A power distribution network fault prediction method and system

Abstract

The application discloses a power distribution network fault prediction method and system, and relates to the field of power distribution network fault prediction.The method comprises the following steps: collecting power fault data and preprocessing the power fault data to obtain a fault data set; constructing a time series prediction model based on the fault data set and fitting the time series prediction model; creating a data frame of a specified prediction time range based on the fitted time series prediction model; generating a prediction result based on the fitted time series prediction model and the data frame, and performing threshold analysis; comparing the prediction result with historical data, and formulating countermeasures; and the application is convenient for maintaining and preventing the power distribution network according to the predictability of the prediction result, reduces the fault rate of the power distribution network, and reduces economic and other losses caused by the fault of the power distribution network.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power distribution network fault prediction, and in particular to a power distribution network fault prediction method and system. BACKGROUND

[0002] The power distribution network is composed of overhead lines, cables, towers, distribution transformers, disconnectors, reactive power compensators and some auxiliary facilities, and plays an important role in distributing electric energy in the power grid. With the rapid development of smart grids, the large-scale uncertain access of distributed power sources makes the fault information of the power distribution network more complex, and the accurate and rapid analysis of faults becomes more difficult. In order to ensure the highly intelligent operation of the power distribution network, it is necessary to monitor the feeder operation data in real time, give timely warning of abnormal situations, quickly locate and handle faults, and predict faults based on existing data. Therefore, a power distribution network fault prediction method is needed.

[0003] The existing power distribution network prediction method mostly ignores the influence of weather and holidays on the fault occurrence rate of the power distribution network when collecting data, resulting in inaccurate prediction results. For example, during a holiday on a sunny day, the load of the power distribution network in the area of a mobile stall or a pedestrian street increases due to the increase in tourists, resulting in an increase in the fault rate. During a holiday with heavy rain, the load of the power distribution network in the area of an indoor supermarket or cinema increases due to the increase in tourists, resulting in an increase in the fault rate. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] Therefore, the technical problem solved by the present application is that the existing power distribution network prediction method mostly ignores the influence of weather and holidays on the fault occurrence rate of the power distribution network when collecting data, resulting in inaccurate prediction results.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a power distribution network fault prediction method, comprising:

[0008] Collecting power failure data and preprocessing to obtain a fault data set;

[0009] Based on the fault data set, a time series prediction model is constructed and fitted;

[0010] Based on the fitted time series prediction model, a data frame of a specified prediction time range is created;

[0011] Based on the fitted time series prediction model and the data frame, a prediction result is generated and threshold analysis is performed;

[0012] The prediction result is compared with historical data to formulate a response measure.

[0013] As a preferred solution of the power distribution network fault prediction method, wherein:

[0014] The comparison of the prediction result with the historical data to formulate a response measure includes:

[0015] The obtained prediction result is compared with the historical data stored in the big data memory to obtain the measures needed to be implemented to cope with the result.

[0016] As a preferred solution of the power distribution network fault prediction method, wherein:

[0017] The construction of the time series prediction model based on the fault data set and the fitting include:

[0018] The time series prediction model is constructed to include seasonal terms, trend terms, residual terms and holiday terms.

[0019] As a preferred solution of the power distribution network fault prediction method, wherein:

[0020] The construction of the time series prediction model based on the fault data set and the fitting also includes:

[0021] In the time series prediction, for all times t, there are two forms, namely, additive form and multiplicative form, wherein the additive form is:

[0022] y t =S t +T t +R t

[0023] Wherein, S t is the seasonal term, T t is the trend term, and R t is the residual term.

[0024] As a preferred solution of the power distribution network fault prediction method, wherein:

[0025] The construction of the time series prediction model based on the fault data set and the fitting also includes:

[0026] The multiplicative form is:

[0027] y t =S t ×T t ×R t

[0028] Taking the logarithm of the multiplication, the additive form and the multiplicative form are equivalent:

[0029] Iny t= InS t + InT t + InR t

[0030] The expression of holiday term ε(t) is added:

[0031] y(t) = g(t) + s(t) + h(t) + ε(t)

[0032] Wherein, g(t) is a trend term, which represents the trend of time series on the non-periodic; s(t) is a seasonal term, which is in units of weeks or years; h(t) is a holiday term, which represents whether there is a holiday in the time range; ε(t) is an error term / residual term;

[0033] The trend term g(t) is a function of logistic regression function and piecewise linear function, the form of logistic regression function is:

[0034] f(x) = C / (1 + e- k(x-m) )

[0035] Wherein, C is the maximum asymptotic value of the curve, k represents the growth rate of the curve, m represents the midpoint of the curve, the special form of logistic regression function is sigmoid function, and the derivative form of sigmoid function is:

[0036] y' = y(1-y).

