CN117708669A - Multi-source data-based adaptive probability neural network high-voltage cable state evaluation method - Google Patents
Multi-source data-based adaptive probability neural network high-voltage cable state evaluation method Download PDFInfo
- Publication number
- CN117708669A CN117708669A CN202311713746.2A CN202311713746A CN117708669A CN 117708669 A CN117708669 A CN 117708669A CN 202311713746 A CN202311713746 A CN 202311713746A CN 117708669 A CN117708669 A CN 117708669A
- Authority
- CN
- China
- Prior art keywords
- data
- state
- cable
- voltage cable
- neural network
- 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
- 238000011156 evaluation Methods 0.000 title claims abstract description 58
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 36
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 35
- 238000012360 testing method Methods 0.000 claims abstract description 14
- 230000007246 mechanism Effects 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000013480 data collection Methods 0.000 claims abstract description 5
- 238000000034 method Methods 0.000 claims description 38
- 230000008859 change Effects 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 8
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 claims description 6
- 230000005856 abnormality Effects 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 5
- 230000036541 health Effects 0.000 claims description 5
- 238000009413 insulation Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 5
- 230000003068 static effect Effects 0.000 claims description 5
- 230000006978 adaptation Effects 0.000 claims description 4
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 claims description 4
- 238000012806 monitoring device Methods 0.000 claims description 4
- 230000007613 environmental effect Effects 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 3
- 238000002759 z-score normalization Methods 0.000 claims description 3
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 claims description 2
- ZAMOUSCENKQFHK-UHFFFAOYSA-N Chlorine atom Chemical compound [Cl] ZAMOUSCENKQFHK-UHFFFAOYSA-N 0.000 claims description 2
- RWSOTUBLDIXVET-UHFFFAOYSA-N Dihydrogen sulfide Chemical compound S RWSOTUBLDIXVET-UHFFFAOYSA-N 0.000 claims description 2
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 2
- 229910002091 carbon monoxide Inorganic materials 0.000 claims description 2
- 239000000460 chlorine Substances 0.000 claims description 2
- 229910052801 chlorine Inorganic materials 0.000 claims description 2
- 229910000037 hydrogen sulfide Inorganic materials 0.000 claims description 2
- 238000013450 outlier detection Methods 0.000 claims description 2
- 239000001301 oxygen Substances 0.000 claims description 2
- 229910052760 oxygen Inorganic materials 0.000 claims description 2
- 230000001960 triggered effect Effects 0.000 claims description 2
- 238000012795 verification Methods 0.000 claims description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 2
- 238000012423 maintenance Methods 0.000 abstract description 7
- 238000012544 monitoring process Methods 0.000 abstract description 7
- 230000003935 attention Effects 0.000 description 4
- 238000009499 grossing Methods 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 238000013399 early diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000001931 thermography Methods 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The invention provides a multi-source data-based adaptive probability neural network high-voltage cable state evaluation method, which comprises the following steps: selecting cable state evaluation indexes, establishing an evaluation index set and state grading, constructing a sensor network layout and data collection mechanism of multi-source data, and collecting multi-source data related to high-voltage cable state evaluation; carrying out data preprocessing on the collected high-voltage cable multi-source data; constructing an APNN self-adaptive probability neural network, dividing the preprocessed sample data as a training set and a testing set, training the neural network by using historical data and known cable fault cases, and training the APNN; and evaluating the state of the high-voltage cable by using the trained APNN network, and outputting an evaluation value of the related state classification. The cable related data is easy to obtain, has strong realizability, is used for intelligent evaluation of the state of the high-voltage cable, and has great potential in improving the accuracy of cable monitoring and reducing the operation and maintenance cost.
Description
Technical Field
The invention relates to the technical field of high-voltage cable maintenance, in particular to a self-adaptive probability neural network high-voltage cable state evaluation method based on multi-source data.
