Power equipment health assessment and fault prediction system

Abstract

The invention discloses a power equipment health assessment and fault prediction system, and relates to the technical field of power equipment monitoring and maintenance, and the system comprises a data collection module, a data processing and storage module, a short-term fault prediction model, a medium-term fault prediction model, a long-term fault prediction model, a model fusion and decision module, and a user interaction interface. By integrating short-term, medium-term and long-term fault prediction models, comprehensive evaluation and fault prediction of the health condition of the power equipment are realized, the short-term model can quickly capture abnormal changes of the current operation state of the equipment and provide instant fault early warning, and the medium-term model is combined with a mechanism model through time sequence analysis, so that the fault early warning efficiency is improved. The performance change trend of the equipment in a period of time in the future is predicted, decision support is provided for planned maintenance work, the long-term model focuses on residual life prediction of the equipment, and important reference is provided for long-term operation and maintenance planning of the equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment monitoring and maintenance, and in particular to a power equipment health assessment and fault prediction system. Background Art

[0002] With the rapid development of the power industry, the health status and fault prediction of power equipment are crucial to ensuring the stable operation of the power system. During operation, power equipment will be affected by various internal and external factors, causing its performance to gradually decline and may eventually fail. Therefore, health assessment and fault prediction of power equipment in order to timely detect potential problems and take corresponding measures has become an important research direction in the power industry.

[0003] Traditional power equipment health assessment and fault prediction mainly rely on manual inspections and regular tests. This method is not only time-consuming and labor-intensive, but also difficult to achieve real-time monitoring and accurate prediction of equipment status. In addition, traditional methods are often based on empirical judgments and lack scientific data support and intelligent analysis methods, resulting in inaccurate assessment results and insufficient fault prediction capabilities. Specifically, traditional technologies have the following shortcomings: First, data collection and processing capabilities are limited, and it is impossible to obtain comprehensive equipment operating status information in real time; second, the fault prediction model is single and cannot adapt to the diversity and complexity of equipment failures; third, the operation and maintenance strategy lacks specificity, and it is impossible to formulate personalized operation and maintenance plans based on equipment health status and fault risk levels.

[0004] Therefore, the development of a power equipment health assessment and fault prediction system will greatly improve the accuracy and reliability of power equipment fault prediction and provide a strong guarantee for the stable operation of the power system. Summary of the invention

[0005] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a power equipment health assessment and fault prediction system, which integrates data collection, processing, storage and short-term, medium-term and long-term fault prediction models, and realizes accurate assessment of equipment health status and fault warning through model fusion and decision-making modules. Compared with traditional technologies, the system can detect potential faults earlier, improve prediction accuracy, reduce unplanned downtime, and optimize operation and maintenance strategies. Through the user interaction interface, operation and maintenance personnel can intuitively obtain equipment health information and operation and maintenance suggestions, thereby improving work efficiency.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a power equipment health assessment and fault prediction system, the system comprising: a data acquisition module, a data processing and storage module, a short-term fault prediction model, a medium-term fault prediction model, a long-term fault prediction model, a model fusion and decision module, and a user interaction interface;

[0007] The data acquisition module is connected to sensors on the power equipment to collect parameter data including electrical, mechanical and environmental aspects, and transmits the data to the data processing and storage module in real time through a wireless communication link;

[0008] The data processing and storage module receives the data transmitted by the data acquisition module, cleans, filters and normalizes the received data, removes noise and abnormal values, extracts feature vectors in the power equipment data through the power data similarity feature extraction formula, and classifies and stores the processed data in short-term, medium-term and long-term data storage areas according to the time scale, and constructs a search index and management system at the same time;

[0009] The short-term fault prediction model is constructed based on the current operation status change data of the power equipment from a few minutes to a few hours. The operation status vector of the power equipment at time t is X(t)=[x1(t),x2(t),…,x n (t)] T , where x i (t) represents the i-th operating parameter, and the formula is: in, is the parameter x i In the past ΔT s The mean value within the time period, σ i 9t-ΔT s ) is the corresponding standard deviation, w i,s is the weight of the i-th parameter in the short-term model, α s is an index of abnormal sensitivity to regulation, when P s (t)>Short-term model fault judgment threshold T s,th , triggering a short-term fault warning, immediately outputting detailed fault warning information, including the fault type, the location of the fault and related component information, and the abnormal feature vector V s (t) = [v s,1 (t),v s,2 (t),…,v s,n (t)] T is passed to the mid-term model, where:

[0010] The mid-term fault prediction model: by deeply mining and analyzing the historical data in the mid-term data storage area, combined with the abnormal feature vector V transmitted by the short-term model s (t) Construct a mid-term failure prediction model, the formula is: Among them, z j (t) is the jth mid-term characteristic parameter in the mid-term data storage area, w i,m and θ j,m are the corresponding weights, βm It is a coefficient related to time decay. Real-time data is input into the model for prediction, and a detailed performance change curve of the equipment in the next 3 days to 2 weeks is output. The estimated time point when the performance drops to the mid-term model fault judgment threshold is predicted and sent to the operation and maintenance personnel to arrange planned maintenance work, and a mid-term key performance indicator vector K is generated. m (t+ΔT m )=[k m,1 (t+ΔT m ),k m,2 (t+ΔT m ),…,k m,p (t+ΔT m )] T Passed to the long-term model, where k m,q (t+ΔT m ) represents the qth mid-term key performance indicator forecast value;

[0011] The long-term fault prediction model: The long-term fault prediction model is constructed by combining theory with a data-driven method. The formula is: Among them, y r (t) is the rth long-term characteristic parameter in the long-term data storage area, ω q,l and φ r,l is the corresponding weight, and the remaining life of the power equipment is predicted through the model, and the results are fed back to the short-term and medium-term models to adjust their fault judgment thresholds and accuracy requirements;

[0012] The model fusion and decision-making module: based on the prediction results of the short-term fault prediction model, the medium-term fault prediction model and the long-term fault prediction model, assign corresponding dynamic weights to them, use the weighted summation strategy to obtain the health assessment index of the overall condition of the equipment, divide the equipment failure into multiple risk levels according to the health assessment index, and formulate personalized operation and maintenance strategies for different levels, receive new data and feedback in real time, and optimize the parameters and structures of each model according to the new data and operation and maintenance feedback;

[0013] The user interaction interface: displays equipment health assessment, fault prediction information and operation and maintenance suggestions in intuitive charts and texts, has a historical data query function, and supports access from multiple terminal devices.

[0014] Furthermore, the sensor used in the data acquisition module:

[0015] Electrical sensors: voltage transformers, current transformers, and power sensors;

[0016] Mechanical sensors: accelerometers, displacement sensors, and strain gauges;

[0017] Environmental sensors: temperature sensor, humidity sensor, air pressure sensor.

[0018] Furthermore, the data processing and storage module extracts the feature vector in the power equipment data by using the power data similarity feature extraction formula. Assume that the original data sequence is D = {d1, d2, ..., d N}, the extracted feature vector F = [f1,f2,…,f M ], the calculation formula is: Among them, k(d i ,d j ) is the data point d i With d j The similarity kernel function between is as follows: σ is the bandwidth parameter of the kernel function, g(d i -d j ) is the weighted function of the data difference, and the formula is: α is the adjustment parameter, λ is the attenuation coefficient, is the mean of the data sequence D.

[0019] Furthermore, the short-term model fault judgment threshold T in the short-term fault prediction model s,th Determine the normal operation data and obtain A s The mean of (t) and standard deviation The calculation formula of the short-term model fault judgment threshold is: Where n is a multiple determined based on the tolerance for failure risk.

[0020] Furthermore, the mid-term model fault judgment threshold T in the mid-term fault prediction model m,th,i For a specific key performance indicator KPI, i The initial design value of this indicator is KPI i,init , the performance degradation ratio limit is α i , for this key performance indicator KPI i The mid-term model fault judgment threshold T m,th,i The calculation formula is: m,th,i =KPI i,init ×(1-α i ).

[0021] Furthermore, in the long-term fault prediction model, the fault judgment thresholds of the short-term model and the mid-term model are adjusted according to the long-term model:

[0022] For the adjustment of the short-term model fault judgment threshold, the remaining life ratio of the equipment evaluated by the long-term model is assumed to be R L 9t), 0≤R L 9t)≤1,R L9t)=1 means the equipment is brand new, R L (t0=0 means the equipment life is exhausted, T s,th is the original short-term model fault judgment threshold, α and β are adjustment parameters, and the new short-term model fault judgment threshold formula is: T s,new,th =T s,th ×(1-α×R L (t))+β×(1-R L (t));

[0023] For the adjustment of the mid-term model fault judgment threshold, the mid-term model fault judgment threshold is set to T m,th,i The equipment performance degradation rate estimated by the long-term model is D L (t), λ and μ are adjustment parameters, and the new medium-term and short-term model fault judgment threshold calculation formula is: T m,new,th =T m,th,i ×(1-λ×D L (t))+μ×D L (t).