[0037] As an optimal solution of the power distribution network fault prediction method, wherein:

[0038] The time series prediction model is constructed and fitted based on the fault data set, which further comprises:

[0039] Holiday influence range parameter κ is introduced i , there are L holidays, the formula is:

[0040]

[0041] Z(t) = (1 {t∈Di} ,···,1 {t∈DL} )

[0042] κ = (κ1,···,κ1) T

[0043] Wherein, κ ~ Normal(0, v 2 ), the greater the value, the greater the influence of holiday on the model; the smaller the value, the smaller the influence of holiday on the model.

[0044] As an optimal solution of the power distribution network fault prediction method, wherein:

[0045] The generating a prediction result based on the fitted time series prediction model and the data frame and performing threshold analysis comprises:

[0046] The prediction value and its confidence interval are obtained by setting the data frame of the prediction time range in the prediction model; the confidence interval of the prediction value is compared with the actual value, and a warning is given when the prediction value exceeds the threshold or the actual value exceeds the confidence interval.

[0047] In a second aspect, an embodiment of the present application provides a power distribution network fault prediction system, comprising:

[0048] A data preparation and preprocessing module is configured to collect power fault data and perform preprocessing to obtain a fault data set;

[0049] A model construction and fitting module is configured to construct a time series prediction model based on the fault data set and perform fitting;

[0050] A future data frame creation module is configured to create a data frame of a specified prediction time range based on the fitted time series prediction model;

[0051] A prediction and analysis module is configured to generate a prediction result based on the fitted time series prediction model and the data frame and perform threshold analysis;

[0052] A unified processing and response measure module is configured to compare the prediction result with historical data and develop a response measure.

[0053] In a third aspect, an embodiment of the present application provides a computing device, comprising:

[0054] A memory and a processor;

[0055] The memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions, so that the one or more processors implement the power distribution network fault prediction method according to any one of the embodiments of the present application.

[0056] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium storing computer executable instructions, which are executed by a processor to implement the power distribution network fault prediction method.

[0057] The application has the beneficial effects that: the application analyzes and processes the internal data of the power grid enterprise in the collected data source platform and the related information on the external environment, i.e. big data, through a crawler, and comprehensively considers the meteorological conditions and holiday conditions of the period and the region to be predicted during the analysis and processing, i.e. the prediction accuracy is improved by enriching and expanding the existing content and field, so as to facilitate the maintenance and prevention of the distribution network according to the prediction result, and thus reduce the failure rate of the distribution network, and reduce the economic and other losses caused by the failure of the distribution network. BRIEF DESCRIPTION OF DRAWINGS

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings needed to be used in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0059] Figure 1 is the overall flowchart of the distribution network fault prediction method described in the application;

[0060] Figure 2 is the internal structure diagram of the trend item of the distribution network fault prediction method described in the application. DETAILED DESCRIPTION

[0061] In order to make the above-mentioned purposes, features and advantages of the application more apparent and easy to understand, the specific embodiments of the application will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor should be within the protection scope of the application.

[0062] In the following description, many specific details are set forth in order to provide a thorough understanding of the application, but the application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the application, therefore the application is not limited to the specific embodiments disclosed below.

[0063] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an independent or alternative embodiment that excludes other embodiments.

[0064] Embodiment 1

[0065] Reference Figure 1For the first embodiment of the present application, the embodiment provides a power distribution network fault prediction method, comprising:

[0066] S1: collecting power failure data and preprocessing to obtain a fault data set;

[0067] S2: based on the fault data set, a time series prediction model is constructed and fitted;

[0068] S3: based on the fitted time series prediction model, a data frame of a specified prediction time range is created;

[0069] S4: based on the fitted time series prediction model and the data frame, a prediction result is generated, and threshold analysis is performed;

[0070] S5: compare the prediction result with the historical data, and formulate countermeasures.

[0071] It should be noted that through steps S1-S5, in the process of collecting, preprocessing, modeling, predicting, analyzing and countermeasure formulation of power failure data, fault prediction and risk management are systematically realized, thereby improving the stability and reliability of power grid operation.