Background
The reliability of the state of the high-voltage cable, which is a key component of the power system, directly affects the stability and safety of the whole power network. These cables are often buried underground or underwater, making their detection and maintenance particularly complex and difficult. In conventional cable condition assessment methods, such as electrical testing, visual inspection, thermal imaging techniques, etc., which are often intermittent and fail to provide real-time data, these methods often rely on experience and judgment of professionals, may be subject to subjective factors, and have limited effectiveness in early fault prediction.
With the development of the internet of things and big data technology, we can now collect data on high voltage cables in real time from multiple sources, including temperature, current, humidity, environmental factors, etc. The comprehensive analysis of the multi-source data can provide a more comprehensive and accurate view of the cable conditions, thereby facilitating a shift from preventative maintenance to predictive maintenance. In this context, probabilistic neural networks show great potential in dealing with uncertainty and non-linearity problems, are particularly suitable for evaluation of complex system states, can learn patterns of cable faults based on historical and real-time data, and show higher flexibility and accuracy in data-driven decision support systems than conventional algorithms.
Considering the variability of the cable running environment and state, the traditional algorithm has the defect of processing dynamic change, abnormal data and non-stationarity data, and an adaptive assessment method is particularly important. The method can automatically adjust the evaluation model according to the real-time data so as to adapt to new data and conditions, and the self-adaptive method can adjust the evaluation parameters in real time, reduce false alarm and missing report and improve the reliability of cable fault prediction. In recent years, artificial intelligence and machine learning are increasingly widely applied in the power industry, and research shows that the state evaluation precision of power equipment can be remarkably improved by combining multi-source data and advanced algorithms.
In summary, the adaptive probabilistic neural network (Adaptive Probabilistic Neural Network, APNN) based on multi-source data is not only an innovation for evaluating the state of the high-voltage cable, but also an important supplement and improvement to the prior art, and the method has great practical value and wide application prospect in improving the accuracy and efficiency of cable monitoring. Therefore, it is necessary to provide a method for evaluating the state of a high-voltage cable of an adaptive probabilistic neural network based on multi-source data.
Disclosure of Invention
In view of the above, it is necessary to provide a method for evaluating the state of a high-voltage cable based on adaptive probabilistic neural network of multi-source data, which improves the sensitivity to early diagnosis of cable faults, can maintain high-level prediction accuracy in complex and variable operating environments, and solves the related problems of insufficient learning and adaptation ability, evaluation accuracy and efficiency to be further improved when the network faces to variable data characteristics in the existing conventional high-voltage cable state evaluation algorithm.
The invention provides a multi-source data-based self-adaptive probability neural network high-voltage cable state evaluation method, which comprises the following steps of:
step S1: selecting cable state evaluation indexes, establishing an evaluation index set and state grading, constructing a sensor network layout and data collection mechanism of multi-source data, and collecting multi-source data related to high-voltage cable state evaluation;
step S2: carrying out data preprocessing on the collected high-voltage cable multi-source data, carrying out outlier detection and missing value processing, and standardizing related data to obtain preprocessed sample data;
step S3: constructing an APNN self-adaptive probability neural network, dividing the preprocessed sample data as a training set and a testing set, training the neural network by using historical data and known cable fault cases, and automatically updating parameters of the APNN according to real-time data flow in the training process so as to adapt to new data modes and change trends;
step S4: and evaluating the state of the high-voltage cable by using the trained APNN network, and outputting an evaluation value of the related state classification.
Further, the evaluation index set established in the step S1 is divided into a static basic index and a dynamic variable index, where the static basic index includes basic information of the device: voltage class, operational age, historical failure rate, load type; the dynamic indexes comprise operation detection state information, operation state information and tunnel environment information, wherein the operation detection state information comprises insulation resistance, eccentricity, partial discharge and dielectric loss; the running state information comprises cable temperature, temperature rise rate, cable load, circulating current and mechanical stress; the tunnel environment information comprises environment humidity, environment temperature, oxygen concentration, methane concentration, carbon monoxide concentration, hydrogen sulfide concentration, chlorine concentration and water level information; the established state grades are classified, and the evaluation grades are specifically classified into five grades of health, attention, abnormality, fault and serious fault.