[0024] Furthermore, the model fusion and decision module uses dynamic weights to calculate the reliability index of each model. Let the short-term model reliability index be R s , calculate the prediction accuracy in the recent time period, the number of correct predictions is N sc , the total number of predictions is N st , the timeliness weight w of the recent prediction t , and w t =e -μ·t , where t is the time interval from the current moment, μ is the attenuation coefficient, and the calculation formula is: Medium-term model reliability index R m , let the mid-term prediction performance change curve be C m (t), the actual performance change curve is C ma (t), in the time period [t1, t2], by calculating the similarity S of the two curves m , and combined with the data integrity index I m , the formula is: R m =S m I m ; Long-term model reliability index R l , assuming that the long-term remaining life prediction value is L p , the actual remaining life is L a , the confidence interval width is W l , the formula is: Based on the reliability index, the weights w assigned to the short-term, medium-term, and long-term models are s 、w m 、w l They are: where β s , β m , β l is the adjustment index.

[0025] Furthermore, the health assessment index of the overall condition of the equipment in the model fusion and decision module is used. Suppose the prediction result of the short-term fault prediction model is P s , the prediction result of the mid-term fault prediction model is P m , the prediction result of the long-term fault prediction model is P l , the revolutionary machine cover pairing weight, the short-term model weight is w s , the medium-term model weight is w m , the long-term model weight is w l , the overall health evaluation index H of the computing equipment is calculated as follows: Among them, k s , k m , k l is the adjustment coefficient.

[0026] Furthermore, the model fusion and decision-making module divides the risk levels and formulates personalized operation and maintenance strategies for different levels. Let the overall health assessment index of the equipment be H, and define the low risk threshold T low , medium risk threshold T mid , and 0<T low <T mid <1:

[0027] When H>T low When the device is in a low risk state, the monitoring frequency f low Set as basic monitoring frequency f base , data is collected every hour, simple inspections are conducted every quarter, and resource allocation mainly focuses on the guarantee of daily maintenance resources;

[0028] When T mid <H≤T low When the device is in a medium risk state, the monitoring frequency is f mid Improve f base ·k mid (k mid >1), maintenance plan is formulated for the next T pm Detailed inspection within the time, T pm It can be determined based on the mid-term model forecast results, and the calculation formula is: ΔT m,th is the time difference between the mid-term performance drop to the threshold, P m (t) is the current medium-term performance forecast, ΔP m is the mid-term performance change rate, and resource allocation increases the reserve ratio of emergency maintenance resources rmid ;

[0029] When H≤T mid When the device is in a high-risk state, the monitoring frequency f high Significantly increased to f base · k high , k high >k mid , immediately start the shutdown maintenance plan, maintenance time T repair Calculated based on the fault type and maintenance resources, the formula is: where t ri is the time required for the ith maintenance step, r i It is the resource sufficiency rate for executing the i-th maintenance step. Resource allocation fully guarantees the human, material and technical support required for maintenance.

[0030] Furthermore, the model fusion and decision module low risk threshold H low and medium risk threshold H mid Determination, assuming that the mean H value of healthy equipment in historical data is μ H , with standard deviation σ H , low risk threshold H low Set to: H low =μ H -k1σ H , where k1 is a coefficient determined based on the risk tolerance, and the medium risk threshold H mid Set to: H mid =μ H -k2σ H , where k2>k1.

[0031] Compared with the existing technology, the power equipment health assessment and fault prediction system has the following beneficial effects:

[0032] 1. The present invention realizes a comprehensive assessment of the health status and fault prediction of power equipment by integrating short-term, medium-term and long-term fault prediction models. The short-term model can quickly capture abnormal changes in the current operating status of the equipment and provide immediate fault warnings. The medium-term model combines time series analysis with mechanism models to predict the performance change trend of the equipment in the future and provide decision support for planned maintenance work. The long-term model focuses on the remaining life prediction of the equipment and provides an important reference for the long-term operation and maintenance planning of the equipment. This multi-time scale fault prediction and health assessment method improves the accuracy and practicality of the prediction.

[0033] 2. The present invention introduces dynamic weight allocation and personalized operation and maintenance strategy formulation functions in the model fusion and decision-making module. By calculating the reliability index of each model, the system can dynamically adjust the weights of short-term, medium-term and long-term models in health assessment, thereby ensuring the objectivity and accuracy of the assessment results. At the same time, the system can also formulate personalized operation and maintenance strategies based on the overall health assessment indicators of the equipment and the division of risk levels, including adjustment of monitoring frequency, formulation of maintenance plans and optimization of resource allocation. This intelligent operation and maintenance decision support function not only improves operation and maintenance efficiency, but also reduces operation and maintenance costs.

