CN109410089B - Low-voltage trip and customer complaint prediction method, device and storage medium - Google Patents
Low-voltage trip and customer complaint prediction method, device and storage medium Download PDFInfo
- Publication number
- CN109410089B CN109410089B CN201811633676.9A CN201811633676A CN109410089B CN 109410089 B CN109410089 B CN 109410089B CN 201811633676 A CN201811633676 A CN 201811633676A CN 109410089 B CN109410089 B CN 109410089B
- Authority
- CN
- China
- Prior art keywords
- prediction
- data
- sample
- result
- low
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000012549 training Methods 0.000 claims description 76
- 238000012795 verification Methods 0.000 claims description 44
- 238000005070 sampling Methods 0.000 claims description 34
- 238000012952 Resampling Methods 0.000 claims description 20
- 238000012545 processing Methods 0.000 claims description 16
- 238000004140 cleaning Methods 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 10
- 238000010200 validation analysis Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 5
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 230000009466 transformation Effects 0.000 abstract description 9
- 238000010586 diagram Methods 0.000 description 8
- 230000005611 electricity Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- 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
-
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to a low-voltage trip and customer complaint prediction method, a device, computer equipment and a storage medium, which can acquire multi-aspect data by acquiring equipment account data, power supply environment attribute data and station area user characteristic data, can increase the prediction accuracy, wash and arrange the equipment account data, the power supply environment attribute data and the station area user characteristic data to acquire prediction characteristic data, can prevent the prediction model chain from identifying errors after the prediction characteristic data is input into the prediction model chain trained based on a classifier chain model, input the prediction characteristic data into the prediction model chain trained based on the classifier chain model to predict whether the low-voltage trip occurs and whether the customer complaint occurs or not, output the prediction result, adopt the prediction model chain to predict the low-voltage trip and the customer complaint, and can utilize the correlation between the low-voltage trip and the customer complaint of a public transformation station area of a power distribution network, the accuracy rate and the prediction coverage rate of the prediction result are effectively improved.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a low-voltage trip and customer complaint prediction method and apparatus, a computer device, and a storage medium.
Background
With the economic development and the increasing living standard of people, the electricity utilization level of industrial and commercial industries and residents is continuously increased. Especially in the high-temperature weather period in summer, the use of a high-power electric appliance causes the electric load to greatly rise, the phenomena that the distribution and transformation capacity cannot meet the increasing demand of the customer electricity consumption occur, the problems of unstable transformer area voltage, fault tripping and the like occur correspondingly, and the complaint quantity of the power supply type customer caused by the problems is high. At present, the processing methods for the customer complaints of low-voltage tripping and power supply in a public transformer area comprise short-term emergency solving mechanisms and long-term solving means such as switch replacement, user line shunt load adjustment, capacity increase reconstruction, public modification and the like. The treatment method mainly comprises post treatment and lacks of pre-judgment work.
With the continuous improvement of the informatization, automation and interaction levels of the intelligent power distribution network, a large amount of power utilization data are accumulated in power enterprises, and scholars at home and abroad propose a plurality of distribution transformer operation state prediction models based on the big data of the power distribution network. Most of the current researches take distribution transformer overload prediction as an entry point, and on the basis, a high-risk platform area which can have fault tripping and customer complaint problems is divided. However, in addition to the problem of heavy overload of distribution and transformation, a large number of low-voltage trips are caused by operation management factors such as uneven load distribution among branches and distribution areas and unbalanced three phases, and equipment factors such as switchgear and line aging, so that the accuracy of predicting the low-voltage trips and customer complaints of the public transformation area of the power distribution network is low.
Disclosure of Invention
In view of the above, it is necessary to provide a low voltage trip and customer complaint prediction method, apparatus, computer device and storage medium capable of improving the accuracy of predicting the low voltage trip and customer complaint of the public transformation area of the power distribution network.
A low voltage trip and customer complaint prediction method, the method comprising:
acquiring equipment account data, power supply environment attribute data and platform area user characteristic data;
cleaning and sorting the equipment standing book data, the power supply environment attribute data and the transformer area user characteristic data to obtain predicted characteristic data;
and inputting the prediction characteristic data into a prediction model chain trained based on a classifier chain model, predicting whether low-voltage tripping occurs and whether customer complaints occur through the prediction model chain, and obtaining a low-voltage tripping prediction result and a customer complaint prediction result.
In one embodiment, the step of inputting the predicted feature data into a prediction model chain trained based on a classifier chain model, predicting whether low-voltage tripping occurs and whether customer complaints occur through the prediction model chain, and obtaining a low-voltage tripping prediction result and a customer complaint prediction result includes:
respectively inputting the prediction characteristic data into each sub prediction model chain in the prediction model chain, and outputting a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result of each sub prediction model chain;
voting is carried out on the preliminary low-voltage trip prediction results, and the determination of the highest vote number is the low-voltage trip prediction result;
and voting the preliminary customer complaint prediction results, wherein the highest vote number is determined as the customer complaint prediction result.
In one embodiment, the step of inputting the predicted feature data into each sub-prediction model chain in the prediction model chain, and outputting the preliminary low-voltage trip prediction result and the preliminary customer complaint prediction result of each sub-prediction model chain includes:
inputting the prediction characteristic data into a first prediction model of the sub-prediction model chain, and outputting a first prediction result;
inputting the predicted characteristic data and the first prediction result into a second prediction model of the sub-prediction model chain, and outputting a second prediction result;
and determining the second prediction result as a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result output by the sub-prediction model chain.
In one embodiment, the training mode of the first prediction model includes:
obtaining sample data, wherein the sample data comprises: a feature data sample and a first result label for the feature data sample;
sampling each sample data based on mixed resampling to obtain a training sample;
inputting the training sample into a first prediction model to be trained to obtain a trained first prediction model;
obtaining a validation sample, the validation sample comprising: verifying the characteristic data sample;
inputting the verification sample into a trained first prediction model, and outputting a verification result;
and when the verification result meets the requirement, obtaining a first prediction model.