Further, the collecting multi-source data related to the high voltage cable status evaluation in the step S1 specifically includes: multiple types of sensors are deployed at key nodes of the cable, including thermocouples, voltage and current sensors, insulation monitoring devices, and environmental monitoring devices, a high frequency sampling strategy is implemented to capture the immediate change in cable status, and data fusion techniques are employed to ensure the consistency in time and space of data collected from different sources.
Further, in the step S2, the data preprocessing is performed on the collected multi-source data of the high voltage cable, which specifically includes:
s21, estimating a missing value by using an interpolation method to finish the processing of missing data, and detecting an abnormal value by using a quarter bit distance method;
s22, normalizing the data by using Z-score normalization to eliminate the differences among influence factor attributes:
where μ is the average and σ is the standard deviation, and the influencing factors include temperature and load.
Further, the step S3 specifically includes:
s31, constructing an APNN adaptive probability neural network suitable for high-voltage cable state evaluation;
s32, dividing the preprocessed sample data into a training set and a testing set respectively, ensuring that a sufficient training sample is used for learning an APNN model, evaluating the performance of the model by using an independent testing set, training a network by using a back propagation algorithm, gradient descent and other technologies to minimize a prediction error, and dynamically adjusting the learning rate in the gradient descent process so as to improve the convergence rate and avoid sinking into a local minimum value; introducing a cross verification method, and performing multiple training and evaluation on the APNN model to verify the generalization capability of the APNN model on different data subsets, so as to ensure the adaptability of the model to diversified data; the method comprises the steps of carrying out a first treatment on the surface of the
S33, network input is monitored in real time, high-frequency sampling of cable state input data is performed, the model is ensured to be capable of timely sensing the change of the cable state, when the dynamic index is detected to be remarkably changed, the self-adaptive adjustment mechanism is triggered, the network automatically adjusts parameters according to the real-time data so as to adapt to new data distribution and characteristics, online learning of the network is achieved, dynamic change of the cable state is continuously adapted, and evolution of the cable state is adapted while evaluation and state prediction are not interrupted.
Furthermore, the network architecture of the APNN adaptive probability neural network is composed of an input layer, a mode layer, a summarizing layer and an output layer, the adaptation of the APNN is realized by dynamically adjusting parameters in the mode layer, and the parameters are updated when the network receives new data so as to better reflect the characteristics of the current data.
Further, the parameters in the mode layer include a mean vector m of the ith training sample i And standard deviation sigma i 。
The beneficial effects brought by adopting the technical scheme are that:
compared with the prior art, reasonable and comprehensive cable state evaluation indexes are selected, an evaluation index set and state grade division are constructed, a sensor network layout and data collection mechanism of multi-source data is built, multi-source data related to high-voltage cable state evaluation is collected, a comprehensive and scientific high-voltage cable state evaluation index set is constructed, so that the cable state evaluation is more comprehensive, the actual running condition of a cable can be accurately reflected, a richer and comprehensive information basis is provided for subsequent state evaluation, potential problems can be found in time, and the reliability of a cable system is improved; the invention utilizes the self-adaptive probability neural network to intelligently evaluate the state of the high-voltage cable, combines the advantages of the probability theory and the neural network, and can efficiently process and analyze a large amount of cable related data compared with the prior art, so that the network can learn and adapt to the changed data characteristics more quickly and accurately when facing the changed data characteristics, thereby improving the evaluation accuracy and efficiency; in practical application, the invention introduces an adaptive mechanism, so that the network can receive new monitoring data in real time, and automatically adjust the structure and parameters thereof to adapt to new data characteristics.