[0034] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0036] Figure 1 This is a flow chart of the power equipment health assessment and fault prediction system;

[0037] Figure 2 The architecture diagram of the power equipment health assessment and fault prediction system. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] Embodiment 1:

[0040] Substation transformer equipment health assessment and fault prediction

[0041] Data collection and processing: In the substation, data collection modules are deployed for transformer equipment. Electrical sensors (voltage transformers, current transformers, power sensors) collect electrical parameters, mechanical sensors (accelerometers, displacement sensors, strain gauges) monitor mechanical status, and environmental sensors (temperature, humidity, and air pressure sensors) obtain environmental information. The collected data is transmitted to the processing and storage module in real time via wireless communication for cleaning, filtering, and normalization to remove noise and outliers. The feature vectors in the power equipment data are extracted through the power data similarity feature extraction formula. Suppose the original data sequence is D = {d1, d2, …, d N}, the extracted feature vector F = [f1,f2,…,f M ], the calculation formula is: Among them, k9d i ,d j ) is the data point d i With d j The similarity kernel function between is calculated as follows: σ is the bandwidth parameter of the kernel function, g(d i -d j ) is the weighted function of the data difference, and the calculation formula is: α is the adjustment parameter, λ is the attenuation coefficient, It is the mean of the data sequence D, and then it is classified and stored according to the time scale.

[0042] Model prediction and analysis:

[0043] Short-term fault prediction: A short-term fault prediction model is constructed based on the current operating status of the transformer from a few minutes to a few hours. The operating status vector of the transformer at time t is X(t) = [x1(t), x2(t), …, x n (t) T , where x i (t) represents the i-th operating parameter. The model is constructed according to the short-term abnormality index calculation formula, which is: in, is the parameter x i In the past ΔT s The mean value within the time period, σ i (t-ΔT s ) is the corresponding standard deviation, w i,s is the weight of the i-th parameter in the short-term model, α s is an index of abnormal sensitivity to regulation, when P s (t)>T s,th (short-term fault threshold), trigger the short-term fault warning, output the fault warning information, and convert the abnormal feature vector V s (t) = [v s,1 (t),vs,2 (t),…,v s,n (t)] T is passed to the mid-term model, where

[0044] Mid-term fault prediction: The mid-term fault prediction model is constructed by combining the time series analysis method with the variable instrument operation mechanism model. The historical data in the mid-term data storage area is deeply mined and analyzed, and the abnormal feature vector V transmitted by the short-term model is combined. s (t) Construct a mid-term failure prediction model, the formula is: Among them, z j (t) is the jth mid-term characteristic parameter in the mid-term data storage area, w i,m and θ j,m are the corresponding weights, β m It is a coefficient related to time decay. Through the historical data optimization model, the independent data evaluation model is used to input real-time data into the prediction, output the performance change curve of the equipment in the next 3 days to 2 weeks and the estimated time point when the predicted performance drops to the warning threshold, and generate the mid-term key performance indicator vector K m (t+ΔT m )=[k m,1 (t+ΔT m ),k m,2 (t+ΔT m ),…,k m,p (t+ΔT m )] T Passed to long-term model;

[0045] Long-term fault prediction: The long-term fault prediction model is constructed by combining theory with data-driven methods. The formula is: Among them, y r (t) is the rth long-term characteristic parameter in the long-term data storage area, ω q,l and The corresponding weight is used to predict the remaining life of the transformer through the model, and to feed back the short-term and medium-term models to adjust the fault judgment threshold and accuracy requirements.

[0046] Model fusion and decision-making, the model fusion and decision-making module assigns dynamic weights based on the prediction results of the short-term, medium-term and long-term fault prediction models, calculates the reliability index of each model, and sets the reliability index of the short-term model as R s , calculate the prediction accuracy in the recent time period, the number of correct predictions is N sc , the total number of predictions is N st , the timeliness weight w of the recent prediction t =e -μ·t , t is the time interval from the current moment, μ is the attenuation coefficient, and the formula is: Medium-term model reliability index R m , let the mid-term prediction performance change curve be C m (t), the actual performance change curve is C ma (t), in the time period [t1, t2], by calculating the similarity S of the two curves m , and combined with the data integrity index I m , the calculation formula is: R m =S m I m ; Long-term model reliability index R l , assuming that the long-term remaining life prediction value is L p , the actual remaining life is L a , the confidence interval width is W l , the calculation formula is: Based on the reliability index, the weights w assigned to the short-term, medium-term, and long-term models are s 、w m 、w l They are: where β s , β m , β l To adjust the index, calculate the overall health assessment index H of the equipment, and assume that the prediction result of the short-term fault prediction model is P s , the prediction result of the mid-term fault prediction model is P m , the prediction result of the long-term fault prediction model is P l , the calculation formula is: where k s , k m , k l is the adjustment coefficient, and the low risk threshold is T low , medium risk threshold is T mid (0 <T low <T mid <1), when T>T low When the equipment is in a low risk state, the monitoring frequency is set to the basic monitoring frequency f base , when T mid <H≤T low When H≤T mid When the equipment is in a high-risk state, the monitoring frequency is greatly increased and the shutdown and maintenance plan is initiated. At the same time, personalized operation and maintenance strategies and resource allocation plans are formulated according to different risk levels.

