CN117609716A - Power equipment residual life remote prediction method and system based on state health information - Google Patents
Power equipment residual life remote prediction method and system based on state health information Download PDFInfo
- Publication number
- CN117609716A CN117609716A CN202311424776.1A CN202311424776A CN117609716A CN 117609716 A CN117609716 A CN 117609716A CN 202311424776 A CN202311424776 A CN 202311424776A CN 117609716 A CN117609716 A CN 117609716A
- Authority
- CN
- China
- Prior art keywords
- data
- model
- life
- power equipment
- health information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000036541 health Effects 0.000 title claims abstract description 37
- 238000012544 monitoring process Methods 0.000 claims abstract description 58
- 238000000491 multivariate analysis Methods 0.000 claims abstract description 23
- 238000004140 cleaning Methods 0.000 claims abstract description 18
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000012423 maintenance Methods 0.000 claims description 22
- 239000011159 matrix material Substances 0.000 claims description 11
- 239000013598 vector Substances 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 10
- 238000002790 cross-validation Methods 0.000 claims description 9
- 238000012417 linear regression Methods 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 238000012795 verification Methods 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 230000010354 integration Effects 0.000 claims description 7
- 230000007246 mechanism Effects 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000013459 approach Methods 0.000 claims description 2
- 238000011156 evaluation Methods 0.000 description 3
- 238000009499 grossing Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 238000002360 preparation method Methods 0.000 description 3
- 238000000513 principal component analysis Methods 0.000 description 3
- 230000000875 corresponding effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000013450 outlier detection Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000012847 principal component analysis method Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a method and a system for remotely predicting the residual life of power equipment based on state health information, wherein the method comprises the following steps: step 1: the state health information of the power equipment is collected, and the data are collected in real time through the sensor and the monitoring system and stored in the database, so that the integrity and consistency of the data are ensured; step 2: preprocessing and cleaning the acquired data to ensure the quality and consistency of the data so as to prepare for multivariate analysis; step 3: multivariate analysis and feature selection to determine key parameters related to the life of the electrical equipment; step 4: selecting key parameters to establish a life model of the power equipment so as to predict the residual life of the power equipment; step 5: verifying the accuracy and reliability of the established life model by using historical data; step 6: the established life prediction model is integrated into a monitoring system of the power equipment for remotely monitoring the equipment state. The invention is helpful to accurately capture the relation between different parameters, thereby improving the reliability of the model.
Description
Technical Field
The invention relates to the technical field of equipment life prediction, in particular to a method and a system for remotely predicting the residual life of power equipment based on state health information.
Background
The state health information of the power equipment refers to data and information related to the operation condition and performance of the power equipment, and is used for evaluating whether the operation state of the equipment is normal or not and whether maintenance or repair is needed, and by collecting and analyzing the state health information, operators and maintenance personnel of the power equipment can monitor the operation state of the equipment better, predict potential problems, and take necessary maintenance measures to ensure the reliability, safety and efficiency of the equipment. But multivariate problems are encountered when predicting remaining life using state health information of the electrical equipment: complex interactions between multiple variables and parameters are ignored.
Disclosure of Invention
In order to solve the problems, the invention provides a method and a system for remotely predicting the residual life of power equipment based on state health information.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
on one hand, the invention discloses a power equipment residual life remote prediction method based on state health information, which comprises the following steps:
step 1: the state health information of the power equipment is collected, and the data are collected in real time through the sensor and the monitoring system and stored in the database, so that the integrity and consistency of the data are ensured;
step 2: preprocessing and cleaning the acquired data to ensure the quality and consistency of the data so as to prepare for multivariate analysis;
step 3: multivariate analysis and feature selection to determine key parameters related to the life of the electrical equipment;
step 4: selecting key parameters to establish a life model of the power equipment so as to predict the residual life of the power equipment;
step 5: verifying the accuracy and reliability of the established life model by using historical data;
step 6: the established life prediction model is integrated into a monitoring system of the power equipment for remotely monitoring the equipment state.
