the critical time point is greater than or equal to the end time point of the maintenance cycle, it indicates that the motor is operating safely within the current cycle.

[0084] The state modeling module, which combines curve type analysis results to predict the subsequent motor operating state and whether a fault exists in the operating trend, also includes the following process:

[0085] If the curve type is nonlinear, a growth analysis is performed on the first-state data set to determine the fastest growth rate of the key operating parameters. The first-state data set consists of historical change data of the key parameters over time. By performing a difference operation on it, the rate of change of the parameters at each moment is calculated, and the maximum growth rate is determined from this.

[0086] Calculate the difference between the current value of the key parameter and the fault threshold to obtain the parameter approximation value. Divide the parameter approximation value by the fastest growth rate to obtain the time it takes for the parameter to reach the fault threshold, i.e., the critical time. Add the critical time to the current time point to obtain the critical time point.

[0087] If the critical time point is less than the end time point of the maintenance cycle, it indicates that the motor will fail within this cycle.

[0088] If the critical time point is greater than or equal to the end time point of the maintenance cycle, it means that the motor will not fail within this cycle.

[0089] Fault Risk Assessment Module: By matching the state trajectory change pattern with the historical fault database, it assesses whether there are potential fault paths. If so, it proceeds to the intervention needs analysis module.

[0090] The technical solution of this embodiment is as follows: By setting the normal operating range of key operating parameters, the sensor operating data of the motor is collected in real time, and combined with historical operating data and current environmental factors, the data is input into a trained digital twin model to construct a virtual mapping of the motor's current operating state. If the key operating parameters deviate from the normal operating range, a process is triggered to determine whether the activation threshold of the twin model has been reached, further generating the evolution trajectory of the current state. By performing trend modeling and type analysis on the evolution trajectory, it is determined whether it shows a trend towards a fault state. If the trend analysis results show potential fault risks, the process enters the fault risk assessment module, providing data support and decision-making basis for subsequent predictive maintenance and control. This not only improves the visualization and trend prediction capabilities of the motor's operating state but also effectively reduces the probability of sudden faults, improving the safety, stability, and maintenance efficiency of equipment operation.

[0091] Example 2

[0092] like Figure 1 As shown, the ultra-efficient explosion-proof motor predictive maintenance system based on digital twins also includes the following modules:

[0093] Intervention Needs Analysis Module: Based on the fault risk level and expected occurrence time, combined with the current production plan and equipment health tolerance, determine whether it is necessary to implement early intervention measures. If so, proceed to the intervention strategy generation module.

[0094] The intervention demand analysis module, by combining the current production plan with equipment health tolerance, determines whether advance intervention measures are needed as follows:

[0095] Obtain current production plan information and the equipment health tolerance of the motor, including the tolerance and remaining health life of the equipment under different health conditions; the equipment health tolerance refers to the boundary of the motor's ability to continue operating within a range of no failure risk or acceptable risk under the current health condition, reflecting the equipment's remaining operating capacity and risk tolerance in the face of potential operating loads or failure trends.

[0096] Set a health score range: for example, 0-100, where: 80-100 indicates good health; 50-79 indicates tolerable but requires attention; 0-49 indicates poor health; health tolerance is defined as the distance from the current score to the lower tolerance limit: TL = current health score - tolerance threshold score;

[0097] Set a maximum tolerable failure probability threshold. If the current predicted failure probability is close to or exceeds this threshold, it is considered to exceed the tolerance: TL = Tolerance threshold probability - Current failure prediction probability.

[0098] If the predicted remaining lifespan of the equipment is less than the required operating time plus the buffer period, it is judged as not having tolerance: TL = predicted remaining lifespan - required operating time * safety factor.

[0099] Combine equipment health status data with production planning data to build equipment health data groups and production task data groups. These two data groups need to be paired at the same point in time or under the same working conditions to ensure data matching.

[0100] The correlation between the health data set and the production plan data set is calculated using the Pearson correlation coefficient. The Pearson correlation coefficient measures the degree of fit between equipment health status and production plan. The formula is:

[0101]

[0102] In the formula, X i Y i These represent the observations from the health status data group and the production planning data group, respectively. These represent the mean values ​​of the health status data group and the production plan data group, respectively.

[0103] The system calculates the gap between the equipment's health status and its tolerance threshold. If the equipment's health status is close to or exceeds the upper limit of its health tolerance, and the production plan places high demands on the equipment's load, it indicates that the equipment may not be able to operate stably and there is a risk of failure. If the Pearson correlation coefficient shows a strong correlation between the health status and the production task, and the equipment's health status is about to exceed the tolerance threshold, the system will determine whether to take early intervention measures, taking into account the urgency of the production plan and the task priority. When the system determines that early intervention is necessary, it will initiate corresponding intervention measures, such as adjusting the production task load, shortening equipment uptime, performing partial maintenance, or activating backup equipment, to ensure continuous equipment operation and reduce the risk of failure.