[0072] Embodiment 2

[0073] Referring to Figure 1 and Figure 2 For an embodiment of the present application, based on the previous embodiment, a power distribution network fault prediction method is provided, comprising:

[0074] In the present application embodiment, the above-mentioned step S1 of collecting power failure data and preprocessing to obtain a fault data set comprises:

[0075] Data acquisition: the internal data of power grid enterprises in the data source platform is collected and preprocessed;

[0076] Data extraction: the data uploaded to the big data platform by external environmental information data is extracted, de-duplicated and preprocessed;

[0077] Data preprocessing includes power distribution network fault influence factor analysis, data mining extraction and sample screening.

[0078] Specifically, data extraction is completed by a keyword extraction algorithm, and the formula of the extraction algorithm is TF-IDF=TF*IDF, wherein TF is the word frequency, which refers to the number of times a certain word appears in an article / the total number of words in the article, and IDF is the inverse document frequency.

[0079] The extraction logic of TF-IDF = TF * IDF is as follows: first, perform word segmentation, part-of-speech filtering, deduplication, and stop word removal data preprocessing operations on the text to be extracted keywords to obtain candidate keywords; second, calculate the TF-IDF values of all candidate keywords; and take the keywords with the top N large TF-IDF values among all candidate keywords as the final keywords of the document.

[0080] In another possible implementation, data extraction can also be done by the following methods:

[0081] Machine learning algorithm: use supervised learning or unsupervised learning algorithm to identify and extract relevant keywords. For example, a classification model can be trained to determine whether a word is a keyword.

[0082] Natural language processing (NLP) technology: use NLP toolkit (such as spaCy, NLTK, etc.) for entity recognition, named entity extraction, etc. to extract keywords with specific meanings.

[0083] Text summarization technology: use text summarization algorithm to automatically generate summary of text and extract important keywords.

[0084] Topic model: such as Latent Dirichlet Allocation (LDA), which can identify topics in text and extract keywords for each topic.

[0085] Statistical method: in addition to TF-IDF, other statistical indicators such as variants of term frequency-inverse document frequency, information gain, chi-square test, etc. can be used to evaluate the importance of keywords.

[0086] Deep learning model: use pre-trained language model (such as BERT, RoBERTa, etc.) to extract keywords, which can capture more complex semantic relationships.

[0087] In the embodiments of the present application, the step S2 of constructing and fitting the time series prediction model based on the fault data set includes:

[0088] A model object for Prophet time series prediction is constructed, including seasonal terms, trend terms, residual terms, and holiday terms; and the fault data set after preprocessing is fitted into the model using the fit method.

[0089] In the Prophet time series prediction, all times t include two forms, namely additive form and multiplicative form, where the additive form is represented as:

[0090] y t = S t + T t + Rt

[0091] Among them, S t For seasonal items, T t For the trend term, R t The remaining terms.

[0092] Multiplication is expressed as follows:

[0093] y t =S t ×T t ×R t

[0094] Taking the logarithm of multiplication, the addition and multiplication forms are equivalent as follows:

[0095] Iny t =InS t +InT t +InR t

[0096] That is, when making a prediction model, the logarithm of y is taken first, and then the time series is decomposed.

[0097] In the model building phase, Prophet time series prediction also includes a holiday term ε(t). The representation of ε(t) with the holiday term is as follows:

[0098] y(t) = g(t) + s(t) + h(t) + ε(t)

[0099] g(t) is the trend term, which represents the trend of the time series in a non-periodic manner; s(t) is the seasonal term, which is in units of weeks or years; h(t) is the holiday term, which indicates whether there are holidays within the time range; ε(t) is the error term / residual term.

[0100] like Figure 2 As shown, the trend term of g(t) is a function of logistic regression and piecewise linearity, and the form of the logistic regression function is as follows:

[0101] f(x) = C / (1+e -k(x-m ))

[0102] Where C is the maximum asymptotic value of the curve, k represents the growth rate of the curve, m represents the midpoint of the curve, and the special form of the logistic regression function is the sigmoid function, and the derivative of the sigmoid function is:

[0103] y′=y(1-y).

[0104] In the model construction stage, the change point position needs to be set in the Prophet time series prediction, and the setting method of the change point position includes manual and automatic selection. When setting the change point, some boundary conditions need to be considered. S change points are set, and the positions of the change points are time stamps j ,1≤j≤S s At these time stamps, the change of the growth rate needs to be given; let there be a vector δ∈R j : where δ j j j∶td＞sj represents the change amount of the growth rate at the time stamp s s j Therefore, the growth rate at the time stamp t is k+a T δ, once the variable k is determined, the other parameter m is adjusted accordingly. When the growth rate k is adjusted, the offset m corresponding to each change point should also be adjusted to connect the last time point of each segment, so through mathematical calculation, the following can be obtained: Substitute the above logistic regression model:

[0105]

[0106] Where: a(t)=(a1(t),···,a s (t)) T ,δ=(δ1,···,δ s ) T ,γ=(γ1,···,γ s ) T .