Further: the step 2 comprises the following steps:
detecting and processing missing data:
deleting the missing data points, if the data volume is sufficient, deleting the data points containing the missing values;
estimating the missing values using interpolation methods to preserve the data points and maintain the continuity of the time series;
identifying and processing outliers;
synchronizing data sampling frequencies of different sensors;
for different units of measurement values, normalization is performed using the formula:
Z＝(X-μ)/σ
wherein X is an original measured value, mu is a mean value, sigma is a standard deviation, and Z is a normalized value;
noise and burst fluctuations are reduced to smooth the data;
a data quality monitoring mechanism is implemented to periodically check the quality of the data and the effectiveness of the cleaning process to ensure that the quality of the data set is maintained continuously.
Further: the step 3 comprises the following steps:
normalizing the data so that different parameters have the same scale;
calculating covariance matrix of data, element C of covariance matrix ij Representing covariance between parameters i and j, covariance calculationThe formula:
where N is the number of samples, X ik Is the value of the ith parameter of the kth sample,is the mean of the i-th parameter;
performing eigenvalue decomposition on the covariance matrix to find a main component;
the feature values are arranged in descending order, and feature vectors corresponding to the first K largest feature values are selected, wherein K is a new dimension
And projecting the original data into the principal component space by using the first K selected feature vectors to obtain new feature vectors.
Further: the step 4 comprises the following steps:
the basic form of the model is as follows:
Y＝β 0 +β 1 X 1 +β 2 X 2 +...+β n X n +ε
wherein: y is the remaining life of the power equipment, X 1 ,X 2 ,...,X n Is the key parameter of choice, beta 0 ,β 1 ,β 2 ,...,β n Is a coefficient of the model representing the effect of each parameter on lifetime; epsilon is the error term of the model, representing random errors that cannot be interpreted by the model;
using the determined key parameters as input variables (X 1 ,X 2 ,...,X n ) And the known lifetime of the electrical equipment as output variable Y;
fitting a linear model using a linear regression algorithm to find the coefficient β 0 ,β 1 ,β 2 ,...,β n Minimizing the error term epsilon of the model;
analysis of model coefficient beta 0 ,β 1 ,β 2 ,...,β n To understand the effect of each parameter on the life of the electrical device;
and inputting key parameter values of the current power equipment by using the established linear regression model, and predicting the residual life of the current power equipment.
Further: the step 5 comprises the following steps:
dividing the historical data set into a training set and a testing set;
for each cross-validated fold, fitting a model using a training set, and evaluating the performance of the model using a testing set;
for each fold, calculating a performance index of the model;
summarizing the performance indexes of each cross verification, and calculating the average performance and the standard deviation thereof;
after cross-validation, a final power equipment life prediction model is selected.
Further: the step 5 comprises the following steps:
deploying the established life prediction model of the power equipment into a monitoring system;
setting a data transmission channel so as to acquire state health information from the power equipment in real time;
preprocessing and cleaning the transmitted data in a monitoring system to ensure the quality and usability of the data;
inputting the equipment state health information acquired in real time into a deployed life prediction model to obtain instant residual life prediction;
setting an alarm mechanism based on the prediction result so as to generate an alarm when the equipment life prediction is problematic or approaches a maintenance threshold;
a user interface is provided in the monitoring system to allow an operator to remotely monitor the status of the device, view life predictions and alarm information.
In another aspect, the invention discloses a system for remotely predicting the remaining life of an electrical device based on status health information, comprising a monitoring system, the monitoring system comprising:
and the data acquisition and integration module: the state health information of the power equipment is collected, and the data are collected in real time through the sensor and the monitoring system and stored in the database, so that the integrity and consistency of the data are ensured;
and the data preprocessing and cleaning module is used for: preprocessing and cleaning the acquired data to ensure the quality and consistency of the data so as to prepare for multivariate analysis;
multivariate analysis and feature selection module: multivariate analysis and feature selection to determine key parameters related to the life of the electrical equipment;
life model module: selecting key parameters to establish a life model of the power equipment so as to predict the residual life of the power equipment;
model verification and optimization module: verifying the accuracy and reliability of the established life model by using historical data;
remote monitoring prediction module: the established life prediction model is integrated into a monitoring system of the power equipment for remotely monitoring the equipment state.
Compared with the prior art, the invention has the following technical progress:
compared with the prior art, the traditional method usually ignores complex interactions between a plurality of variables and parameters, while the designed method uses multivariate analysis and feature selection to determine which parameters are most critical to life prediction, and helps to capture the relationship between different parameters more accurately, thereby improving the reliability of the model.