[0104] Intervention strategy generation module: Based on the state control parameters fed back from the digital twin model and combined with the specific working conditions, an intervention plan is generated.

[0105] The process of generating an intervention plan in the intervention strategy generation module, based on specific working conditions, is as follows:

[0106] After modeling and predicting the motor's operating status, the digital twin model outputs key control parameters, such as current operating load, safety margin, vibration level, temperature rise trend, and cooling efficiency. These parameters are used to characterize the current operating health status and potential risks of the system.

[0107] Synchronously collect operating condition information related to equipment operation, including: current production cycle and task load; historical maintenance records and current maintenance status of equipment; environmental conditions; and available resources.

[0108] Compare the control parameters suggested by the digital twin model with the current operating conditions to determine whether there are any conflicts or potential risks, and whether the intervention trigger conditions are met, such as: insufficient safety margin; key parameters approaching or exceeding the warning threshold; changes in operating conditions leading to parameter instability.

[0109] Locate and match applicable intervention strategy libraries, and select the intervention measures most suitable for the current scenario, including: reducing load operation; partial shutdown for inspection; advance lubrication and maintenance; activating backup modules or allocating production line resources, etc.

[0110] Based on the selected strategy, combined with resource scheduling and production tasks, a complete intervention plan is formed, including intervention content, time arrangement, resource requirements, and execution steps. An impact assessment is conducted on the generated intervention plan to analyze its impact on production continuity, operation and maintenance costs, and equipment lifespan. If necessary, it is submitted for manual review for confirmation or adjustment. After the final plan is confirmed, it is automatically distributed to relevant personnel and system modules for execution, such as scheduling downtime, pushing maintenance work orders, and allocating necessary spare parts.

[0111] Example 3

[0112] Please see Figure 2 As shown, the predictive maintenance method for ultra-efficient explosion-proof motors based on digital twins includes the following steps:

[0113] S1: When the key operating parameters of the motor deviate from the normal operating range, determine whether the activation threshold of the digital twin model has been reached. If it has, enter the state modeling module.

[0114] The process of determining whether the activation threshold of the digital twin model has been reached when key operating parameters of the motor deviate from the normal operating range is as follows:

[0115] Set key operating parameters, including real-time parameter value P. real Operating condition warning value P alert With the activation threshold P of the digital twin model max The formula for calculating the deviation of the current parameter from the normal operating range is as follows:

[0116] ΔP alert =P real -P alert

[0117] If ΔP alert >0 indicates that the motor operating parameters have exceeded the warning value and need to be further evaluated to see if they are close to the activation threshold;

[0118] Calculate the difference between the current parameter and the activation threshold, i.e., the activation proximity value ΔP. margin :

[0119] ΔP margin =P max -P real

[0120] If the activation proximity value is less than the activation proximity threshold P thresh Then it is necessary to predict the state of the current parameters;

[0121] If the activation proximity value is greater than or equal to the activation proximity threshold P thresh If so, there is no need to predict the state of the current parameters.

[0122] S2: Based on historical operating data, real-time sensor data, and environmental factors, the system calls the digital twin model to construct a virtual mapping of motor operation, generates the current state evolution trajectory, predicts whether there is a fault in the operating trend, and if so, proceeds to the fault risk assessment module.

[0123] The process of calling the digital twin model to construct a virtual mapping of motor operation and generating the current state evolution trajectory is as follows:

[0124] Construct a multi-dimensional time-series curve of the motor state based on the current key operating parameters of the motor, with the X-axis representing time and the Y-axis representing each operating parameter;

[0125] Based on historical operating data, a state evolution trajectory curve of the motor under standard conditions is constructed as a reference curve;

[0126] On the current state timing curve of the motor and the reference state trajectory curve, select several sampling points with the same time point;

[0127] Obtain the operating parameter values ​​of the motor's current state curve at different sampling points and integrate them into the first state data group;

[0128] Obtain the values ​​of each running parameter at the same sampling point of the reference state trajectory curve, and integrate them into a second state data group;

[0129] Calculate the cosine similarity between the first state data set and the second state data set to obtain the trajectory similarity value.

[0130] S3: By matching the state trajectory change pattern with the historical fault database, assess whether there are potential fault paths. If there is a clear trend, proceed to the intervention requirement analysis module.

[0131] S4: Based on the fault risk level and the expected occurrence time, combined with the current production plan and equipment health tolerance, determine whether it is necessary to implement early intervention measures. If so, proceed to the intervention strategy generation module.

[0132] The process of determining whether to implement early intervention measures by combining the current production plan with the equipment health tolerance is as follows:

[0133] Obtain current production plan information and the equipment health tolerance of the motor, including the tolerance of the equipment under different health conditions and the remaining health life;

[0134] The correlation between the health data set and the production plan data set is calculated using the Pearson correlation coefficient. The Pearson correlation coefficient measures the degree of fit between equipment health status and production plan. The formula is:

[0135]

[0136] In the formula, X i Y i These represent the observations from the health status data group and the production planning data group, respectively. These represent the mean values ​​of the health status data group and the production plan data group, respectively.