[0107] The segmented linear function is expressed as:

[0108] g(t)=(k+a(t)δ)·t+(m+a(t) T γ)

[0109] k is the growth rate, δ is the change amount of the growth rate, m is the offset parameter, and γ is the offset parameter. The biggest difference between the segmented linear function and the logistic function is that the setting of γ is different. In the segmented linear function, γ=(γ1,···,γ s ) T ,γ j =-s j δ j .

[0110] The seasonal term model in the model construction stage uses Fourier series to simulate the periodicity of the time series: P = 365.25 represents a period of a year, P = 7 represents a period of a week, and the holiday term represents the influence of the i-th holiday. Di represents a period of time before and after the holiday, and a parameter κ is needed to represent the influence range of the holiday, i.e., L holidays are set. The formula is as follows: i

[0111]

[0112] Z(t) = (1 {t∈Di} ,···,1 {t∈DL} )

[0113] κ = (κ1,···,κ1) T

[0114] where κ ~ Normal(0, v 2 ) and the normal distribution is affected by v = holidays_prior_scale, and the default value is 10. The larger the value, the greater the influence of the holiday on the model; the smaller the value, the smaller the influence of the holiday on the model.

[0115] The form of the Fourier series is as follows: N represents the number of cycles that are expected to be used in the model.

[0116] For example, for a sequence with a period of a year (P = 365.25), N = 10; for a sequence with a period of a week (P = 7), N = 3.

[0117] When N = 10,

[0118] When N = 3,

[0119] Therefore, the seasonal term of the time series is s(t) = X(t)β, and the initialization of β is β ~ Normal(0, σ 2 ). The larger the value of σ, the more obvious the effect of the season; the smaller the value, the less obvious the effect of the season.

[0120] In the embodiments of the present application, the step S3 of creating a data frame of a specified prediction time range based on the fitted time series prediction model includes:

[0121] A data frame containing future prediction dates is created using the make_future_data frame method, and the prediction time range is specified.

[0122] ​For example, if we want to predict the failure data for the next 365 days, we can set periods = 365 and specify the frequency (e.g., freq = 'D' means by day).

[0123] It should be noted that this step ensures the accuracy of the predicted time range and provides a clear time frame for subsequent prediction. By setting appropriate time intervals, the precision and practicality of the prediction can be improved.

[0124] In the embodiments of the present application, the step S4 of generating a prediction result based on the fitted time series prediction model and the data frame and performing threshold analysis includes:

[0125] By setting the data frame of the prediction time range in the prediction model, the prediction value and its confidence interval are obtained; the confidence interval of the prediction value is compared with the actual value, and a warning is given when the prediction value exceeds the threshold or the actual value exceeds the confidence interval.

[0126] In another possible implementation, the predict method is used to generate a prediction result, including the prediction value, confidence interval, etc.

[0127] The plot method is used to visualize the prediction result, including the actual value, prediction value and confidence interval.

[0128] Set a threshold (e.g., failure rate exceeds a certain value), and trigger an alarm when the prediction result reaches or exceeds the threshold.

[0129] It should be noted that the visualization of the prediction result provided by this step helps to intuitively understand the future trend and uncertainty; threshold analysis can timely discover potential failure risks and take measures in advance.

[0130] In the embodiments of the present application, the step S5 of comparing the prediction result with the historical data to develop countermeasures includes:

[0131] The obtained prediction result is compared with the historical data stored in the big data memory to obtain the measures needed to be implemented to cope with the result.

[0132] Specifically, the prediction result is compared with the historical data to analyze the accuracy and deviation of the prediction.

[0133] According to the comparison result, the performance of the prediction model is evaluated, and the model is optimized if necessary.

[0134] Develop specific countermeasures, such as increasing maintenance frequency, adjusting operating parameters, etc.

[0135] It should be noted that the step verifies the accuracy of the prediction by comparing historical data, improves the reliability of the decision, formulates targeted countermeasures, and can effectively prevent faults and improve the stability and safety of system operation.

[0136] Embodiment 3

[0137] The above is a schematic scheme of the power distribution network fault prediction method of the embodiment. It should be noted that the technical scheme of the power distribution network fault prediction system belongs to the same concept as the technical scheme of the power distribution network fault prediction method described above. The technical details of the power distribution network fault prediction system in this embodiment are not described in detail, and can be referred to the description of the technical scheme of the power distribution network fault prediction method described above.