In one embodiment, the training mode of the second prediction model includes:
obtaining sample data, wherein the sample data comprises: a characteristic data sample and a first result label and a second result label of the characteristic data sample;
sampling each sample data based on mixed resampling to obtain a training sample;
inputting the training sample into a second prediction model to be trained to obtain a trained second prediction model;
obtaining a validation sample, the validation sample comprising: verifying the characteristic data sample;
inputting the verification sample into a trained second prediction model, and outputting a verification result;
and when the verification result meets the requirement, obtaining a second prediction model.
In one embodiment, the step of sampling each sample data based on mixed resampling to obtain a training sample includes:
sampling most sample data in the sample data by adopting NCL undersampling to obtain a first training sample of a training sample;
sampling a few types of sample data in the sample data by adopting SMOTE oversampling to obtain a second training sample of the training sample.
In one embodiment, the step of performing sampling processing on a plurality of types of sample data in each sample data by using NCL undersampling to obtain a first training sample of training samples includes:
traversing each sample data to carry out data cleaning to obtain most types of sample data;
carrying out normalization processing on each plurality of types of sample data to obtain each processed sample data;
calculating Euclidean distance between the processed sample data to obtain a distance matrix of the processed sample data;
sequencing elements in an upper triangle of the distance matrix based on the similarity to obtain similarity arrangement of each sample data;
and sequentially and randomly selecting one sample data in every two sample data according to the sequence of the similarity arrangement to obtain a first training sample of the training samples.
A low voltage trip and customer complaint prediction device, the device comprising:
the data acquisition module is used for acquiring equipment ledger data, power supply environment attribute data and station area user characteristic data;
the data processing module is used for cleaning and sorting the equipment standing book data, the power supply environment attribute data and the transformer area user characteristic data to obtain predicted characteristic data;
and the prediction module is used for inputting the prediction characteristic data into a prediction model chain trained based on a classifier chain model, predicting whether low-voltage tripping occurs or not and whether customer complaints occur or not through the prediction model chain, and obtaining a low-voltage tripping prediction result and a customer complaint prediction result.
A computer device comprising a memory storing a computer program and a processor implementing the steps of the method when executing the computer program.
A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method.
According to the low-voltage trip and customer complaint prediction method, the device, the computer equipment and the storage medium, various data can be obtained by obtaining equipment account data, power supply environment attribute data and station area user characteristic data, the prediction accuracy can be improved, the equipment account data, the power supply environment attribute data and the station area user characteristic data are cleaned and sorted to obtain the prediction characteristic data, the prediction characteristic data can be input into a prediction model chain trained based on a classifier chain model to avoid the recognition error of the prediction model chain, the prediction characteristic data are input into the prediction model chain to predict whether the low-voltage trip occurs and whether the customer complaint occurs or not, the prediction result is output, the prediction model chain is adopted to predict the low-voltage trip and the customer complaint, and the correlation between the low-voltage trip and the customer complaint of a public transformation station area of a power distribution network can be utilized, the accuracy rate and the prediction coverage rate of the prediction result are effectively improved.
Drawings
FIG. 1 is a diagram of an example implementation of a low voltage trip and customer complaint prediction method;
FIG. 2 is a schematic flow chart diagram of a low voltage trip and customer complaint prediction method in one embodiment;
FIG. 3 is a schematic flow chart diagram of a low voltage trip and customer complaint prediction method in one embodiment;
FIG. 4 is a schematic flow chart diagram of a low voltage trip and customer complaint prediction method in another embodiment;
FIG. 5 is a schematic flow chart illustrating the training of a first predictive model of the low voltage trip and customer complaint prediction method in one embodiment;
FIG. 6 is a schematic flow chart illustrating the training of a second predictive model of the low voltage trip and customer complaint prediction method in another embodiment;
FIG. 7 is a block diagram of the low voltage trip and customer complaint prediction unit of one embodiment;
FIG. 8 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The low-voltage trip and customer complaint prediction method provided by the application can be applied to the application environment as shown in FIG. 1. Wherein the terminal 102 communicates with the server 104 via a network. The server 104 acquires the equipment account data, the power supply environment attribute data and the platform area user characteristic data through the terminal 102; the server 104 cleans and sorts the equipment standing book data, the power supply environment attribute data and the station area user characteristic data to obtain predicted characteristic data; the server 104 inputs the prediction characteristic data into a prediction model chain trained based on the classifier chain model, and predicts whether low-voltage tripping occurs and whether customer complaints occur through the prediction model chain to obtain a low-voltage tripping prediction result and a customer complaint prediction result. The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the server 104 may be implemented by an independent server or a server cluster formed by a plurality of servers.
In one embodiment, as shown in fig. 2, a low voltage trip and customer complaint prediction method is provided, which is illustrated by applying the method to the server in fig. 1, and includes steps S220 to S260:
step S220, obtaining equipment account data, power supply environment attribute data and platform area user characteristic data.
Wherein, equipment standing book data includes: the transformer capacity can refer to distribution transformer rated capacity (kVA), the number of low-voltage users can refer to the number of low-voltage users in a distribution area, the operation duration can refer to the distribution transformer operation duration (year), and the switch type can refer to the type of trip switch equipment. The power supply environment attribute data includes: the data of temperature, distribution transformation load rate, holiday label and the like, wherein the temperature can be the average temperature in the daytime (DEG C), the distribution transformation load rate can be the highest daily load rate of distribution transformation, and the holiday label can be whether the holiday is legal holiday or not. The station area user characteristic data comprises: the electricity consumption property can refer to residents, industry, business and synthesis, and the regional characteristic can refer to urban areas, towns and villages in cities. The equipment account data refers to the equipment account data of a public transformer area of the power distribution network, the power supply environment attribute data refers to the power supply environment attribute data of the public transformer area of the power distribution network, and the transformer area user characteristic data refers to the transformer area user characteristic data of the public transformer area of the power distribution network.