Drawings
FIG. 1 is a schematic flow chart of cable joint temperature prediction provided by the invention;
fig. 2 is a schematic flow chart of step S1 provided in the present invention;
FIG. 3 is a schematic flow chart of step S2 according to the present invention;
fig. 4 is a schematic flow chart of step S3 provided in the present invention;
FIG. 5 is a schematic diagram of an APNN adaptive probabilistic neural network provided by the invention;
Detailed Description
Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form a part hereof, and together with the description serve to explain the principles of the invention, and are not intended to limit the scope of the invention.
Example 1
The embodiment of the invention provides a multi-source data-based adaptive probability neural network high-voltage cable state evaluation method, and in combination with fig. 1, fig. 1 is a schematic flow chart of cable joint temperature prediction provided by the invention, and the method comprises the following steps:
s1, reasonable and comprehensive cable state evaluation indexes are selected, an evaluation index set and state grade division are constructed, a sensor network layout and data collection mechanism of multi-source data is built, multi-source data related to high-voltage cable state evaluation are collected, and as seen in combination with FIG. 2, FIG. 2 is a schematic flow chart of step S1 provided by the invention, and specific steps in step S1 comprise the following steps S11-S13.
S11, constructing an evaluation index set, dividing the evaluation index set into a static basic index and a dynamic variable index, wherein specific state information types and state parameters are shown in a table 1, and the statistics comprises 21 state quantities.
TABLE 1
S12, classifying the state grade of the high-voltage cable, and classifying the evaluation grade into five grades of health, attention, abnormality, fault and serious fault;
s13, deploying various sensors such as thermocouples, voltage and current sensors, insulation monitoring equipment and environment monitoring equipment (such as humidity and temperature sensors) at key nodes of the cable, and acquiring 200 groups of high-voltage cables to comprise 21 state parameter related historical data in S11 to construct a data set as a specific embodiment. A high-frequency sampling strategy is implemented to capture the immediate change of the cable state, ensure the timeliness and accuracy of the data, and a data fusion technology is adopted to ensure the consistency of the data collected from different sources in time and space.
S2, preprocessing data of the collected multi-source data of the high-voltage cable, detecting abnormal values and processing missing values, standardizing related data, and referring to FIG. 3, FIG. 3 is a schematic flow chart of the step S2 provided by the invention, wherein specific steps in the step S2 comprise the following steps S21-S22:
s21, estimating the missing value to finish the processing of the missing data by using an interpolation method, and detecting an abnormal value by using a quarter bit distance (Interquartile Range, IQR) method:
calculate the first quartile Q 1 And a third quartile Q 3 Calculating the quartile range iqr=q 3 -Q 1 Abnormal value judgment: x is x<Q 1 -1.5 XIQR or x>Q 3 +1.5×iqr, i.e., a value in the data less than the lower limit or greater than the upper limit, is regarded as an outlier.
S22, eliminating differences among attributes of influencing factors such as temperature, load and the like by using a prediction model, and normalizing data by using Z-score normalization:
where μ is the mean and σ is the standard deviation.
S3, constructing an APNN self-adaptive probability neural network, dividing the preprocessed sample data as a training set and a testing set, training the neural network by using historical data and known cable fault cases, and automatically updating parameters of the APNN according to real-time data flow in the training process so as to adapt to new data modes and change trends;
preferably, as seen in fig. 4, fig. 4 is a schematic flow chart of step S3 provided in the present invention, and specific steps in step S3 include the following steps S31-S33:
s31, constructing an APNN self-adaptive probability neural network suitable for high-voltage cable state evaluation, and combining with FIG. 5, FIG. 5 is a schematic diagram of the APNN self-adaptive probability neural network provided by the invention, wherein a specific network architecture can be expressed as follows:
1. input layer: an input vector x is received, where x is a feature vector representing various parameters of the system under evaluation, such as temperature, current, etc. of the high voltage cable.
2. Mode layer: each neuron represents a training sample and describes the data by a probability distribution function (typically gaussian distribution) whose output y for the ith mode neuron i Is a function of the input x, typically expressed using a gaussian function:
wherein m is i Is the mean vector, sigma, of the ith training sample i Is the standard deviation.
3. Summarizing layer: the outputs of the pattern layers are classified and summarized, and for each class Cj, the output Sj thereof is the output y of all pattern neurons belonging to the class i And:
4. output layer: and providing a final judgment or classification result based on the output of the summarization layer. If there are K categories, the output of the output layer is a K-dimensional vector, where each element represents the probability that the input x belongs to the corresponding category.
5. An adaptive mechanism: the adaptation of APNN is achieved by dynamically adjusting parameters in the mode layer (e.g., m i Sum sigma i ) To achieve this. The network updates these parameters as new data is received to better reflect the characteristics of the current data.
As a specific embodiment, in the present invention, for the high-voltage cable state evaluation, the constructed APNN network structure is 21×200×5×1, the numbers of neurons of the input layer, the hidden layer, the summation layer and the output layer corresponding to each other are respectively, when the smoothing factor σ is less than 0.10, the APNN performance is better, and in the present invention, the initialized smoothing factor σ=0.05 in the APNN model, for the convenience of network calculation, output state values 1, 2, 3, 4, 5 are respectively set to represent five different states of health, attention, abnormality, failure and serious.
S32, dividing the preprocessed sample data into a training set and a testing set, wherein 80% of the sample data are randomly extracted to be used as the training sample set, and the remaining 20% of the sample data are used as the testing set. In the process of network training, the network is trained by a back propagation algorithm, gradient descent and other technologies so as to minimize the prediction error. Meanwhile, the generalization capability of the model is evaluated by adopting a k-fold cross validation method, a data set is divided into k subsets, the model is trained k times, one subset is used as a test set each time, the rest is used as a training set, and the steps are circulated to enable each subset to be opportunistically used as the test set, so that the stability of the model in the unseen data is ensured.
S33, monitoring network input in real time, triggering a self-adaptive adjustment mechanism when a significant change of a data mode is detected, and automatically adjusting parameters of the network according to real-time data so as to adapt to new data distribution and characteristics. The online learning of the network is realized to continuously adapt to the dynamic change of the cable state.
As a specific embodiment, in the training process of the model, calculating and comparing the difference between the current training error and the last training error, and comparing the current training error and the last training error, if the former is smaller than the latter, increasing the smoothing factor; if the former is larger than the latter and exceeds 1.03 times of the last error, training of the next sample is carried out; otherwise the smoothing factor is reduced.
S4, evaluating the state of the high-voltage cable by using the trained APNN network, and outputting an evaluation value of the related state classification, namely evaluating the state of the high-voltage cable as a corresponding grade of health, attention, abnormality, fault and severity.
In order to evaluate the state evaluation method of the self-adaptive probability neural network high-voltage cable based on multi-source data, as a specific embodiment, the effect of the model provided by the invention on a test data set is counted, the prediction accuracy of the state evaluation result of the high-voltage cable is shown in table 2, the average accuracy of the state evaluation of the APNN model high-voltage cable constructed by the invention can be up to 90.16% as shown in table 2, the network parameters of the model are timely adjusted according to the result error feedback in the training process, so that the state evaluation accuracy is high, and the advanced data processing technology and the self-adaptive machine learning algorithm are combined, so that the cable state can be accurately evaluated and maintenance decision can be timely guided, thereby improving the reliability and safety of the power system, and showing great potential in the aspects of improving the accuracy of cable monitoring and reducing the operation and maintenance cost.
TABLE 2
The present invention is not limited to the above-mentioned embodiments, and any changes or substitutions that can be easily understood by those skilled in the art within the technical scope of the present invention are intended to be included in the scope of the present invention.
Claims (7)
1. The adaptive probability neural network high-voltage cable state evaluation method based on the multi-source data is characterized by comprising the following steps of:
step S1: selecting cable state evaluation indexes, establishing an evaluation index set and state grading, constructing a sensor network layout and data collection mechanism of multi-source data, and collecting multi-source data related to high-voltage cable state evaluation;
step S2: carrying out data preprocessing on the collected high-voltage cable multi-source data, carrying out outlier detection and missing value processing, and standardizing related data to obtain preprocessed sample data;
step S3: constructing an APNN self-adaptive probability neural network, dividing the preprocessed sample data as a training set and a testing set, training the neural network by using historical data and known cable fault cases, and automatically updating parameters of the APNN according to real-time data flow in the training process so as to adapt to new data modes and change trends;
step S4: and evaluating the state of the high-voltage cable by using the trained APNN network, and outputting an evaluation value of the related state classification.
2. The method for evaluating the state of a high-voltage cable of an adaptive probabilistic neural network based on multi-source data according to claim 1, wherein the set of evaluation indexes established in the step S1 is divided into static basic indexes and dynamic variable indexes, wherein the static basic indexes include basic information of equipment: voltage class, operational age, historical failure rate, load type; the dynamic indexes comprise operation detection state information, operation state information and tunnel environment information, wherein the operation detection state information comprises insulation resistance, eccentricity, partial discharge and dielectric loss; the running state information comprises cable temperature, temperature rise rate, cable load, circulating current and mechanical stress; the tunnel environment information comprises environment humidity, environment temperature, oxygen concentration, methane concentration, carbon monoxide concentration, hydrogen sulfide concentration, chlorine concentration and water level information; the established state grades are classified, and the evaluation grades are specifically classified into five grades of health, attention, abnormality, fault and serious fault.
3. The method for evaluating the state of the high-voltage cable of the adaptive probabilistic neural network based on multi-source data according to claim 1, wherein the collecting the multi-source data related to the state evaluation of the high-voltage cable in step S1 specifically comprises: multiple types of sensors are deployed at key nodes of the cable, including thermocouples, voltage and current sensors, insulation monitoring devices, and environmental monitoring devices, a high frequency sampling strategy is implemented to capture the immediate change in cable status, and data fusion techniques are employed to ensure the consistency in time and space of data collected from different sources.
4. The method for evaluating the state of the high-voltage cable of the adaptive probabilistic neural network based on multi-source data according to claim 1, wherein the step S2 of preprocessing the collected multi-source data of the high-voltage cable specifically comprises the following steps:
s21, estimating a missing value by using an interpolation method to finish the processing of missing data, and detecting an abnormal value by using a quarter bit distance method;
s22, normalizing the data by using Z-score normalization to eliminate the differences among influence factor attributes:
where μ is the average and σ is the standard deviation, and the influencing factors include temperature and load.
5. The method for evaluating the state of the high-voltage cable of the adaptive probabilistic neural network based on multi-source data according to claim 1, wherein the step S3 specifically comprises:
s31, constructing an APNN adaptive probability neural network suitable for high-voltage cable state evaluation;
s32, dividing the preprocessed sample data into a training set and a testing set respectively, ensuring that a sufficient training sample is used for learning an APNN model, evaluating the performance of the model by using an independent testing set, training a network by using a back propagation algorithm, gradient descent and other technologies to minimize a prediction error, and dynamically adjusting the learning rate in the gradient descent process so as to improve the convergence rate and avoid sinking into a local minimum value; introducing a cross verification method, and performing multiple training and evaluation on the APNN model to verify the generalization capability of the APNN model on different data subsets, so as to ensure the adaptability of the model to diversified data; the method comprises the steps of carrying out a first treatment on the surface of the
S33, network input is monitored in real time, high-frequency sampling of cable state input data is performed, the model is ensured to be capable of timely sensing the change of the cable state, when the dynamic index is detected to be remarkably changed, the self-adaptive adjustment mechanism is triggered, the network automatically adjusts parameters according to the real-time data so as to adapt to new data distribution and characteristics, online learning of the network is achieved, dynamic change of the cable state is continuously adapted, and evolution of the cable state is adapted while evaluation and state prediction are not interrupted.
6. The method for evaluating the state of a high-voltage cable of an adaptive probabilistic neural network based on multi-source data according to claim 5, wherein the network architecture of the APNN adaptive probabilistic neural network is composed of an input layer, a mode layer, a summary layer and an output layer, the adaptation of the APNN is achieved by dynamically adjusting parameters in the mode layer, and the network updates the parameters when receiving new data so as to better reflect the characteristics of the current data.
7. The method for evaluating the state of a high-voltage cable of an adaptive probabilistic neural network based on multi-source data of claim 6, wherein said parameters in said pattern layer comprise the mean vector m of the ith training sample i And standard deviation sigma i 。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311713746.2A CN117708669A (en) | 2023-12-14 | 2023-12-14 | Multi-source data-based adaptive probability neural network high-voltage cable state evaluation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311713746.2A CN117708669A (en) | 2023-12-14 | 2023-12-14 | Multi-source data-based adaptive probability neural network high-voltage cable state evaluation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117708669A true CN117708669A (en) | 2024-03-15 |
Family
ID=90150923
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311713746.2A Pending CN117708669A (en) | 2023-12-14 | 2023-12-14 | Multi-source data-based adaptive probability neural network high-voltage cable state evaluation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117708669A (en) |
-
2023
- 2023-12-14 CN CN202311713746.2A patent/CN117708669A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Peimankar et al. | Evolutionary multi-objective fault diagnosis of power transformers | |
| Malik et al. | Fuzzy reinforcement learning based intelligent classifier for power transformer faults | |
| CN112131212A (en) | Hybrid cloud scene-oriented time sequence data anomaly prediction method based on ensemble learning technology | |
| Sharma et al. | Review of artificial intelligence techniques application to dissolved gas analysis on power transformer | |
| Erginel | Fuzzy rule-based $\tilde p $ and $ n\tilde p $ control charts | |
| CN111241673A (en) | Health state prediction method for industrial equipment in noisy environment | |
| CN116579768B (en) | Power plant on-line instrument operation and maintenance management method and system | |
| Basuki | Online dissolved gas analysis of power transformers based on decision tree model | |
| Tang et al. | Prediction of bearing performance degradation with bottleneck feature based on LSTM network | |
| CN117193222A (en) | Intelligent quality control system based on industrial Internet of things and big data and control method thereof | |
| CN113782113B (en) | Method for identifying gas fault in transformer oil based on deep residual error network | |
| Jin et al. | State-of-the-art review on asset management methodologies for oil-immersed power transformers | |
| Mao et al. | Comparative study on prediction of fuel cell performance using machine learning approaches | |
| CN117633722B (en) | Detection control method and system based on intelligent detection robot | |
| Esmaeili Nezhad et al. | A review of the applications of machine learning in the condition monitoring of transformers | |
| CN114595883A (en) | Oil-immersed transformer residual life personalized dynamic prediction method based on meta-learning | |
| CN113158448A (en) | Self-adaptive calculation method for equipment threshold of ship system | |
| CN117371207A (en) | Extra-high voltage converter valve state evaluation method, medium and system | |
| Dang et al. | seq2graph: Discovering dynamic non-linear dependencies from multivariate time series | |
| Eo et al. | Deep learning framework with essential pre-processing techniques for improving mixed-gas concentration prediction | |
| CN117708669A (en) | Multi-source data-based adaptive probability neural network high-voltage cable state evaluation method | |
| Jin et al. | Deep Machine Learning-Based Asset Management Approach for Oil-Immersed Power Transformers Using Dissolved Gas Analysis | |
| CN113887571A (en) | Electronic equipment fault prediction method for improving SVR algorithm | |
| Shamsudin et al. | Investigation on online DGA monitoring to determine transformer health index using machine learning | |
| CN113743534B (en) | Transformer oil gas composite imaging identification method based on depth residual error network |
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 |