By considering a plurality of parameters in the modeling process, the method can more comprehensively analyze the state of the equipment and fully utilize available data, thereby realizing more accurate prediction of the service life of the equipment. This helps to optimize the overall performance of the plant, reducing downtime and maintenance costs.
The method integrates the established life prediction model into a monitoring system of the power equipment, so that an operator can remotely monitor the state of the equipment in real time. This allows problems to be discovered earlier and maintenance measures to be taken more timely, helping to extend the service life of the equipment and reducing maintenance costs. By using historical data and real-time data, the method provides data-driven decision support for operators. Operators can take appropriate maintenance measures according to the suggestions of the model, so that the risk of subjective decision making is reduced. The method has flexibility and can be suitable for different types of power equipment and environments. By retraining the model and adjusting the parameters, the predictive model can be customized to suit the characteristics of different devices.
The method improves the accuracy and reliability of life prediction by comprehensively considering the correlations among a plurality of parameters and variables, and is beneficial to optimizing the operation and maintenance of equipment and reducing the cost. The method is not only helpful for solving the problem of multiple variables, but also integrates the monitoring and predicting system tightly, thereby bringing greater value for equipment management and maintenance.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention.
In the drawings:
FIG. 1 is a flow chart of the present invention.
Detailed Description
The following embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. Embodiments of the present invention will be described below with reference to the accompanying drawings.
Example 1
As shown in fig. 1, the invention discloses a method for remotely predicting the residual life of power equipment based on state health information, which comprises the following steps:
step 1, data acquisition and integration
First, status health information of the electrical equipment needs to be collected, including multiple parameters such as voltage, current, temperature, vibration, humidity, maintenance record, etc. These data can be collected in real time by sensors and monitoring systems and stored in a centralized database to ensure data integrity and consistency.
Step 2, data preprocessing and cleaning
Preprocessing and cleaning the collected data, including processing missing data, removing outliers, synchronizing data of different sampling frequencies, normalizing measured values of different units, etc., to ensure data quality and consistency, in preparation for multivariate analysis.
Step 3, multivariate analysis and feature selection
A principal component analysis method is used to analyze correlations between a plurality of parameters to determine which parameters are most critical to power equipment life predictions. Principal component analysis is used to reduce dimensionality, combining multiple, highly correlated parameters into fewer principal components, thereby alleviating the complexity of the multivariate problem.
Step 4, establishing a life model
Based on the analysis result in step 3, the most critical parameters are selected to establish a power equipment life model, and the model integrates information of a plurality of parameters to predict the residual life of the power equipment.
Step 5, model verification and optimization
Historical data is used to verify the accuracy and reliability of the established life model. The performance of the model can be evaluated by using cross-validation and other techniques, and the model can be optimized according to the validation results to improve the prediction accuracy.
Step 6, remote prediction and monitoring system
And integrating the established life prediction model into a monitoring system of the power equipment. The system can receive the equipment state health information in real time, use the model to predict the residual life, generate an alarm or suggest maintenance measures, enable operators to monitor the equipment state remotely, take measures in time, prolong the service life of the equipment and reduce the maintenance cost.
The method can effectively solve the problem of multiple variables, establishes a life model by using multiple parameters and identifies interaction among key parameters, and also allows remote real-time monitoring of equipment states and early discovery of potential problems, thereby improving reliability and maintenance efficiency of the power equipment.
Specifically, step 1 includes:
at this stage, how to implement data collection and integration of power equipment status health information to solve the problem of multiple variables will be described in detail.
Sensor deployment: various sensors are installed to monitor the status of the electrical equipment. For example, a current sensor, a voltage sensor, a temperature sensor, a vibration sensor, a humidity sensor, etc. Each sensor is responsible for monitoring one or more specific parameters.
And a data acquisition system: the data acquisition system is arranged, so that data can be acquired from the sensor in real time. These systems may employ communication protocols such as Modbus, OPC, etc. to transfer data to a centralized storage and processing location.
Data storage and integration: the collected data is stored in a centralized database to ensure the security and availability of the data. The database may use a relational database (e.g., mySQL, postgreSQL) or a NoSQL database (e.g., mongoDB) to store different types of data.
Data cleaning and calibration: and (3) performing data cleaning and calibration, including detecting and processing missing values, removing abnormal data, performing unit conversion and calibration on data of different sensors, and ensuring the quality and consistency of the data.