And step S240, cleaning and sorting the equipment account data, the power supply environment attribute data and the transformer area user characteristic data to obtain predicted characteristic data.
Wherein, clear up equipment standing book data, power supply environment attribute data and platform district user characteristic data's data, if: and rechecking and checking the equipment account data, the power supply environment attribute data and the station area user characteristic data, deleting repeated data information, and correcting existing error data. And (3) performing data arrangement on the cleaned equipment standing book data, the power supply environment attribute data and the station area user characteristic data, if: the data of the capacity, the number of low-voltage users and the commissioning time duration of the transformer can be processed in a box-dividing processing mode to obtain predicted characteristic data, temperature, distribution transformer load rate and electricity utilization property, can be processed in a standardized processing mode to obtain predicted characteristic data, and can be processed in a coding processing mode to obtain predicted characteristic data.
And step S260, inputting the prediction characteristic data into a prediction model chain trained based on a classifier chain model, predicting whether low-voltage tripping occurs and whether customer complaints occur through the prediction model chain, and obtaining a low-voltage tripping prediction result and a customer complaint prediction result.
In order to fully utilize the correlation among labels, the Classifier chain adds the prediction result obtained by each basic Classifier to the characteristic variable space of all the subsequent basic classifiers in the prediction process, provides prediction information for other labels, and forms a chain-shaped Classifier. The low voltage trip prediction results may be: low or no voltage trip occurs; the customer complaint prediction results may be: a customer complaint occurred or a customer complaint did not occur. The low-voltage trip prediction result refers to a predicted low-voltage trip prediction result of a public transformer area of the power distribution network, and the customer complaint prediction result refers to a predicted customer complaint prediction result of a power supply customer of the public transformer area of the power distribution network. Such as: the prediction targets are two related labels of low-voltage tripping of the public transformer area and customer complaints of the public transformer area, the characteristic variable is assumed to be X, whether the low-voltage tripping occurs on the prediction label is y1 and whether the customer service complaint occurs on the prediction label is y2, the prediction result y1 can be obtained by inputting the characteristic variable X into one prediction model in a prediction model chain, the prediction result y1 and the characteristic variable X are input into the other prediction model in the prediction model chain, the prediction result y2 can be obtained, the prediction result y2 can be obtained by inputting the characteristic variable X into one prediction model in the prediction model chain, the prediction result y2 and the characteristic variable X are input into the other prediction model in the prediction model chain, the prediction result y1 can be obtained, and the prediction results of y1 and y2 can be output.
In the low-voltage trip and customer complaint prediction method, various data are acquired by acquiring equipment ledger data, power supply environment attribute data and station area user characteristic data, the accuracy of prediction can be increased, the equipment standing book data, the power supply environment attribute data and the station area user characteristic data are cleaned and sorted to obtain predicted characteristic data, the prediction characteristic data can be input into a prediction model chain trained based on a classifier chain model to avoid the recognition error of the prediction model chain, the prediction characteristic data is input into the prediction model chain trained based on the classifier chain model to predict whether low-voltage tripping occurs and whether each customer complaint occurs, a prediction result is output, the prediction model chain is adopted to predict the low-voltage tripping and the customer complaint, the correlation between the low-voltage tripping and the customer complaints can be utilized, and the accuracy rate of the prediction result and the prediction coverage rate can be effectively improved.
In one embodiment, the step of inputting the prediction feature data into a prediction model chain trained based on a classifier chain model, predicting whether low-voltage tripping occurs and whether customer complaints occur through the prediction model chain, and obtaining a low-voltage tripping prediction result and a customer complaint prediction result comprises the following steps:
respectively inputting the prediction characteristic data into each sub-prediction model chain in the prediction model chain, and outputting a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result of each sub-prediction model chain; voting is carried out on the preliminary low-voltage trip prediction results, and the determination of the highest vote number is the low-voltage trip prediction result; and voting the preliminary customer complaint prediction results, wherein the highest vote number is determined as the customer complaint prediction result.
As shown in fig. 3, the prediction model chain may include a plurality of sub-prediction model chains, the prediction feature data is input into each sub-prediction model chain, each sub-prediction model chain outputs a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result, and the preliminary low-voltage trip prediction result and the preliminary customer complaint prediction result are voted to obtain the prediction result. Such as: supposing that the prediction model chain comprises a sub-prediction model chain 1, a sub-prediction model chain 2, a sub-prediction model chain 3 and a sub-prediction model chain 4, respectively inputting prediction characteristic data into the sub-prediction model chain 1, the sub-prediction model chain 2, the sub-prediction model chain 3 and the sub-prediction model chain 4, wherein the sub-prediction model chain 1 outputs a preliminary low-voltage trip prediction result as the occurrence of low-voltage trip, and the preliminary customer complaint prediction result is the occurrence of customer complaint; the sub-prediction model chain 2 outputs a preliminary low-voltage trip prediction result as that no low-voltage trip occurs, and a preliminary customer complaint prediction result is that customer complaints occur; the sub-prediction model chain 3 outputs a preliminary low-voltage trip prediction result as the occurrence of low-voltage trip, and a preliminary customer complaint prediction result is the occurrence of customer complaints; the sub-prediction model chain 4 outputs a preliminary low-voltage trip prediction result as the occurrence of low-voltage trip, and a preliminary customer complaint prediction result is that no customer complaint occurs; voting is carried out based on the prediction results of the sub prediction model chains, the number of votes with low-voltage tripping is 3, the number of votes without low-voltage tripping is 1, the number of votes with customer complaints is 3, and the number of votes without customer complaints is 1, then the low-voltage tripping prediction result output by the prediction model chain is the occurrence of low-voltage tripping, and the customer complaint prediction result output by the prediction model chain is the occurrence of customer complaints. The prediction model chain is adopted to predict the low-voltage tripping operation and the customer complaints, and the correlation between the low-voltage tripping operation and the customer complaints can be utilized to effectively improve the accuracy rate of the prediction result and the prediction coverage rate.
In one embodiment, the step of inputting the prediction feature data into each sub-prediction model chain in the prediction model chain, and outputting the preliminary low-voltage trip prediction result and the preliminary customer complaint prediction result of each sub-prediction model chain includes:
inputting the predicted characteristic data into a first prediction model of the sub-prediction model chain, and outputting a first prediction result; inputting the predicted characteristic data and the first prediction result into a second prediction model of the sub-prediction model chain, and outputting a second prediction result; and determining the second prediction result as a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result output by the sub-prediction model chain.
Wherein, the first prediction model in each sub prediction model chain can be a model for predicting the preliminary low-voltage trip prediction result and can also be a model for predicting the preliminary customer complaint prediction result, when the first predictive model in the chain of sub-predictive models is the model used to predict the preliminary low voltage trip prediction, the second prediction model is used for predicting the prediction result of the preliminary customer complaint, the prediction characteristic data is input into the first prediction model, the first prediction model predicts the prediction result of the preliminary low-voltage trip, the prediction result of the preliminary low-voltage trip (namely the first prediction result) is output, the prediction result of the preliminary low-voltage trip and the prediction characteristic data are input into the second prediction model, the second prediction model predicts the prediction result of the preliminary customer complaint based on the prediction result of the preliminary low-voltage trip and the prediction characteristic data, and outputs a preliminary low voltage trip prediction result and a preliminary customer complaint prediction result (i.e., a second prediction result).
When the first prediction model in the sub-prediction model chain is used for predicting the preliminary customer complaint prediction result, the second prediction model is used for predicting the preliminary low-voltage trip prediction result, the prediction characteristic data is input into the first prediction model, the first prediction model predicts the preliminary customer complaint prediction result, the preliminary customer complaint prediction result (namely the first prediction result) is output, the preliminary customer complaint prediction result and the prediction characteristic data are input into the second prediction model, the second prediction model predicts the preliminary low-voltage trip prediction result on the basis of the preliminary customer complaint prediction result and the prediction characteristic data, and the preliminary low-voltage trip prediction result and the preliminary customer complaint prediction result (namely the second prediction result) are output. The prediction model chain is adopted to predict the low-voltage tripping operation and the customer complaints, and the correlation between the low-voltage tripping operation and the customer complaints can be utilized to effectively improve the accuracy rate of the prediction result and the prediction coverage rate.
In one embodiment, as shown in fig. 4, the predicted feature data is input into the first prediction model of each sub-prediction model chain, and each first prediction result is output; inputting the predicted characteristic data and each first prediction result into a second prediction model corresponding to each sub-prediction model chain, and outputting each second prediction result; and correspondingly determining each second prediction result as a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result output by each sub-prediction model chain, and voting each preliminary low-voltage trip prediction result and each preliminary customer complaint prediction result to obtain a prediction result. The prediction model chain is adopted to predict the low-voltage tripping operation and the customer complaints, and the correlation between the low-voltage tripping operation and the customer complaints can be utilized to effectively improve the accuracy rate of the prediction result and the prediction coverage rate.
Assuming that the prediction model chain comprises a sub-prediction model chain 1 and a sub-prediction model chain 2, wherein a first prediction model in the sub-prediction model chain 1 is a model for predicting a preliminary low-voltage trip prediction result, a first prediction model in the sub-prediction model chain 2 is a model for predicting a preliminary customer complaint prediction result, prediction characteristic data is input into the first prediction model in the sub-prediction model chain 1, the output preliminary low-voltage trip prediction result is that low-voltage trip occurs, the occurrence low-voltage trip and the prediction characteristic data are input into a second prediction model in the sub-prediction model chain 1, and the output preliminary low-voltage trip prediction result and the preliminary customer complaint prediction result are as follows: low voltage tripping occurs and customer complaints occur; inputting the predicted characteristic data into a first prediction model in the sub-prediction model chain 2, outputting a preliminary customer complaint prediction result as a customer complaint, inputting the customer complaint and the predicted characteristic data into a second prediction model in the sub-prediction model chain 2, and outputting a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result as follows: when low-voltage tripping occurs and customer complaints occur, voting is carried out based on the prediction results of the sub-prediction model chain 1 and the sub-prediction model chain 2, the number of votes with low-voltage tripping is 2 votes, the number of votes without low-voltage tripping is 0 vote, the number of votes with customer complaints is 2 votes, and the number of votes without customer complaints is 0 vote, then the low-voltage tripping prediction result and the customer complaint prediction result output by the prediction model chain are as follows: low voltage trips occur and customer complaints occur.
In one embodiment, the first prediction model is trained by: obtaining each sample data, the sample data comprising: a first result label of the characteristic data sample and the characteristic data sample; sampling each sample data based on mixed resampling to obtain a training sample; inputting a training sample into a first prediction model to be trained to obtain the trained first prediction model; obtaining a verification sample, the verification sample comprising: verifying the characteristic data sample; inputting the verification sample into the trained first prediction model, and outputting a verification result; and when the verification result meets the requirement, obtaining a first prediction model.
When the first prediction model is trained, the characteristic data samples in the sample data are used as characteristic variables, and the first result labels of the characteristic data samples in the sample data are used as prediction labels. The characteristic data samples comprise a large amount of equipment standing book data, a large amount of power supply environment attribute data and a large amount of platform area user characteristic data, the first result labels of the characteristic data samples refer to result labels set according to first results corresponding to the characteristic data samples, as shown in fig. 5, sampling processing is carried out on the sample data based on mixed resampling to obtain training samples, the training samples are input into a first prediction model to be trained, and the first prediction model to be trained is trained based on the characteristic data samples in the training samples and the first result labels of the characteristic data samples to obtain the trained first prediction model; obtaining a verification sample, the verification sample comprising: verifying the characteristic data sample; inputting the verification sample into the trained first prediction model, and outputting a verification result; and when the verification result meets the requirement, obtaining a first prediction model.
And when the first result label of the characteristic data sample is the result label set according to the low-voltage trip result (namely the first result) corresponding to each characteristic data sample, obtaining the first prediction model as the model for predicting the preliminary low-voltage trip prediction result, and when the first result label of the characteristic data sample is the result label set according to the customer complaint result (namely the first result) corresponding to each characteristic data sample, obtaining the first prediction model as the model for predicting the preliminary customer complaint prediction result. Sampling processing is carried out through mixed resampling to obtain training samples, and the problems of important information loss and information redundancy are avoided.
In one embodiment, the training of the second prediction model includes: obtaining each sample data, the sample data comprising: the characteristic data sample and a first result label and a second result label of the characteristic data sample; sampling each sample data based on mixed resampling to obtain a training sample; inputting the training sample into a second prediction model to be trained to obtain a trained second prediction model; obtaining a verification sample, the verification sample comprising: verifying the characteristic data sample; inputting the verification sample into the trained second prediction model, and outputting a verification result; and when the verification result meets the requirement, obtaining a second prediction model.
When the second prediction model is trained, the feature data samples in the sample data and the first result labels of the feature data samples are used as feature variables, and the second result labels of the feature data samples in the sample data are used as prediction labels. The characteristic data samples comprise a large amount of equipment account data, a large amount of power supply environment attribute data and a large amount of platform area user characteristic data, a first result label of each characteristic data sample refers to a result label set according to a first result corresponding to each characteristic data sample, a second result label of each characteristic data sample refers to a result label set according to each characteristic data sample and a second result corresponding to the first result of each characteristic data sample, the first result can refer to a customer complaint result or a low-voltage trip result, when the first result is the customer complaint result, the second result is the low-voltage trip result, and when the first result is the low-voltage trip result, the second result is the customer complaint result,
as shown in fig. 6, sampling each sample data based on mixed resampling to obtain a training sample, inputting the training sample into a second prediction model to be trained, and training the second prediction model to be trained based on a feature data sample in each training sample, a first result label of the feature data sample, and a second result label of the feature data sample to obtain a trained second prediction model; obtaining a verification sample, the verification sample comprising: verifying the characteristic data sample; inputting the verification sample into the trained second prediction model, and outputting a verification result; and when the verification result meets the requirement, obtaining a second prediction model.
And when the first result label of the characteristic data sample is the result label of the customer complaint result, obtaining a second prediction model which is the model for predicting the preliminary low-voltage trip prediction result. The mixed resampling method based on the combination of NCL undersampling and SMOTE oversampling can effectively reduce the influence of two problems, namely, a large amount of added synthesized samples of an oversampling algorithm, possibly causing information redundancy of a few types of samples, and easy loss of part of important information in a plurality of types of samples of the undersampling algorithm on prediction accuracy and prediction coverage.
Wherein, before mixed resampling, the target proportion of the training samples of the minority class and the majority class in the training sample set formed by resampling needs to be confirmed, and the number of over-sampling and under-sampling addition and removal is calculated accordingly. Assuming that the number of training samples in the minority class and the number of training samples in the majority class in an unbalanced data set are n1 and n0, the target ratio of the training samples in the minority class to the training samples in the majority class formed by resampling is k 1: k0, mixed resampling with a combination of undersampling and SMOTE oversampling, the number of training samples N that need to be added and removed for oversampling and undersampling can be calculated from equation 1:
N＝round[(k1·n0)-(k0·n1)](1)
in formula 1, round means rounding the calculation result by rounding; wherein k1+ k0 is 1, so the mixed resampling target ratio can be controlled only by setting the parameter k 1; to ensure that the number of majority training samples in the training sample set is not less than the minority training samples and that the N value is not negative, the upper limit value of k1 is set to 0.5, and the lower limit value is the minority proportion in the sample set: n1/(n1+ n 0).
In one embodiment, the step of performing sampling processing on each sample data based on mixed resampling to obtain a training sample includes: sampling most sample data in each sample data by adopting NCL undersampling to obtain a first training sample of a training sample; sampling a few types of sample data in each sample data by adopting SMOTE oversampling to obtain a second training sample of the training samples.
The NCL undersampling means that all the few types of sample data are retained and the majority of sample data existing in the neighborhood of the sample data are cleaned. Basic NCL undersampling is divided into two steps: step one, traversing each sample in a sample data set, finding out three nearest samples, if any sample x belongs to a majority class and at least two of the three nearest samples are a minority class, determining x as noise data, and cleaning the noise data; and step two, if the sample x belongs to the minority class and at least two of the three nearest samples are the majority class, cleaning the majority class samples in the adjacent samples.
SMOTE oversampling is an oversampling method proposed by Chawla et al to solve the unbalanced data problem, and its main idea is to synthesize a new sample between two adjacent samples by a random linear interpolation method, thereby obtaining a specified number of samples. In the present application, the SMOTE algorithm is used to oversample a few classes of samples in the sample data. Synthesis of sample XnewThe calculation method is as formula 2:
in equation 2: rand (0, 1) represents a random number for the interval (0, 1), x is any few class samples,
is a random one of the k nearest neighbors of x. In this application, k is set to 5 by default, and the number of oversamples is N.
In one embodiment, the step of sampling a plurality of types of sample data in each sample data by NCL undersampling to obtain a first training sample of the training samples includes: traversing each sample data to carry out data cleaning to obtain most types of sample data; carrying out normalization processing on each plurality of types of sample data to obtain each processed sample data; calculating Euclidean distances among the processed sample data to obtain a distance matrix of the processed sample data; sequencing elements in an upper triangle of the distance matrix based on the similarity to obtain similarity arrangement of each sample data; and sequentially and randomly selecting one sample data in every two sample data according to the sequence of the similarity arrangement to obtain a first training sample of the training samples.