[0047] In summary, this embodiment takes the substation transformer equipment as the core, constructs a comprehensive data collection system, uses multiple types of sensors to collect electrical, mechanical, and environmental data and standardizes processing and storage. Short-term fault prediction captures sudden faults in real time based on the operating parameter vector and the abnormality index formula; the medium-term model integrates multivariate data and mechanism knowledge to predict the medium-term performance trend; the long-term model estimates the remaining life based on characteristic parameters and feeds back to optimize the previous model. The model fusion module evaluates the health of the equipment based on the prediction results, accurately divides the risk level, sets the monitoring frequency, plans maintenance inspection plans, and allocates resources based on the risk level, forming a closed-loop intelligent operation and maintenance architecture from data to decision-making, which comprehensively guarantees the reliable operation of the transformer, reduces the operation and maintenance costs, and improves the stability of the power grid, providing an efficient and accurate example for substation operation and maintenance management, and has significant engineering application and promotion value.

[0048] Embodiment 2:

[0049] Health Assessment and Fault Prediction of Steam Turbine Generator Units in Thermal Power Plants

[0050] Data acquisition and processing: Data acquisition modules are carefully laid out in the steam turbine generator area of ​​the thermal power plant. Electrical sensors (voltage transformers and current transformers accurately measure the power generation voltage and current parameters, and power sensors monitor the power output stability and power quality of the unit in real time) capture electrical operating parameters in all directions, mechanical sensors (accelerometers are tightly attached to the turbine shaft neck and bearing seat to sensitively sense vibration conditions; displacement sensors accurately monitor the axial and radial displacement of turbine blades to strictly control the blade running trajectory; strain gauges are firmly attached to the turbine cylinder and key parts of the pipeline to accurately measure stress and strain changes, providing a key basis for the health assessment of the mechanical structure) carefully monitor the mechanical performance status, and environmental sensors (temperature sensors are densely distributed around the machine room and cooling system, humidity sensors sense the ambient humidity in real time, and air pressure sensors stably monitor the air pressure in the machine room to lay a solid foundation for analyzing the impact of environmental factors on the unit. The environmental information is collected comprehensively through the high-speed wireless communication link. The massive data are instantly transmitted to the data processing and storage module with the help of high-speed wireless communication links. After strict cleaning (eliminating abnormal data points caused by electromagnetic interference and transient pulses), efficient filtering (using a combination of low-pass, high-pass and band-pass filters to filter out electrical noise and mechanical vibration clutter), and normalization (unifying various parameters into a standard numerical range to eliminate dimensional differences), they are intelligently classified and stored according to the time scale. The short-term data storage area accurately retains the high-frequency sampling data every minute within the past 2 hours. The medium-term storage area orderly archives the average data every 10 minutes within the past week. The long-term storage area systematically integrates the statistical values ​​and extreme values ​​of key parameters every month for the past year. At the same time, an intelligent search index and an efficient management system are built to ensure that data retrieval is fast, accurate, safe and reliable.

[0051] Model prediction and analysis:

[0052] Short-term fault prediction: A short-term fault prediction model is constructed based on the current 30-minute operating state vector of the steam turbine generator set. Assume that the operating state vector at time t is X(t) = [x1(t), x2(t), …, x n (t)] T , x i (t) is the key operating parameter (such as turbine speed deviation, lubricating oil temperature change rate, bearing vibration amplitude), and the model is constructed according to the short-term abnormality index formula, which is: in, is the parameter x i In the past ΔT s Mean, σ i (t-ΔT s ) is the standard deviation, w i,s is the parameter weight, α s Adjust the sensitivity, such as the vibration amplitude x of the turbine bearing at a certain moment j (t) mutation, calculated P s (t)>T s,th (The threshold is scientifically set based on the statistical analysis of normal operating data and risk tolerance), which immediately triggers a sharp fault warning, outputs the fault type (such as bearing wear fault), approximate location (a certain bearing number) and component information in detail, and generates the abnormal feature vector V s (t) Efficiently transmit to mid-term models to facilitate in-depth analysis;

[0053] Medium-term fault prediction: Deeply integrate the time series analysis method with the complex operation mechanism model of the steam turbine generator set to build a medium-term fault prediction model, deeply mine the historical data rich mine in the medium-term data storage area, and combine the short-term abnormal feature vector V s (t), accurately predict according to the formula, the formula is: Accurate prediction, where z j (t) is the mid-term characteristic parameter, w i,m and θ j,m is the weight, β m Correlate time decay, such as predicting the unit efficiency decline curve in the next 10 days, accurately determine the time point when key performance indicators (such as KPI corresponding to heat rate) reach the warning threshold, and the operation and maintenance team scientifically arranges a deep maintenance plan in 5 days (checking the scaling of the turbine flow part, adjusting the shaft alignment, and optimizing the lubrication system) based on this, and generates a mid-term key performance indicator vector K m (t+ΔT m ) is accurately passed to the long-term model;