[0137] S5: Based on the state control parameters fed back from the digital twin model and combined with the specific working conditions, generate an intervention plan.

[0138] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0139] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A predictive maintenance system for ultra-high efficiency explosion-proof electric motor based on digital twin, characterized in that, Comprise: Running state recognition module: when the motor key operation parameter deviates from the normal working condition interval, judge whether the digital twin model activation threshold is reached, if reached, enter the state modeling module; The process of judging whether the digital twin model activation threshold is reached is: Set the key operation parameter, calculate the deviation degree of the current parameter relative to the normal operation interval; If the deviation degree is greater than 0, it means that the motor operation parameter has exceeded the warning value, and whether it is close to the activation threshold is evaluated; The evaluation of whether it is close to the activation threshold includes: Calculate the difference between the current parameter and the activation threshold to get the activation proximity value; If the activation proximity value is less than the activation proximity threshold, the current parameter needs to be state predicted; If the activation proximity value is greater than or equal to the activation proximity threshold, the current parameter does not need to be state predicted; State modeling module: based on historical operation data, real-time sensor data and environmental factors, call the digital twin model to construct the virtual mapping of motor operation, and generate the current state evolution track, predict whether the running trend exists fault, if exists, enter the fault risk assessment module; The process of generating the current state evolution track is: According to the current key operation parameter of the motor, a motor state multi-dimensional time sequence curve is constructed with X axis as time and Y axis as each operation parameter; Based on the historical operation data, the state evolution track curve of the motor in the standard state is constructed as a reference curve; Select several sampling points with the same time on the motor current state time sequence curve and the reference state trajectory curve; Get the operation parameter values of the motor current state curve at different sampling points and integrate them into the first state data set; Get the operation parameter values of the reference state trajectory curve at the same sampling points and integrate them into the second state data set; Calculate the cosine similarity between the first state data set and the second state data set to get the trajectory similarity value; Set the curve similarity threshold value: If the trajectory similarity value is less than the curve similarity threshold value, the curve type is linear; If the trajectory similarity value is greater than or equal to the curve similarity threshold value, the curve type is nonlinear; The process of predicting whether the running trend exists fault includes: If the curve type is linear, use the least square method to fit the change trend curve of the motor key operation parameter to get the function equation of the fitted straight line; Take the fault threshold value of the operation parameter as the input of the function equation to predict the time point when the operation parameter reaches the fault threshold, i.e. the critical time point; The process of predicting whether the running trend exists fault also includes: If the curve type is nonlinear, perform growth analysis on the first state data set to determine the fastest growth rate of the key operation parameter. The first state data set is the historical change data of the key parameter in the time dimension, and the maximum growth rate is determined by difference operation. Calculate the difference between the current value of the key parameter and the fault threshold to get the parameter proximity value, and divide the parameter proximity value by the fastest growth rate to get the time length of the parameter reaching the fault threshold, i.e. the critical time length, and the critical time point is the critical time length plus the current time point. ​ Failure risk assessment module: match the state trajectory change pattern with the historical failure library to assess whether there is a potential failure path, if there is, enter the intervention demand analysis module; Intervention demand analysis module: according to the failure risk level and the predicted occurrence time, combined with the current production plan and the equipment health tolerance, judge whether it is necessary to implement the early intervention measures, if necessary, enter the intervention strategy generation module; Intervention strategy generation module: according to the state control parameters fed back by the digital twin model, combined with the specific working condition, generate the intervention scheme.

2. The digital-twin-based predictive maintenance system for ultra-high- efficiency explosion-proof electric machines according to claim 1, characterized in that, Compare the critical time point with the current maintenance cycle end time point: If the critical time point is less than the maintenance cycle end time point, it means that the motor running state is not safe within the current cycle; If the critical time point is greater than or equal to the maintenance cycle end time point, it means that the motor running state is safe within the current cycle.

3. The digital-twin-based ultra-high-efficiency explosion-proof motor predictive maintenance system of claim 1, wherein, The process of judging whether it is necessary to implement the early intervention measures is: Get the current production plan information and the equipment health tolerance of the motor; Combine the health status data of the equipment with the production plan data to construct the equipment health data group and the production task data group; Calculate the correlation between the health data group and the production plan data group using Pearson correlation coefficient.

4. The digital-twin-based ultra-high-efficiency explosion-proof motor predictive maintenance system of claim 1, wherein, The process of generating intervention scheme is: Collect the working condition information related to equipment operation, find and match the intervention strategy library, select the intervention measures suitable for the current scene, according to the selected strategy, combined with the resource scheduling situation and the production task, form the intervention scheme.

Images