The step of traversing each sample data to perform data cleaning and obtaining a plurality of types of sample data comprises the following steps: step one, traversing each sample in a sample data set, finding out three nearest samples, if any sample x belongs to a majority class and at least two of the three nearest samples are a minority class, determining x as noise data, and cleaning the noise data; and step two, if the sample x belongs to the minority class and at least two of the three nearest samples are the majority class, cleaning the majority class samples in the nearest samples, and determining the remaining majority class samples as the majority class samples. There may be more than one element in the upper triangle, and each element in the upper triangle represents a distance calculation between two samples, which is expressed as a similarity.
Most types of sample data obtained by sampling based on the basic NCL undersampling are found to be few in the number of most types cleaned by the NCL, and the classification performance of the few types cannot be obviously improved after the undersampling. Therefore, on the basis of traversing each sample data to perform data cleaning and obtain a plurality of types of sample data, normalization processing is performed on the plurality of types of sample data to obtain each processed sample data, and the Euclidean distance between each pair of samples in the plurality of types of sample data is calculated to obtain the distance matrix of each processed sample data. Wherein the distance between the p-th and q-th samples is denoted as DIp，q，DIp，qSmaller indicates higher similarity between sample p and sample q. Then, the similarity of every two samples is sequenced from high to low according to the upper triangular element of the DI (distance matrix), the similarity arrangement of each sample data is obtained, one of every pair of samples corresponding to the similarity is sequentially and randomly selected according to the arrangement sequence to be cleaned until the sum of the number of most types of samples cleaned in the previous two steps reaches N, the NCL undersampling is stopped, and the rest samples are determined as the first training samples of the training samples. The problem that the number of most classes cleaned by NCL is less and the classification performance of few classes cannot be obviously improved after undersampling is solved.
It should be understood that, although the steps in the flowchart of fig. 2 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a portion of the steps in fig. 2 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
In one embodiment, as shown in fig. 7, there is provided a low voltage trip and customer complaint prediction device comprising: a data acquisition module 310, a data processing module 320, and a prediction module 330, wherein:
the data acquisition module 310 is configured to acquire device ledger data, power supply environment attribute data, and platform area user feature data;
the data processing module 320 is configured to clean and sort the device ledger data, the power supply environment attribute data, and the platform area user feature data to obtain predicted feature data;
and the prediction module 330 is configured to input the prediction feature data into a prediction model chain trained based on a classifier chain model, and predict whether low-voltage tripping occurs and whether customer complaints occur through the prediction model chain to obtain a low-voltage tripping prediction result and a customer complaint prediction result.
In one embodiment, the prediction module 330 further comprises: the preliminary prediction unit is used for respectively inputting the prediction characteristic data into each sub-prediction model chain in the prediction model chain and outputting a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result of each sub-prediction model chain; the low-voltage tripping voting unit is used for voting the preliminary low-voltage tripping prediction results, and the determination of the highest vote number is the low-voltage tripping prediction result; and the customer complaint voting unit is used for voting the preliminary customer complaint prediction results, and the highest vote number is determined as the customer complaint prediction result.
In one embodiment, the preliminary prediction unit is further to: inputting the predicted characteristic data into a first prediction model of the sub-prediction model chain, and outputting a first prediction result; inputting the predicted characteristic data and the first prediction result into a second prediction model of the sub-prediction model chain, and outputting a second prediction result; and determining the second prediction result as a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result output by the sub-prediction model chain.
In one embodiment, the low voltage trip and customer complaint prediction device further comprises: the sample data acquisition module is used for acquiring each sample data, and the sample data comprises: a first result label of the characteristic data sample and the characteristic data sample; the sample sampling module is used for sampling each sample data based on mixed resampling to obtain a training sample; the model training module is used for inputting a training sample into a first prediction model to be trained to obtain the trained first prediction model; a verification module for obtaining a verification sample, the verification sample comprising: and (3) verifying the sample by using the characteristic data, inputting the verification sample into the trained first prediction model, outputting a verification result, and obtaining the first prediction model when the verification result meets the requirement.
In one embodiment, the sample data obtaining module is further configured to: obtaining each sample data, the sample data comprising: the characteristic data sample and a first result label and a second result label of the characteristic data sample; the sample sampling module is further configured to: sampling each sample data based on mixed resampling to obtain a training sample; the model training module is further to: inputting the training sample into a second prediction model to be trained to obtain a trained second prediction model; the verification module is further to: obtaining a verification sample, the verification sample comprising: verifying the characteristic data sample; inputting the verification sample into the trained second prediction model, and outputting a verification result; and when the verification result meets the requirement, obtaining a second prediction model.
In one embodiment, the sample sampling module comprises: the first sample sampling unit is used for sampling most sample data in each sample data by adopting NCL undersampling to obtain a first training sample of the training sample; and the second sample sampling unit is used for sampling a few types of sample data in each sample data by SMOTE oversampling to obtain a second training sample of the training sample.
In one embodiment, the first sample sampling unit is further configured to: traversing each sample data to carry out data cleaning to obtain most types of sample data; carrying out normalization processing on each plurality of types of sample data to obtain each processed sample data; calculating Euclidean distances among the processed sample data to obtain a distance matrix of the processed sample data; sequencing elements in an upper triangle of the distance matrix based on the similarity to obtain similarity arrangement of each sample data; and sequentially and randomly selecting one sample data in every two sample data according to the sequence of the similarity arrangement to obtain a first training sample of the training samples.
Specific limitations regarding the low voltage trip and customer complaint prediction means can be found in the limitations regarding the low voltage trip and customer complaint prediction method described above and will not be described in detail herein. The various modules in the low voltage trip and customer complaint prediction devices described above may be implemented in whole or in part by software, hardware, and combinations thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 8. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used to store sample data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a low voltage trip and customer complaint prediction method.
Those skilled in the art will appreciate that the architecture shown in fig. 8 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
A computer apparatus comprising a memory storing a computer program and a processor implementing the steps of a low voltage trip and customer complaint prediction method when the computer program is executed.
A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of a low voltage trip and customer complaint prediction method.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A low voltage trip and customer complaint prediction method, the method comprising:
acquiring equipment account data, power supply environment attribute data and platform area user characteristic data;
cleaning and sorting the equipment standing book data, the power supply environment attribute data and the transformer area user characteristic data to obtain predicted characteristic data;
inputting the prediction characteristic data into a prediction model chain trained based on a classifier chain model, predicting whether low-voltage tripping occurs and whether customer complaints occur through the prediction model chain, and obtaining a low-voltage tripping prediction result and a customer complaint prediction result;
the step of inputting the prediction characteristic data into a prediction model chain trained based on a classifier chain model, predicting whether low-voltage tripping occurs and whether customer complaints occur through the prediction model chain, and obtaining a low-voltage tripping prediction result and a customer complaint prediction result comprises the following steps:
respectively inputting the prediction characteristic data into a first prediction model of each sub-prediction model chain in the prediction model chain, and outputting a first prediction result; inputting the predicted characteristic data and the first prediction result into a second prediction model of the sub-prediction model chain, and outputting a second prediction result; determining the second prediction result as a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result output by the sub-prediction model chain, wherein the first prediction result is the preliminary low-voltage trip prediction result or the preliminary customer complaint prediction result;
voting the preliminary customer complaint prediction results, determining the highest vote number as the customer complaint prediction result, wherein the customer complaint prediction result refers to the predicted customer complaint prediction result of the power supply type customer in the public transformer area of the power distribution network;
voting is carried out on each preliminary low-voltage trip prediction result, the highest vote number is determined as a low-voltage trip prediction result, and the low-voltage trip prediction result refers to a predicted low-voltage trip prediction result of a public transformer area of the power distribution network.
2. The method of claim 1, wherein the first predictive model is trained by:
obtaining sample data, wherein the sample data comprises: a feature data sample and a first result label for the feature data sample;
sampling each sample data based on mixed resampling to obtain a training sample;
inputting the training sample into a first prediction model to be trained to obtain a trained first prediction model;
obtaining a validation sample, the validation sample comprising: verifying the characteristic data sample;
inputting the verification sample into a trained first prediction model, and outputting a verification result;
and when the verification result meets the requirement, obtaining a first prediction model.
3. The method of claim 1, wherein the second predictive model is trained by:
obtaining sample data, wherein the sample data comprises: a characteristic data sample and a first result label and a second result label of the characteristic data sample;
sampling each sample data based on mixed resampling to obtain a training sample;
inputting the training sample into a second prediction model to be trained to obtain a trained second prediction model;
obtaining a validation sample, the validation sample comprising: verifying the characteristic data sample;
inputting the verification sample into a trained second prediction model, and outputting a verification result;
and when the verification result meets the requirement, obtaining a second prediction model.
4. The method according to any one of claims 2 or 3, wherein the step of sampling each sample data based on mixed resampling to obtain training samples comprises:
sampling most sample data in the sample data by adopting NCL undersampling to obtain a first training sample of a training sample;
sampling a few types of sample data in the sample data by adopting SMOTE oversampling to obtain a second training sample of the training sample.
5. The method of claim 4, wherein the step of sampling a majority of sample data in each of the sample data using NCL undersampling to obtain a first training sample of training samples comprises:
traversing each sample data to carry out data cleaning to obtain most types of sample data;
carrying out normalization processing on each plurality of types of sample data to obtain each processed sample data;
calculating Euclidean distance between the processed sample data to obtain a distance matrix of the processed sample data;
sequencing elements in an upper triangle of the distance matrix based on the similarity to obtain similarity arrangement of each sample data;
and sequentially and randomly selecting one sample data in every two sample data according to the sequence of the similarity arrangement to obtain a first training sample of the training samples.
6. The method according to claim 2, wherein when the first result label of the characteristic data sample is a result label set according to the low voltage trip result corresponding to each characteristic data sample, the first prediction model is a model for predicting a preliminary low voltage trip prediction result;
and when the first result label of the characteristic data sample is a result label set according to the customer complaint result corresponding to each characteristic data sample, the first prediction model is a model for predicting a preliminary customer complaint prediction result.
7. The method of claim 3, wherein the second predictive model is a model that predicts preliminary customer complaint prediction results when the first outcome label of the signature data sample is an outcome label for low voltage trip outcomes, and wherein the second predictive model is a model that predicts preliminary low voltage trip prediction results when the first outcome label of the signature data sample is an outcome label for customer complaint outcomes.
8. A low voltage trip and customer complaint prediction device, characterized in that it comprises:
the data acquisition module is used for acquiring equipment ledger data, power supply environment attribute data and station area user characteristic data;
the data processing module is used for cleaning and sorting the equipment standing book data, the power supply environment attribute data and the transformer area user characteristic data to obtain predicted characteristic data;
the prediction module is used for inputting the prediction characteristic data into a prediction model chain trained based on a classifier chain model, predicting whether low-voltage tripping occurs or not and whether customer complaints occur or not through the prediction model chain, and obtaining a low-voltage tripping prediction result and a customer complaint prediction result;
the prediction module comprises: the preliminary prediction unit is used for respectively inputting the prediction characteristic data into each sub-prediction model chain in the prediction model chain and outputting a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result of each sub-prediction model chain; the low-voltage tripping voting unit is used for voting the preliminary low-voltage tripping prediction results, and the determination of the highest vote number is the low-voltage tripping prediction result; the customer complaint voting unit is used for voting the preliminary customer complaint prediction results, and the highest vote number is determined as the customer complaint prediction result;
the preliminary prediction unit is further to: inputting the predicted characteristic data into a first prediction model of the sub-prediction model chain, and outputting a first prediction result; inputting the predicted characteristic data and the first prediction result into a second prediction model of the sub-prediction model chain, and outputting a second prediction result; and determining the second prediction result as a preliminary low-voltage trip prediction result and a preliminary customer complaint prediction result output by the sub-prediction model chain, wherein the first prediction result is the preliminary low-voltage trip prediction result or the preliminary customer complaint prediction result.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811633676.9A CN109410089B (en) | 2018-12-29 | 2018-12-29 | Low-voltage trip and customer complaint prediction method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811633676.9A CN109410089B (en) | 2018-12-29 | 2018-12-29 | Low-voltage trip and customer complaint prediction method, device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109410089A CN109410089A (en) | 2019-03-01 |
| CN109410089B true CN109410089B (en) | 2020-11-03 |
Family
ID=65462757
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811633676.9A Active CN109410089B (en) | 2018-12-29 | 2018-12-29 | Low-voltage trip and customer complaint prediction method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109410089B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135614A (en) * | 2019-03-26 | 2019-08-16 | 广东工业大学 | It is a kind of to be tripped prediction technique based on rejecting outliers and the 10kV distribution low-voltage of sampling techniques |
| CN110033140A (en) * | 2019-04-22 | 2019-07-19 | 广东工业大学 | A kind of distribute-electricity transformer district tripping prediction technique, system and device |
| CN111382897A (en) * | 2019-10-25 | 2020-07-07 | 广州供电局有限公司 | Transformer area low-voltage trip prediction method and device, computer equipment and storage medium |
| CN111178957B (en) * | 2019-12-23 | 2023-04-14 | 广西电网有限责任公司 | Method for early warning sudden increase of electric quantity of electricity consumption customer |
| CN115879586B (en) * | 2022-01-11 | 2024-01-02 | 北京中关村科金技术有限公司 | Complaint prediction optimization method and device based on ablation experiment and storage medium |
| CN115456210B (en) * | 2022-08-22 | 2024-04-12 | 国网浙江省电力有限公司杭州市临安区供电公司 | Power consumption complaint early warning method based on cascading logistic regression Bayesian algorithm |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104376510A (en) * | 2014-12-05 | 2015-02-25 | 国家电网公司 | Method of predicting and accessing level of wildfire-caused trip risk in power transmission lines |
| CN106980929A (en) * | 2017-01-05 | 2017-07-25 | 国网福建省电力有限公司 | A kind of power failure complaint risk Forecasting Methodology based on random forest |
| CN107506849A (en) * | 2017-07-24 | 2017-12-22 | 国网江西省电力公司电力科学研究院 | A kind of intelligent optimization distribution transforming, which has a power failure, studies and judges system |
-
2018
- 2018-12-29 CN CN201811633676.9A patent/CN109410089B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104376510A (en) * | 2014-12-05 | 2015-02-25 | 国家电网公司 | Method of predicting and accessing level of wildfire-caused trip risk in power transmission lines |
| CN106980929A (en) * | 2017-01-05 | 2017-07-25 | 国网福建省电力有限公司 | A kind of power failure complaint risk Forecasting Methodology based on random forest |
| CN107506849A (en) * | 2017-07-24 | 2017-12-22 | 国网江西省电力公司电力科学研究院 | A kind of intelligent optimization distribution transforming, which has a power failure, studies and judges system |
Non-Patent Citations (1)
| Title |
|---|
| 厦门配电网可靠性提升措施的研究与应用;王乐;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20150131;第二章、第三章 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109410089A (en) | 2019-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109410089B (en) | Low-voltage trip and customer complaint prediction method, device and storage medium | |
| Xiong et al. | Optimal planning of storage in power systems integrated with wind power generation | |
| Xydas et al. | Forecasting electric vehicle charging demand using support vector machines | |
| Kulkarni et al. | A spiking neural network (SNN) forecast engine for short-term electrical load forecasting | |
| Hedayati-Mehdiabadi et al. | Reserve policy optimization for scheduling wind energy and reserve | |
| CN110389269B (en) | Low-voltage distribution area topological relation identification method and device based on current optimization matching | |
| CN113853528A (en) | Power utilization abnormality recognition method, device and computer readable storage medium | |
| Razghandi et al. | Short-term load forecasting for smart home appliances with sequence to sequence learning | |
| CN115034519A (en) | Method and device for predicting power load, electronic equipment and storage medium | |
| EP4193443A1 (en) | Power distribution system reconfigurations for multiple contingencies | |
| Souza et al. | Immune system memetic algorithm for power distribution network design with load evolution uncertainty | |
| CN111210260A (en) | Electricity price data processing method and device, computer equipment and storage medium | |
| CN110728395A (en) | Main transformer short-term power load calculation method and device, computer and storage medium | |
| CN116562398B (en) | Power load prediction model training method, prediction method, electronic device and medium | |
| CN113222281A (en) | Power distribution network short-term load prediction method and device based on improved AlexNet-GRU model | |
| Li et al. | A probabilistic data-driven method for response-based load shedding against fault-induced delayed voltage recovery in power systems | |
| Wang et al. | An ultra-short-term forecasting model for high-resolution solar irradiance based on SOM and deep learning algorithm | |
| CN108269007B (en) | Method and device for analyzing transformer data, computer equipment and storage medium | |
| Schleifer et al. | Hybrid renewable energy systems: The value of storage as a function of PV-wind variability | |
| CN112926777B (en) | Day-ahead load prediction method and system considering commercial building energy consumption behaviors | |
| Chen et al. | Distributed generation placement for power distribution networks | |
| CN110766286B (en) | Power distribution network topology checking method and device, computer equipment and readable storage medium | |
| Dong et al. | Image-based processing mechanism for peak load forecasting in smart grids | |
| Lotfi et al. | Incorporating Mobile Energy Resources in Optimal Power Flow Models Considering Geographical and Road Network Data | |
| Xu et al. | An LSTM approach for predicting the short-time passenger flow of urban bus |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200924 Address after: 510620 Tianhe District, Guangzhou, Tianhe South Road, No. two, No. 2, No. Applicant after: Guangzhou Power Supply Bureau of Guangdong Power Grid Co.,Ltd. Address before: 510620 Tianhe District, Guangzhou, Tianhe South Road, No. two, No. 2, No. Applicant before: GUANGZHOU POWER SUPPLY Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |