Method for identifying intermittent electricity stealing based on electric quantity line loss relationship
By extracting the multi-dimensional characteristic indicators of user power and station area line loss timing data and building a discriminant model, the problem of too many false positive users in intermittent power theft identification in the existing technology is solved, and a more efficient identification process is achieved, saving manpower and time costs.
Patent Information
- Application Number
- CN202411957922.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-30
- Publication Date
- 2025-04-29
AI Technical Summary
There are a large number of false positive suspect users in existing methods of identifying intermittent power theft, which makes manual screening consume a lot of time and manpower.
By extracting multi-dimensional characteristic indicators of user power and station area line loss timing data, and building an efficient discriminant model, the proportion of false positive suspect users is reduced.
Significantly reduce the proportion of false positives in the list of suspected users of intermittent power theft, saving labor and time costs.
Smart Images
Figure CN119904246A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electric energy metering and collection, and in particular to a method for identifying intermittent electricity theft based on the relationship between electric quantity and line loss. Background Art
[0002] The existing method for identifying intermittent electricity theft is mainly to calculate the Pearson correlation coefficient between the user's electricity consumption time series and the substation line loss time series. If the correlation coefficient is negative and less than a specified threshold, the user is considered to be suspected of intermittent electricity theft.
[0003] The existing method for identifying intermittent electricity theft is mainly based on a single correlation indicator and sets a fixed threshold based on experience to screen suspected users. Due to the complexity of the power and line loss time series data, there are a large number of false positive suspected users (normal users) in the list of suspected electricity theft users obtained by this method. Further manual screening is required to eliminate a large number of false positive users. Summary of the invention
[0004] 1. Technical issues to be resolved
[0005] In view of the shortcomings of the prior art, the present invention provides a method for identifying intermittent electricity theft based on the relationship between electricity consumption and line loss. Based on the user's electricity consumption and the time series data of the substation line loss, more dimensional feature indicators are extracted, and based on these feature indicators, an efficient discrimination model is constructed, which significantly reduces the proportion of false positives in the output list of suspected intermittent electricity theft users, thereby saving a lot of manpower and time costs.
[0006] (II) Technical solution
[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for identifying intermittent electricity theft based on the relationship between power line loss, specifically comprising the following steps:
[0008] S1. Data preprocessing: Get the daily line loss value data of the most recent N days in the station area, as well as the daily electricity data of the user on the corresponding date, and then fill the missing values for the station area line loss time series data and the user electricity time series data. The station area line loss time series data obtained after data preprocessing is expressed as:
[0009] L=[l1,l2,l3,…,l N ];
[0010] Then the user power time series data is expressed as:
[0011] E=[e1,e2,e3,…,e N ];
[0012] S2. Model feature index extraction: specifically includes the following steps:
[0013] T1, calculate the time series difference and the time series absolute difference;
[0014] T2, screening time series subsequences;
[0015] T3, characteristic index calculation;
[0016] S3, constructing a discriminant model based on feature indicators: Based on the feature indicator set {F L ,F E ,F P ,F R}, use the indicator logic judgment method to build a rule discrimination model, or mark a batch of training samples in advance, and then use the machine learning algorithm to build a binary classification machine learning model. If F L >0.3∧F E >0.3∧-1.2 <F P <0.8∧F R If the condition >0.05 is met, it is considered that the user is stealing electricity intermittently.
[0017] Preferably, in step S1, N≥30, and data with daily line loss value less than or equal to 0 are forcibly set to missing values.
[0018] Preferably, in step S1, a linear interpolation algorithm is used to fill missing values in the substation line loss time series data and the user power time series data.
[0019] Preferably, the specific steps of calculating the timing difference and the timing absolute difference in step T1 are as follows:
[0020] Based on the time series data of the line loss in the substation area and the time series data of the user's power consumption, the first-order difference of the time series and the first-order absolute difference of the time series are calculated respectively. The first-order difference series of the line loss is expressed as:
[0021]
[0022] The first-order absolute difference sequence of line loss is expressed as:
[0023]
[0024] Similarly, the first-order difference series of electric quantity is expressed as:
[0025]
[0026] The first-order absolute difference series of electric quantity is expressed as:
[0027]
[0028] Preferably, the step T2 of screening the time sequence subsequence specifically comprises the following steps:
[0029] Since it is only meaningful to analyze the relationship between power consumption change and line loss change when the user's status changes (from power theft to normal, or from normal to power theft), the screening rules for the candidate time set corresponding to the user's status change are defined as follows:
[0030]
[0031] The absolute difference weight subsequence of line loss is obtained as follows:
[0032]
[0033] in,
[0034] Preferably, the absolute difference weight subsequence of electric quantity is:
[0035]
[0036] in,
[0037] Preferably, the line loss power differential ratio subsequence is:
[0038]
[0039]
[0040] Preferably, the specific steps of calculating the characteristic index in step T3 are as follows:
[0041] By combining the line loss absolute difference weight subsequence and the line loss power difference ratio subsequence, the line loss absolute difference weighted index is obtained:
[0042]
[0043] By combining the power absolute difference weight subsequence and the line loss power difference ratio subsequence, the power absolute difference weighted index is obtained:
[0044]
[0045] Based on the line loss power differential ratio subsequence, the line loss power differential ratio weighted index is obtained:
[0046]
[0047] Similarly, based on the line loss power differential ratio subsequence, the user state switching frequency ratio index is obtained:
[0048]
[0049] in, represents the indicative function, if If the condition is met, the result is 1; otherwise, it is 0.