Data integration: the data from the different sensors are integrated into a unified data structure for subsequent multivariate analysis. This may require data alignment to ensure data consistency in time and space.
Timestamp and identifier: a time stamp and device identifier are added to the data to track the time and source device of each data point for subsequent data analysis and traceability.
And (3) data quality monitoring: a data quality monitoring mechanism is implemented to detect and address problems in data acquisition or transmission in time. This may include an anomaly detection algorithm to automatically discover data anomalies.
Through the steps, the data collected from a plurality of sensors can be effectively integrated, and a reliable data basis is provided for subsequent multivariate analysis and establishment of a life prediction model. This process also solves part of the multivariate problem because it ensures that the data of different sensors can be analyzed in the same dataset.
Specifically, step 2 includes:
in this step, it will be described in detail how the collected power plant status health information is data pre-processed and cleaned in preparation for multivariate analysis.
Processing missing data: first, missing data is detected and processed. For missing data points, the following method can be used to process:
deletion of missing data points: if the amount of data is sufficient, it may be considered to delete the data points that contain the missing values.
Interpolation filling: interpolation methods (such as linear interpolation or polynomial interpolation) are used to estimate the missing values to preserve the data points and maintain the continuity of the time series.
Outlier detection and processing: outliers are identified and processed to prevent them from adversely affecting subsequent analysis. Outliers can be detected and processed using the following method:
standard deviation method: data points falling outside a few standard deviations are identified based on the mean and standard deviation and are rejected or corrected.
The box diagram method comprises the following steps: outliers are detected and processed using the box plot, and data points that lie outside the box plot are considered outliers.
Replacement value or correction: for a detected outlier, it may be chosen to be replaced with a reasonable value or corrected to a near value to reduce the impact of the outlier.
Data synchronization: if the data sampling frequency of the different sensors is different, it is necessary to synchronize them for subsequent multivariate analysis. Interpolation or resampling methods can be used to synchronize the data to the same point in time.
Data normalization: for different units of measured values, normalization is required to ensure that they have the same dimensions, using the formula:
Z＝(X-μ)/σ
where X is the raw measurement, μ is the mean, σ is the standard deviation, and Z is the normalized value.
Smoothing data: data smoothing techniques, such as moving average or exponential smoothing, may be applied to reduce noise and sudden fluctuations to make the data more suitable for analysis.
And (3) data quality monitoring: a data quality monitoring mechanism is implemented to periodically check the quality of the data and the effectiveness of the cleaning process to ensure that the quality of the data set is maintained continuously.
Through the steps, the collected data can be ensured to have high quality, consistency and analyzability after being preprocessed and cleaned, and a reliable data base is provided for subsequent multivariate analysis. This helps solve the problem of multivariate, ensuring the accuracy of the comparison and correlation analysis between different parameters.
Specifically, step 3 includes:
data normalization: before applying the principal component analysis, it is ensured that the data is normalized so that the different parameters have the same dimensions. The normalization method has been described in step 2.
Covariance matrix calculation: a covariance matrix of the data is calculated. Covariance matrix describes the relationship between different parameters and their variances, element C of covariance matrix ij Representing the covariance between parameters i and j, the covariance calculation formula:
where N is the number of samples, X ik Is the value of the ith parameter of the kth sample,is the mean of the i-th parameter.
Eigenvalue and eigenvector calculation: and carrying out eigenvalue decomposition on the covariance matrix to find a principal component. The eigenvalues represent the contribution of the variance in each principal component direction, and the eigenvectors represent the principal component directions.
Sorting and selecting characteristic values: the feature values are arranged in descending order, and feature vectors corresponding to the first K largest feature values are selected, wherein K is a new dimension. These feature vectors constitute a new feature space, called principal component space.
Projection data: and projecting the original data into the principal component space by using the first K selected feature vectors to obtain new feature vectors. These new feature vectors can be used for subsequent modeling and analysis.
By principal component analysis, information of multiple parameters can be integrated into fewer principal components, reducing dimensions and retaining the most important information. This helps solve the problem of multiple variables while ensuring high interpretability and accuracy of predictions of power plant life.