[0138] The embodiment also provides a system based on the power distribution network fault prediction method, comprising:

[0139] A data preparation and preprocessing module is configured to collect power fault data and perform preprocessing to obtain a fault data set.

[0140] A model construction and fitting module is configured to construct a time series prediction model based on the fault data set and perform fitting.

[0141] A future data frame creation module is configured to create a data frame of a specified prediction time range based on the fitted time series prediction model.

[0142] A prediction and analysis module is configured to generate a prediction result based on the fitted time series prediction model and the data frame, and perform threshold analysis.

[0143] A unified adjustment processing and countermeasure module is configured to compare the prediction result with historical data and formulate countermeasures.

[0144] The embodiment also provides a computing device suitable for the power distribution network fault prediction method, comprising:

[0145] A memory and a processor; the memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions to implement the power distribution network fault prediction method proposed in the above embodiment.

[0146] The embodiment also provides a storage medium having a computer program stored thereon, which is executed by a processor to implement the power distribution network fault prediction method proposed in the above embodiment.

[0147] The storage medium proposed in the embodiment belongs to the same inventive concept as the power distribution network fault prediction method proposed in the above embodiment. The technical details not described in detail in the embodiment can be referred to the above embodiment, and the embodiment has the same beneficial effects as the above embodiment.

[0148] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A power distribution network fault prediction method characterized by, The method comprises the following steps: collecting power failure data and preprocessing to obtain a failure data set; based on the failure data set, a time series prediction model is constructed and fitted; the step of constructing a time series prediction model based on the failure data set and fitting the model comprises the following steps: constructing a time series prediction model including seasonal terms, trend terms, residual terms and holiday terms; in the time series prediction, for all times t, there are two forms, namely additive form and multiplicative form, wherein the additive form is expressed as: ； wherein, is a seasonal term, is a trend term, is a residual term; the multiplicative form is expressed as: ； taking the logarithm of the multiplication, the additive form and the multiplicative form are equivalent to: ； Adding a holiday item The manifestation is: ； wherein, is a trend term, which represents the changing trend of the time series over the non-periodic; is a seasonal term, which is in units of week or year; is a holiday term, which represents whether there is a holiday in the time range; is an error term / residual term; The trend term is a function of a logistic regression function and a piecewise linear function, the logistic regression function being of the form: ； wherein, is the maximum asymptote of the curve, k represents the growth rate of the curve, m represents the midpoint of the curve, the special form of the logistic regression function is the sigmoid function, and the derivative form of the sigmoid function is: ； Introducing holiday influence range parameter , provided with holidays, the formula is expressed as: ； wherein, Normal (0, v 2 ), the greater the value, the greater the impact of the holiday on the model; the smaller the value, the smaller the impact of the holiday on the model; based on the fitted time series prediction model, a data frame of a specified prediction time range is created; based on the fitted time series prediction model and the data frame, a prediction result is generated and threshold analysis is performed; comparing the prediction result with historical data to develop countermeasures; the step of comparing the prediction result with historical data to develop countermeasures comprises the following steps: comparing the obtained prediction result with historical data stored in the big data to obtain measures needed to be taken to deal with the result.

2. The power distribution grid failure prediction method of claim 1, wherein, the step of generating a prediction result based on the fitted time series prediction model and the data frame, and performing threshold analysis comprises the following steps: by setting the data frame of the prediction time range in the prediction model, the prediction value and its confidence interval are obtained; the confidence interval of the prediction value is compared with the actual value, and a warning is given when the prediction value exceeds the threshold or the actual value exceeds the confidence interval.

3. A system employing the power distribution network fault prediction method according to any one of claims 1 to 2, characterized by, The method comprises the following steps: a data preparation and preprocessing module for collecting power failure data and preprocessing to obtain a failure data set; a model construction and fitting module for constructing a time series prediction model based on the failure data set and fitting the model; a future data frame creation module for creating a data frame of a specified prediction time range based on the fitted time series prediction model; a prediction and analysis module for generating a prediction result based on the fitted time series prediction model and the data frame, and performing threshold analysis; a unified processing and countermeasure module for comparing the prediction result with historical data to develop countermeasures.

4. A computing device, comprising: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions, which realize the steps of the power distribution network failure prediction method in any one of claims 1 to 2 when executed by the processor.

5. A computer readable storage medium storing computer executable instructions, which realize the steps of the power distribution network failure prediction method in any one of claims 1 to 2 when executed by the processor.

Images