[0050] Preferably, the machine learning algorithm in step S3 is one of a support vector machine algorithm, a random forest algorithm or an artificial neural network algorithm.
[0051] (III) Beneficial effects
[0052] The present invention provides a method for identifying intermittent electricity theft based on the relationship between electricity and line loss. Compared with the prior art, the method has the following beneficial effects: the method for identifying intermittent electricity theft based on the relationship between electricity and line loss specifically comprises the following steps: S1, data preprocessing: obtaining the daily line loss value data of the most recent N terraces and the daily electricity data of the user on the corresponding date, and then filling the missing values of the terrace line loss time series data and the user electricity time series data, respectively; S2, extracting model feature indicators; S3, constructing a discriminant model based on the feature indicators: based on the feature indicator set {F L ,F E ,F P ,F R}, use the indicator logic judgment method to build a rule discrimination model, or mark a batch of training samples in advance, and then use the machine learning algorithm to build a binary classification machine learning model. If F L >0.3∧F E >0.3∧-1.2 <F P <0.8∧F R >0.05, the user is considered to have intermittent electricity theft. By extracting more dimensional feature indicators based on the user's electricity consumption and the time series data of the line loss in the substation area, and based on these feature indicators, the discriminant model constructed can significantly reduce the proportion of false positives of intermittent electricity theft suspects. Based on the user's electricity consumption and the time series data of the line loss in the substation area, the calculation of four feature indicator sets {F L ,F E ,F P ,F R}, and based on the feature indicator set, an efficient intermittent electricity theft identification model can be further constructed, thereby saving a lot of manpower and time costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION
[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0055] See also Figure 1 The embodiment of the present invention provides two technical solutions: a method for identifying intermittent electricity theft based on the relationship between power line loss, specifically including the following embodiments:
[0056] Embodiment 1: A method for identifying intermittent electricity theft based on the relationship between power consumption and line loss, specifically comprising the following steps:
[0057] S1. Data preprocessing: Get the daily line loss value data of the most recent N days in the station area, as well as the daily electricity data of the user on the corresponding date, and then fill the missing values for the station area line loss time series data and the user electricity time series data. The station area line loss time series data obtained after data preprocessing is expressed as:
[0058] L=[l1,l2,l3,…,l N ];
[0059] Then the user power time series data is expressed as:
[0060] E=[e1,e2,e3,…,e N ];
[0061] S2. Model feature index extraction: specifically includes the following steps:
[0062] T1, calculate the time series difference and the time series absolute difference;
[0063] T2, screening time series subsequences;
[0064] T3, characteristic index calculation;
[0065] S3, constructing a discriminant model based on feature indicators: Based on the feature indicator set {F L ,F E ,F P ,F R}, use the indicator logic judgment method to build a rule discrimination model, or mark a batch of training samples in advance, and then use the machine learning algorithm to build a binary classification machine learning model. If F L >0.3∧F E >0.3∧-1.2 <F P <0.8∧F R If the condition >0.05 is met, it is considered that the user is stealing electricity intermittently.
[0066] In this embodiment of the present invention, in step S1, N≥30, and data with a daily line loss value less than or equal to 0 are forcibly set to missing values.
[0067] In the embodiment of the present invention, in step S1, a linear interpolation algorithm is used to fill missing values in the substation line loss time series data and the user power time series data.
[0068] Embodiment 2: The technical solution of the embodiment of the present invention is different from that of embodiment 1 in that: the specific steps of calculating the timing difference and the timing absolute difference in step T1 are as follows:
[0069] Based on the time series data of the line loss in the substation area and the time series data of the user's power consumption, the first-order difference of the time series and the first-order absolute difference of the time series are calculated respectively. The first-order difference series of the line loss is expressed as:
[0070]
[0071] The first-order absolute difference sequence of line loss is expressed as:
[0072]
[0073] Similarly, the first-order difference series of electric quantity is expressed as:
[0074]
[0075]
[0076] The first-order absolute difference series of electric quantity is expressed as:
[0077]
[0078] In the embodiment of the present invention, step T2 of screening the time sequence subsequence specifically includes the following steps:
[0079] Since it is only meaningful to analyze the relationship between power consumption change and line loss change when the user's status changes (from power theft to normal, or from normal to power theft), the screening rules for the candidate time set corresponding to the user's status change are defined as follows:
[0080]
[0081] The absolute difference weight subsequence of line loss is obtained as follows:
[0082]
[0083] in,
[0084] In the embodiment of the present invention, the absolute difference weight subsequence of electric quantity is:
[0085]
[0086] in,
[0087] In the embodiment of the present invention, the line loss power differential ratio subsequence is:
[0088]
[0089] In the embodiment of the present invention, the specific steps of calculating the characteristic index in step T3 are as follows:
[0090] By combining the line loss absolute difference weight subsequence and the line loss power difference ratio subsequence, the line loss absolute difference weighted index is obtained:
[0091]
[0092] By combining the power absolute difference weight subsequence and the line loss power difference ratio subsequence, the power absolute difference weighted index is obtained:
[0093]
[0094] Based on the line loss power differential ratio subsequence, the line loss power differential ratio weighted index is obtained:
[0095]
[0096] Similarly, based on the line loss power differential ratio subsequence, the user state switching frequency ratio index is obtained:
[0097]
[0098] in, represents the indicative function, if If the condition is met, the result is 1; otherwise, it is 0.
[0099] In this embodiment of the present invention, the machine learning algorithm in step S3 is a support vector machine algorithm.
[0100] In summary, the present invention extracts more dimensional feature indicators based on the user power consumption and the time series data of the line loss in the substation area, and the discriminant model constructed based on these feature indicators can significantly reduce the false positive ratio of the suspected users of intermittent power theft output. Based on the user power consumption and the time series data of the line loss in the substation area, the present invention proposes to calculate four feature indicator sets {F L ,F E ,F P ,F R}, and based on the feature indicator set, an efficient intermittent electricity theft identification model can be further constructed, thereby saving a lot of manpower and time costs.
[0101] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.
[0102] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0103] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A method for identifying intermittent electricity theft based on the relationship between power line loss, characterized by: The specific steps include: S1. Data preprocessing: Get the daily line loss value data of the most recent N days in the station area, as well as the daily electricity data of the user on the corresponding date, and then fill the missing values for the station area line loss time series data and the user electricity time series data. The station area line loss time series data obtained after data preprocessing is expressed as: L＝[l1，l2，l3，...，l N ]； Then the user power time series data is expressed as: And=[e1,e2,e3,...,e N ]; S2. Model feature index extraction: specifically includes the following steps: T1, calculate the time series difference and the time series absolute difference; T2, screening time series subsequences; T3, characteristic index calculation; S3, constructing a discriminant model based on feature indicators: Based on the feature indicator set {F L , F E , F P , F R }, use the indicator logic judgment method to build a rule discrimination model, or mark a batch of training samples in advance, and then use the machine learning algorithm to build a binary classification machine learning model. If F L >0.3∧F E >0.3∧-1.2<F P <0.8∧F R >0.05, the user is considered to be stealing electricity intermittently.
2. A method for identifying intermittent electricity theft based on the relationship between power line loss according to claim 1, characterized in that: In the step S1, N≥30, and for data whose daily line loss value is less than or equal to 0, it is forcibly set to a missing value.
3. The method for identifying intermittent electricity theft based on the relationship between power line loss according to claim 1 is characterized by: In the step S1, a linear interpolation algorithm is used to fill missing values in the substation line loss time series data and the user power time series data.
4. The method for identifying intermittent electricity theft based on the relationship between power line loss according to claim 1 is characterized by: The specific steps of calculating the timing difference and the timing absolute difference in step T1 are as follows: Based on the time series data of the line loss in the substation area and the time series data of the user's power consumption, the first-order difference of the time series and the first-order absolute difference of the time series are calculated respectively. The first-order difference series of the line loss is expressed as: The first-order absolute difference sequence of line loss is expressed as: Similarly, the first-order difference series of electric quantity is expressed as: The first-order absolute difference series of electric quantity is expressed as:
5. The method for identifying intermittent electricity theft based on the relationship between power line loss according to claim 4 is characterized by: The step T2 of screening the time sequence subsequence specifically comprises the following steps: Since the relationship between power change and line loss change is analyzed only when the user status changes, the screening rules for the candidate time set corresponding to the user status change are as follows: The absolute difference weight subsequence of line loss is obtained as follows: in, 6. A method for identifying intermittent electricity theft based on the relationship between power line loss according to claim 5, characterized in that: The absolute difference weight subsequence of electric quantity is: in, 7. A method for identifying intermittent electricity theft based on the relationship between power line loss according to claim 6, characterized in that: The line loss power differential ratio subsequence is:
8. The method for identifying intermittent electricity theft based on the relationship between power line loss according to claim 7 is characterized by: The specific steps of calculating the characteristic index in step T3 are as follows: By combining the line loss absolute difference weight subsequence and the line loss power difference ratio subsequence, the line loss absolute difference weighted index is obtained: By combining the power absolute difference weight subsequence and the line loss power difference ratio subsequence, the power absolute difference weighted index is obtained: Based on the line loss power differential ratio subsequence, the line loss power differential ratio weighted index is obtained: Similarly, based on the line loss power differential ratio subsequence, the user state switching frequency ratio index is obtained: in, represents the indicative function, if If the condition is met, the result is 1; otherwise, it is 0.
9. The method for identifying intermittent electricity theft based on the relationship between power line loss according to claim 1, characterized in that: The machine learning algorithm in step S3 is one of a support vector machine algorithm, a random forest algorithm or an artificial neural network algorithm.