Specifically, step 4 includes:
the input variables are selected key parameters, the output variables are life of the power equipment, and the basic form of the model is as follows:
Y＝β 0 +β 1 X 1 +β 2 X 2 +...+β n X n +ε
wherein: y is the remaining life (or life prediction) of the power equipment, X 1 ,X 2 ,...,X n Is a key parameter of choice, these parameters being determined in step 3, beta 0 ,β 1 ,β 2 ,...,β n Is a coefficient of the model representing the effect of each parameter on lifetime; epsilon is the error term of the model and represents a random error that cannot be interpreted by the model.
The following implementation steps are as follows:
data preparation: using the key parameters determined in step 3 as input variables (X 1 ,X 2 ,...,X n ) And the known lifetime of the electrical equipment as output variable Y.
Fitting a linear regression model: fitting a linear model using a linear regression algorithm, e.g. least squares, to find the coefficient beta 0 ,β 1 ,β 2 ,...,β n The error term epsilon of the model is minimized.
Model evaluation: the performance of the model is evaluated using suitable evaluation criteria, such as Mean Square Error (MSE), root Mean Square Error (RMSE), decision coefficient (R-squared), etc. These metrics can help determine the fit and prediction accuracy of the model.
Model interpretation: analysis of model coefficient beta 0 ,β 1 ,β 2 ,...,β n To understand the effect of each parameter on the life of the electrical device. The positive and negative coefficients represent the positive and negative effects of the parameter on lifetime, and the magnitude of the coefficient represents the intensity of the effect.
Predicting the service life of the power equipment: and inputting key parameter values of the current power equipment by using the established linear regression model, and predicting the residual life of the current power equipment.
By means of this established linear regression model, the remaining life of the electrical equipment can be estimated from the values of the key parameters. This model integrates information for multiple parameters, helping to solve the problem of multiple variables, while providing a balance of interpretability and predictive performance.
Specifically, step 5 includes:
in this step, the accuracy and reliability of the established power equipment life model is verified using a cross-validation technique, and the model is optimized according to the verification result.
Cross-validation is a commonly used model evaluation method that can evaluate the performance of a model on unseen data and help discover problems with the model and how to improve. The following implementation steps are as follows:
data segmentation: the historical dataset is divided into a training set and a testing set. Typically, the data can be divided into K subsets using K-fold cross-validation, with K-1 for training and the remaining 1 for testing. This process is repeated K times to ensure that each subset is used for testing once.
Model training and verification: for each cross-validated fold, a training set is used to fit the model, and a testing set is used to evaluate the performance of the model. For the linear regression model, the model established in step 4 is used.
Calculating performance indexes: for each fold, performance metrics of the model, such as Mean Square Error (MSE), root Mean Square Error (RMSE), decision coefficient (R-squared), etc., are calculated. These metrics may help understand the behavior of the model on different subsets of data.
Summarizing model performance: and summarizing the performance indexes of each cross verification, and calculating the average performance and the standard deviation thereof. This can help to understand the overall performance and stability of the model.
Model optimization: if the performance of the model is not satisfactory, model optimization may be considered. The optimization method may include adjusting model parameters, trying different feature selection strategies, adding more data, etc. Re-cross-validation is required after each optimization to evaluate improved performance.
Final model selection: after cross-validation and possibly multiple optimizations, a final power plant life prediction model is selected that performs best on historical data.
By using cross-validation, the performance of the model can be more fully understood, avoiding the over-fit or under-fit problems on a single dataset. This helps to ensure generalized performance of the model over unseen data and improves accuracy and reliability of power equipment life predictions.
Specifically, step 6 includes:
in this step, it is explained in detail how the established power equipment lifetime prediction model is integrated into the monitoring system of the power equipment to enable remote prediction and monitoring.
Model deployment: first, the established life prediction model of the power equipment is deployed into a monitoring system. This may be achieved by embedding the model into the software of the monitoring system or by creating an API interface. Ensuring that the model is able to receive and process new state health information in real time.
Data transmission and real-time monitoring: and setting a data transmission channel so as to acquire state health information from the power equipment in real time. This may be achieved by a sensor, monitoring device or industrial internet of things (IIoT) connection. Ensuring that the data can be securely transmitted to the monitoring system.
Data preprocessing: the transmitted data is pre-processed and purged in the monitoring system similar to that in step 2 to ensure the quality and availability of the data. This includes processing missing data, removing outliers, normalizing and data synchronization.
Model real-time prediction: and inputting the equipment state health information acquired in real time into a deployed life prediction model to obtain instant residual life prediction. The model will return a prediction indicating the expected lifetime of the device.