[0054] Long-term fault prediction: The long-term fault prediction model is constructed by combining theory with data-driven methods. The formula is: Among them, y r(t) is the long-term characteristic parameter (such as the cumulative operating time of the unit, the number of major maintenance times, and the high-temperature creep damage assessment value), ω q,l , The remaining life of the unit is accurately predicted with weights. For example, for a unit that has been in operation for 15 years, has undergone three major overhauls, and whose long-term characteristic parameters have changed to specific values, the remaining life is predicted to be three years. Based on this key result, the short- and medium-term models are accurately fed back, the short-term fault threshold is dynamically optimized, the sensitivity is intelligently adjusted according to the remaining life ratio, and the medium-term performance threshold and prediction accuracy are finely corrected to ensure that the operation and maintenance strategy is closely adapted to the characteristics of each stage of the unit's aging process, effectively reducing the risk of failure, extending the service life of equipment, and improving the operating efficiency of thermal power plants.

[0055] Model fusion and decision-making: The model fusion and decision-making module efficiently integrates the results of short-term, medium-term and long-term fault prediction models, accurately assigns dynamic weights, and scientifically calculates the reliability indicators of each model. The short-term model is based on the recent prediction accuracy and the number of correct predictions N. sc With the total number of times N st Ratio combined with timeliness weight w t =e -μ·t , μ is the attenuation coefficient, determined by the formula: The medium-term model is based on the predicted performance curve and the actual similarity S m and Data Integrity I m Calculation formula: R m =S m I m ; The long-term model determines R by the formula based on the remaining life prediction deviation and the confidence interval width l , the formula is: Based on the reliability index, the weight is accurately assigned according to the formula: β s , β m , β l To adjust the index, calculate the overall health assessment index of the equipment, the formula is: H = w s · k s , k m , k l is the adjustment coefficient, and the risk level is accurately divided according to it, and the low risk threshold T is set low , medium risk threshold T mid , 0 <T low <T mid <1, when H>T low When the risk is low, the monitoring frequency is set to the normal value, and resource allocation focuses on daily maintenance material supply. mid <H≤T lowWhen H≤T, the monitoring frequency is increased (collection every 30 minutes, in-depth inspection every week), and maintenance resources are prepared within 7 days according to the medium-term forecast plan; when H≤T mid When the risk is high, the monitoring frequency increases dramatically (collection every 10 minutes, immediate shutdown for maintenance), and manpower, spare parts, and technical experts are deployed in all directions. Equipment is repaired quickly according to the emergency repair plan to ensure stable power generation in thermal power plants and improve the reliability and economic benefits of power supply.

[0056] In summary, this embodiment focuses on the steam turbine generator set in a thermal power plant and constructs a comprehensive, multi-level data-driven health assessment and fault prediction system. Electrical, mechanical and environmental data are collected through sensors deployed in all directions. After rigorous processing and classified storage, a solid data foundation is provided for models at each stage. The short-term model keenly captures the hidden dangers of sudden failures based on real-time parameters. The medium-term model integrates historical trends and short-term anomalies to accurately outline the performance evolution trajectory. The long-term model accurately quantifies the remaining life according to the characteristics of the unit's entire life cycle. The three work together and are dynamically optimized. The model fusion module condenses comprehensive health indicators and accurately defines risk levels based on a scientific weight allocation mechanism, and then customizes operation and maintenance planning and allocates resources according to risk characteristics. This system forms an efficient closed-loop operation and maintenance strategy, which effectively ensures the stable and efficient operation of the unit, significantly reduces the risk of unplanned downtime, deeply optimizes the allocation of operation and maintenance resources, and continuously improves the economic benefits and power supply reliability of thermal power plants. It sets an innovative example for the operation and maintenance management of key equipment in the energy industry, and has great industry promotion value and far-reaching practical significance.

[0057] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. Power equipment health assessment and fault prediction system, characterized in that: The system includes: a data acquisition module, a data processing and storage module, a short-term fault prediction model, a medium-term fault prediction model, a long-term fault prediction model, a model fusion and decision module, and a user interaction interface; The data acquisition module is connected to sensors on the power equipment to collect parameter data including electrical, mechanical and environmental aspects, and transmits the data to the data processing and storage module in real time through a wireless communication link; The data processing and storage module receives the data transmitted by the data acquisition module, cleans, filters and normalizes the received data, removes noise and abnormal values, extracts feature vectors in the power equipment data through the power data similarity feature extraction formula, and classifies and stores the processed data in short-term, medium-term and long-term data storage areas according to the time scale, and constructs a search index and management system at the same time; The short-term fault prediction model is constructed based on the current operation status change data of the power equipment from a few minutes to a few hours. The operation status vector of the power equipment at time t is X(t)=[x1(t),x2(t),…,x n (t)] T , where x i (t) represents the uth operating parameter, and the formula is: in, is the parameter x i In the past ΔT s The mean value within the time period, σ i (t-ΔT s ) is the corresponding standard deviation, w i,s is the weight of the i-th parameter in the short-term model, α s is an index of abnormal sensitivity to regulation, when P s (t)>Short-term model fault judgment threshold T s,th , triggering a short-term fault warning, immediately outputting detailed fault warning information, including the fault type, the location of the fault and related component information, and the abnormal feature vector V s (t) = [v s,1 (t),v s,2 (t),…,v s,n (t)] T is passed to the mid-term model, where: The mid-term fault prediction model: by deeply mining and analyzing the historical data in the mid-term data storage area, combined with the abnormal feature vector V transmitted by the short-term model s (t) Construct a mid-term failure prediction model, the formula is: Among them, z j (t) is the jth mid-term characteristic parameter in the mid-term data storage area, w i,m and θ j,m are the corresponding weights, β m It is a coefficient related to time decay. Real-time data is input into the model for prediction, and a detailed performance change curve of the equipment in the next 3 days to 2 weeks is output. The estimated time point when the performance drops to the mid-term model fault judgment threshold is predicted and sent to the operation and maintenance personnel to arrange planned maintenance work, and a mid-term key performance indicator vector K is generated. m (t+ΔT m )=[k m,1 (t+ΔT m ),k m,2 (t+ΔT m ),…,k m,p (t+ΔT m )] T Passed to the long-term model, where k m,q (t+ΔT m ) represents the qth mid-term key performance indicator forecast value; The long-term fault prediction model: The long-term fault prediction model is constructed by combining theory with a data-driven method. The formula is: Among them, y r (t) is the rth long-term characteristic parameter in the long-term data storage area, ω q,l and is the corresponding weight, and the remaining life of the power equipment is predicted through the model, and the results are fed back to the short-term and medium-term models to adjust their fault judgment thresholds and accuracy requirements; The model fusion and decision-making module: based on the prediction results of the short-term fault prediction model, the medium-term fault prediction model and the long-term fault prediction model, assign corresponding dynamic weights to them, use the weighted summation strategy to obtain the health assessment index of the overall condition of the equipment, divide the equipment failure into multiple risk levels according to the health assessment index, and formulate personalized operation and maintenance strategies for different levels, receive new data and feedback in real time, and optimize the parameters and structures of each model according to the new data and operation and maintenance feedback; The user interaction interface: displays equipment health assessment, fault prediction information and operation and maintenance suggestions in intuitive charts and texts, has a historical data query function, and supports access from multiple terminal devices.

2. The power equipment health assessment and fault prediction system according to claim 1, characterized in that: Sensors used in the data acquisition module: Electrical sensors: voltage transformers, current transformers, and power sensors; Mechanical sensors: accelerometers, displacement sensors, and strain gauges; Environmental sensors: temperature sensor, humidity sensor, air pressure sensor.

3. The power equipment health assessment and fault prediction system according to claim 1, characterized in that: The data processing and storage module extracts the feature vector in the power equipment data by using the power data similarity feature extraction formula. Assume that the original data sequence is D={d1, d2, ..., d N }, the extracted feature vector F = [f1,f2,…,f M ], the calculation formula is: Among them, k(d i ,d j ) is the data point d i With d j The similarity kernel function between is as follows: σ is the bandwidth parameter of the kernel function, g(d i -d j ) is the weighted function of the data difference, and the formula is: α is the adjustment parameter, λ is the attenuation coefficient, is the mean of the data sequence D.

4. The power equipment health assessment and fault prediction system according to claim 1, characterized in that: The short-term model fault judgment threshold T in the short-term fault prediction model s,th Determine the normal operation data and obtain A s The mean of (t) and standard deviation The calculation formula of the short-term model fault judgment threshold is: Where n is a multiple determined based on the tolerance for failure risk.

5. The power equipment health assessment and fault prediction system according to claim 1, characterized in that: The mid-term model fault judgment threshold T in the mid-term fault prediction model m,th,i For a specific key performance indicator KPI, i The initial design value of this indicator is KPI i,init , the performance degradation ratio limit is α i , for this key performance indicator KPI i The mid-term model fault judgment threshold T m,th,i The calculation formula is: m,th,i =KPI i,init ×(1-α i ).

6. The power equipment health assessment and fault prediction system according to claim 1, characterized in that: In the long-term fault prediction model, the fault judgment thresholds of the short-term model and the mid-term model are adjusted according to the long-term model: For the adjustment of the short-term model fault judgment threshold, the remaining life ratio of the equipment evaluated by the long-term model is assumed to be R L (t), 0≤R L (t)≤1,R L (t) = 1 means the equipment is brand new, R L (t) = 0 means the equipment life is exhausted, T s,th is the original short-term model fault judgment threshold, α and β are adjustment parameters, and the new short-term model fault judgment threshold formula is: T s,new,th =T s,th ×(1-α×R L (t))+β×(1-R L (t)); For the adjustment of the mid-term model fault judgment threshold, the mid-term model fault judgment threshold is set to T m,th,i The equipment performance degradation rate estimated by the long-term model is D L (t), λ and μ are adjustment parameters, and the new medium-term and short-term model fault judgment threshold calculation formula is: T m,new,th =T m,th,i ×(1-λ×D L (t))+μ×D L (t).

7. The power equipment health assessment and fault prediction system according to claim 1, characterized in that: In the model fusion and decision module, dynamic weights are used to calculate the reliability index of each model. The short-term model reliability index is R s , calculate the prediction accuracy in the recent time period, the number of correct predictions is N sc , the total number of predictions is N st , the timeliness weight w of the recent prediction t , and w t =e -μ·t , where t is the time interval from the current moment, μ is the attenuation coefficient, and the calculation formula is: Medium-term model reliability index R m , let the mid-term prediction performance change curve be C m (t), the actual performance change curve is C ma (t), in the time period [t1, t2], by calculating the similarity S of the two curves m , and combined with the data integrity index I m , the formula is: R m =S m I m ; Long-term model reliability index R l , assuming that the long-term remaining life prediction value is L p , the actual remaining life is L a , the confidence interval width is W l , the formula is: Based on the reliability index, the weights w assigned to the short-term, medium-term, and long-term models are s 、w m 、w l They are: where β s , β m , β l is the adjustment index.

8. The power equipment health assessment and fault prediction system according to claim 1, characterized in that: The health assessment index of the overall condition of the equipment in the model fusion and decision-making module, assuming that the prediction result of the short-term fault prediction model is P s , the prediction result of the mid-term fault prediction model is P m , the prediction result of the long-term fault prediction model is P l , the revolutionary machine cover pairing weight, the short-term model weight is w s , the medium-term model weight is w m , the long-term model weight is w l , the overall health evaluation index H of the computing equipment is calculated as follows: Among them, k s , k m , k l is the adjustment coefficient.

9. The power equipment health assessment and fault prediction system according to claim 1, characterized in that: The risk level division in the model fusion and decision-making module and the formulation of personalized operation and maintenance strategies for different levels, assuming that the overall health assessment index of the equipment is H, define the low risk threshold T low , medium risk threshold T mid , and 0<T lo w<T mid <1: When H>T low When the device is in a low risk state, the monitoring frequency f low Set as basic monitoring frequency f buse , data is collected every hour, simple inspections are conducted every quarter, and resource allocation mainly focuses on the guarantee of daily maintenance resources; When T mid <H≤T low When the device is in a medium risk state, the monitoring frequency is f mid Improve f base ·k mid (k mid >1), maintenance plan is formulated for the next T pm Detailed inspection within the time, T pm It can be determined based on the mid-term model forecast results, and the calculation formula is: ΔT m,th is the time difference between the mid-term performance drop to the threshold, P m (t) is the current medium-term performance forecast, ΔP m is the mid-term performance change rate, and resource allocation increases the reserve ratio of emergency maintenance resources r mid ; When H≤T mid When the device is in a high-risk state, the monitoring frequency f high Significantly increased to f base ·k high , k high >k mid , immediately start the shutdown maintenance plan, maintenance time T repair Calculated based on the fault type and maintenance resources, the formula is: where t ri is the time required for the ith maintenance step, r i It is the resource sufficiency rate for executing the i-th maintenance step. Resource allocation fully guarantees the human, material and technical support required for maintenance.

10. The power equipment health assessment and fault prediction system according to claim 9, characterized in that: The model fusion and the low risk threshold H in the decision module low and medium risk threshold H mid Determination, assuming that the mean H value of healthy equipment in historical data is μ H , with standard deviation σ H , low risk threshold H low Set to: H low =μ H -k1σ H , where k1 is a coefficient determined based on the risk tolerance, and the medium risk threshold H mid Set to: H mid =μ H -k2σ H , where k2>k1.

Images