Alarm generation and maintenance advice: based on the prediction result, an alert mechanism is set to generate an alert when the equipment life prediction is problematic or approaching a maintenance threshold. Suggested maintenance measures may also be provided to the operator to extend the equipment life.
Remote monitoring and decision support: a user interface is provided in the monitoring system to allow an operator to remotely monitor the status of the device, view life predictions and alarm information. The operator may take appropriate action, such as maintaining in advance, changing operating strategies, or adjusting equipment loads, based on the recommendations of the model.
Historical data record: for each device, historical state health information, life predictions, and maintenance records are recorded. This helps build maintenance history and optimize predictive models for the device.
Periodic maintenance and updating: and the monitoring system is maintained regularly, so that the accuracy and the performance of the model are ensured. If new data is available, the data can be used to update and retrain the model to accommodate the actual operating conditions of the device.
By integrating the life prediction model of the power equipment into the monitoring system, operators can know the state of the equipment in real time and take measures in time so as to prolong the service life of the equipment and reduce the maintenance cost. Such a remote monitoring and prediction system improves the reliability and efficiency of the power equipment.
Example two
The invention discloses a power equipment residual life remote prediction system based on state health information, which comprises a monitoring system, wherein the monitoring system comprises:
and the data acquisition and integration module: the state health information of the power equipment is collected, and the data are collected in real time through the sensor and the monitoring system and stored in the database, so that the integrity and consistency of the data are ensured;
and the data preprocessing and cleaning module is used for: preprocessing and cleaning the acquired data to ensure the quality and consistency of the data so as to prepare for multivariate analysis;
multivariate analysis and feature selection module: multivariate analysis and feature selection to determine key parameters related to the life of the electrical equipment;
life model module: selecting key parameters to establish a life model of the power equipment so as to predict the residual life of the power equipment;
model verification and optimization module: verifying the accuracy and reliability of the established life model by using historical data;
remote monitoring prediction module: the established life prediction model is integrated into a monitoring system of the power equipment for remotely monitoring the equipment state.
Finally, it should be noted that: the foregoing description is only a preferred embodiment of the present invention, and the present invention is not limited thereto, but it is to be understood that modifications and equivalents of some of the technical features described in the foregoing embodiments may be made by those skilled in the art, although the present invention has been described in detail with reference to the foregoing embodiments. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.
Claims (7)
1. The power equipment residual life remote prediction method based on the state health information is characterized by comprising the following steps of:
step 1: the state health information of the power equipment is collected, and the data are collected in real time through the sensor and the monitoring system and stored in the database, so that the integrity and consistency of the data are ensured;
step 2: preprocessing and cleaning the acquired data to ensure the quality and consistency of the data so as to prepare for multivariate analysis;
step 3: multivariate analysis and feature selection to determine key parameters related to the life of the electrical equipment;
step 4: selecting key parameters to establish a life model of the power equipment so as to predict the residual life of the power equipment;
step 5: verifying the accuracy and reliability of the established life model by using historical data;
step 6: the established life prediction model is integrated into a monitoring system of the power equipment for remotely monitoring the equipment state.
2. The method for remote prediction of remaining life of electrical equipment based on status health information according to claim 1, wherein the step 2 comprises:
detecting and processing missing data:
deleting the missing data points, if the data volume is sufficient, deleting the data points containing the missing values;
estimating the missing values using interpolation methods to preserve the data points and maintain the continuity of the time series;
identifying and processing outliers;
synchronizing data sampling frequencies of different sensors;
for different units of measurement values, normalization is performed using the formula:
Z＝(X-μ)/σ
wherein X is an original measured value, mu is a mean value, sigma is a standard deviation, and Z is a normalized value;
noise and burst fluctuations are reduced to smooth the data;
a data quality monitoring mechanism is implemented to periodically check the quality of the data and the effectiveness of the cleaning process to ensure that the quality of the data set is maintained continuously.
3. The method for remote prediction of remaining life of electrical equipment based on status health information according to claim 2, wherein the step 3 comprises:
normalizing the data so that different parameters have the same scale;
calculating covariance matrix of data, element C of covariance matrix ij Representing the covariance between parameters i and j, the covariance calculation formula:
where N is the number of samples, X ik Is the value of the ith parameter of the kth sample,is the mean of the i-th parameter;
performing eigenvalue decomposition on the covariance matrix to find a main component;
the feature values are arranged in descending order, and feature vectors corresponding to the first K largest feature values are selected, wherein K is a new dimension
And projecting the original data into the principal component space by using the first K selected feature vectors to obtain new feature vectors.
4. The method for remote prediction of remaining life of electrical equipment based on status health information according to claim 3, wherein the step 4 comprises:
the basic form of the model is as follows:
Y＝β 0 +β 1 X 1 +β 2 X 2 +...+β n X n +ε
wherein: y is the remaining life of the power equipment, X 1 ,X 2 ,...,X n Is the key parameter of choice, beta 0 ,β 1 ,β 2 ,...,β n Is a coefficient of the model representing the effect of each parameter on lifetime; epsilon is the error term of the model, representing random errors that cannot be interpreted by the model;
using the determined key parameters as input variables (X 1 ,X 2 ,...,X n ) And the known lifetime of the electrical equipment as output variable Y;
fitting a linear model using a linear regression algorithm to find the coefficient β 0 ,β 1 ,β 2 ,...,β n Minimizing the error term epsilon of the model;
analysis of model coefficient beta 0 ,β 1 ,β 2 ,...,β n To understand the effect of each parameter on the life of the electrical device;
and inputting key parameter values of the current power equipment by using the established linear regression model, and predicting the residual life of the current power equipment.
5. The method for remote prediction of remaining life of electrical equipment based on status health information according to claim 4, wherein the step 5 comprises:
dividing the historical data set into a training set and a testing set;
for each cross-validated fold, fitting a model using a training set, and evaluating the performance of the model using a testing set;
for each fold, calculating a performance index of the model;
summarizing the performance indexes of each cross verification, and calculating the average performance and the standard deviation thereof;
after cross-validation, a final power equipment life prediction model is selected.
6. The method for remote prediction of remaining life of electrical equipment based on status health information according to claim 5, wherein said step 5 comprises:
deploying the established life prediction model of the power equipment into a monitoring system;
setting a data transmission channel so as to acquire state health information from the power equipment in real time;
preprocessing and cleaning the transmitted data in a monitoring system to ensure the quality and usability of the data;
inputting the equipment state health information acquired in real time into a deployed life prediction model to obtain instant residual life prediction;
setting an alarm mechanism based on the prediction result so as to generate an alarm when the equipment life prediction is problematic or approaches a maintenance threshold;
a user interface is provided in the monitoring system to allow an operator to remotely monitor the status of the device, view life predictions and alarm information.
7. The utility model provides a power equipment residual life remote prediction system based on state health information which characterized in that includes monitoring system, monitoring system includes:
and the data acquisition and integration module: the state health information of the power equipment is collected, and the data are collected in real time through the sensor and the monitoring system and stored in the database, so that the integrity and consistency of the data are ensured;
and the data preprocessing and cleaning module is used for: preprocessing and cleaning the acquired data to ensure the quality and consistency of the data so as to prepare for multivariate analysis;
multivariate analysis and feature selection module: multivariate analysis and feature selection to determine key parameters related to the life of the electrical equipment;
life model module: selecting key parameters to establish a life model of the power equipment so as to predict the residual life of the power equipment;
model verification and optimization module: verifying the accuracy and reliability of the established life model by using historical data;
remote monitoring prediction module: the established life prediction model is integrated into a monitoring system of the power equipment for remotely monitoring the equipment state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311424776.1A CN117609716A (en) | 2023-10-30 | 2023-10-30 | Power equipment residual life remote prediction method and system based on state health information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311424776.1A CN117609716A (en) | 2023-10-30 | 2023-10-30 | Power equipment residual life remote prediction method and system based on state health information |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117609716A true CN117609716A (en) | 2024-02-27 |
Family
ID=89950544
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311424776.1A Pending CN117609716A (en) | 2023-10-30 | 2023-10-30 | Power equipment residual life remote prediction method and system based on state health information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117609716A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117761444A (en) * | 2024-02-22 | 2024-03-26 | 深圳普泰电气有限公司 | Method and system for monitoring service life of surge protector |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222290A (en) * | 2020-01-13 | 2020-06-02 | 浙江工业大学 | Large-scale equipment residual service life prediction method based on multi-parameter feature fusion |
| CN114757435A (en) * | 2022-04-29 | 2022-07-15 | 南京国电南自维美德自动化有限公司 | Fan gear box aging prediction method |
| CN116432071A (en) * | 2021-12-31 | 2023-07-14 | 中核武汉核电运行技术股份有限公司 | Rolling bearing residual life prediction method |
| CN116738868A (en) * | 2023-08-16 | 2023-09-12 | 青岛中德智能技术研究院 | Rolling bearing residual life prediction method |
| CN116770930A (en) * | 2023-06-21 | 2023-09-19 | 金松果新材料科技有限公司 | Integrated sewage discharge system for modularized building |
-
2023
- 2023-10-30 CN CN202311424776.1A patent/CN117609716A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222290A (en) * | 2020-01-13 | 2020-06-02 | 浙江工业大学 | Large-scale equipment residual service life prediction method based on multi-parameter feature fusion |
| CN116432071A (en) * | 2021-12-31 | 2023-07-14 | 中核武汉核电运行技术股份有限公司 | Rolling bearing residual life prediction method |
| CN114757435A (en) * | 2022-04-29 | 2022-07-15 | 南京国电南自维美德自动化有限公司 | Fan gear box aging prediction method |
| CN116770930A (en) * | 2023-06-21 | 2023-09-19 | 金松果新材料科技有限公司 | Integrated sewage discharge system for modularized building |
| CN116738868A (en) * | 2023-08-16 | 2023-09-12 | 青岛中德智能技术研究院 | Rolling bearing residual life prediction method |
Non-Patent Citations (1)
| Title |
|---|
| 韦艳芳: ""基于多域退化特征的智能装备部件寿命预测研究"", 《中国优秀硕士学位论文全文数据库 基础科学辑》, 15 February 2021 (2021-02-15), pages 002 - 854 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117761444A (en) * | 2024-02-22 | 2024-03-26 | 深圳普泰电气有限公司 | Method and system for monitoring service life of surge protector |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111222290B (en) | Multi-parameter feature fusion-based method for predicting residual service life of large-scale equipment | |
| Javed et al. | State of the art and taxonomy of prognostics approaches, trends of prognostics applications and open issues towards maturity at different technology readiness levels | |
| Paolanti et al. | Machine learning approach for predictive maintenance in industry 4.0 | |
| US20220237060A1 (en) | Abnormality detection system, abnormality detection method, abnormality detection program, and method for generating learned model | |
| US9933338B2 (en) | Health management system, fault diagnosis system, health management method, and fault diagnosis method | |
| Jung et al. | Vibration analysis for IoT enabled predictive maintenance | |
| JP5530045B1 (en) | Health management system and health management method | |
| KR101948604B1 (en) | Method and device for equipment health monitoring based on sensor clustering | |
| Compare et al. | Predictive maintenance by risk sensitive particle filtering | |
| NO343417B1 (en) | System and method for condition assessment of downhole tools | |
| CN112416643A (en) | Unsupervised anomaly detection method and unsupervised anomaly detection device | |
| CN117609716A (en) | Power equipment residual life remote prediction method and system based on state health information | |
| CN116629627A (en) | Intelligent detection system of power transmission on-line monitoring device | |
| CN117216481A (en) | Remote monitoring method and system for electric appliance | |
| CN113888353A (en) | Energy efficiency diagnosis method, system and medium for distributed photovoltaic power generation equipment | |
| CN116756909A (en) | Early warning diagnosis system of thermal power plant based on data model and mechanism model | |
| Sen et al. | Taming data quality in AI-enabled industrial internet of things | |
| CN117289085A (en) | Multi-line fault analysis and diagnosis method and system | |
| CN117114454B (en) | DC sleeve state evaluation method and system based on Apriori algorithm | |
| Abdelli et al. | Machine learning based data driven diagnostic and prognostic approach for laser reliability enhancement | |
| CN117290685A (en) | Power plant power equipment expert diagnosis system and method based on historical data | |
| CN117435908A (en) | Multi-fault feature extraction method for rotary machine | |
| Xin et al. | Dynamic probabilistic model checking for sensor validation in Industry 4.0 applications | |
| Xin et al. | Run-time probabilistic model checking for failure prediction: A smart lift case study | |
| CN117272844B (en) | Method and system for predicting service life of distribution board |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |