Distribution Low-Voltage Outlet Leakage Monitoring System and Method

Abstract

The present invention relates to the technical field of distribution maintenance, specifically a low-voltage outgoing line leakage monitoring system and method for distribution. First, data is collected and an electricity consumption habit model is constructed for users. Then, the power outage moment and the daily fluctuation events are determined. Then, based on the daily fluctuation events, the interval time of the power fluctuation occurring after the daily fluctuation events is predicted in the electricity consumption habit model and the predicted time difference is calculated. The predicted time differences of all users are obtained, and the user with the shortest predicted time difference is determined as the leakage user. The present invention can predict the users who may cause faults when the load switch trips due to leakage faults or short-circuit faults, so as to make a reasonable power transmission sequence when the power supply is restored, and avoid multiple power outages caused by the inability to determine the faulty users during the process of restoring power supply, which causes inconvenience to electricity users.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution maintenance, and specifically to a leakage monitoring system and method for low-voltage outgoing lines of power distribution. Background Art

[0002] In the residential power supply system, the power source introduced from the upper-level power distribution system is mostly three-phase power, which is distributed to users after passing through the low-voltage power distribution system. When distributing power to users, usually one load switch supplies power to multiple users. For example, one load switch may supply power to 10 users or 15 users. Therefore, when the load switch trips due to a fault in the electrical equipment of a user, if it is not known which user caused the trip, the trip may occur again after power-on. Frequent tripping will not only affect the electrical equipment but also the user's power consumption experience, resulting in inconvenience to the user's power consumption life. At the same time, during the power-on process, if it is not known which user caused the power outage, when powering on, the user who caused the power outage may not be the first to be powered on. For example, after 5 users have been successfully powered on, the sixth user to be powered on is the user who caused the power outage. At this time, the first 5 users who have been powered on may already be in normal use after power-on, and when closing the switch for the sixth user, it may cause the load switch to trip, causing the first 5 users to lose power again. Therefore, after the main load switch for powering users trips, it is very necessary to obtain the information of the user who caused the trip, which is also an important prerequisite for providing high-quality power supply services and reducing damage to electrical equipment. Currently, only the form of trial power-on one by one is used for judgment, and then the user who caused the trip is excluded to supply power to other normal users first. After assisting in troubleshooting the tripping user, power is supplied to this user again. Therefore, designing a method to identify the user who caused the trip has become an urgent requirement. Summary of the Invention

[0003] The technical problem to be solved by the present invention is: to provide a leakage user discrimination method for discriminating the user who caused the trip.

[0004] The technical solution for the technical problem to be solved by the present invention is: a leakage user discrimination method, including the following steps:

[0005] Step 1: Build an electricity consumption habit model for each user served by the load switch one by one;

[0006] Step 2: Determine the power outage moment and the load curve during a set time period before the power outage moment, and obtain the daily fluctuation event Dj according to the obtained load curve during the set time period;

[0007] Step 3: Input the daily fluctuation event Dj of the corresponding user into the user's electricity consumption habit model to predict the interval time Tb of the power fluctuation after the daily fluctuation event, and calculate the predicted time difference Tc = |Ta - Tb|, where Ta is the interval time from after the daily fluctuation event Dj to the power outage moment;

[0008] Step 4: Repeat Step 3 until the predicted time differences Tc of all users are obtained, and the user with the shortest predicted time difference is determined as the user with electricity leakage.

[0009] Preferably, in Step 1:

[0010] S1.1 Collect the daily load change data of the user to form a data set;

[0011] S1.2 Perform power decomposition on the collected daily load change data, and generate a fluctuation event Dwt for each power fluctuation;

[0012] S1.3 Apply the PrefixSpan algorithm to all the fluctuation events Dwt to perform frequent pattern mining on the user and generate an electricity consumption habit model.

[0013] Preferably, the fluctuation event Dwt includes three data, namely the fluctuation power Wd, the fluctuation time Td, and the fluctuation direction Df.

[0014] Preferably, select the first daily fluctuation event D1 and input it into the user's electricity consumption habit model; obtain all the frequent sequence patterns where the daily fluctuation event D1 is located, and obtain a predicted time difference according to each frequent sequence pattern.

[0015] Preferably, select two or more consecutive daily fluctuation events to form a daily fluctuation event sequence and input it into the user's electricity consumption habit model, and calculate the predicted time difference after obtaining the corresponding frequent sequence pattern.

[0016] Preferably, the length of the set time period is 0.5 - 3 hours.

[0017] A method for restoring power supply after a fault trip includes the following steps:

[0018] Step 1: Build an electricity consumption habit model for each user served by the load switch one by one;

[0019] Step 2: Obtain the power outage moment Tt and the daily fluctuation event Dj before this moment;

[0020] Step 3: Input the daily fluctuation event Dj into the user's electricity consumption habit model to obtain the predicted time difference Tc of this user.

[0021] Step 4: Sort the predicted time differences from small to large, and restore power supply to the users in the order from small to large.

[0022] Preferably, in step 1:

[0023] S1.1 Collect the daily load change data of users to form a data set;

[0024] S1.2 Perform power decomposition on the collected daily load change data, and generate a fluctuation event Dwt for each power fluctuation;

[0025] S1.3 Apply the PrefixSpan algorithm to all the fluctuation events Dwt to mine frequent patterns of users and generate an electricity consumption habit model.

[0026] The low-voltage distribution outgoing line leakage monitoring system includes a data acquisition system, a model construction system, and a prediction and analysis system; the data acquisition system obtains the electricity consumption load data of users; the model construction system constructs an electricity consumption habit model of users by using the collected data; the prediction and analysis system obtains the power outage moment and the daily fluctuation events according to the collected data, and calculates the predicted time difference according to the electricity consumption habit model; the prediction and analysis system determines the user with the smallest predicted time difference among each user as the user causing the tripping fault.

[0027] The beneficial effects of the present invention are as follows:

[0028] The present invention can predict the users who may cause faults when the load switch trips due to leakage faults or short-circuit faults, so as to make a reasonable power transmission sequence when restoring power supply, and avoid multiple power outages caused by the inability to determine the faulty users during the process of restoring power supply, which causes inconvenience to electricity users. Description of the Drawings

[0029] Figure 1 It is a schematic flowchart of an embodiment of the present invention. Detailed Embodiments

[0030] To make the technical solutions and beneficial effects of the present invention clearer, the following further elaborates on the embodiments of the present invention.

[0031] A method for discriminating leakage users. The present invention determines the electricity consumption situation of users at the power outage moment by obtaining the electricity consumption habits of users through big data analysis, and determines which user caused the power outage due to leakage when starting electrical appliances at the power outage moment. Furthermore, during the process of restoring power supply, the problem of multiple power outages caused by restoring power supply to the power-off users is avoided.

[0032] First, an electricity consumption habit model is constructed for each user connected to the load switch. Specifically, it includes the following steps.

[0033] S1.1 After the startup discrimination, data collection is carried out. The data obtained can be electrical energy data or current data. The electrical energy data can be collected through an electric meter or extracted from the electrical energy database in the power dispatching and monitoring system. The current data can be collected through a current transformer. The data obtained is a large amount of daily load change data, that is, the daily load curve of the user.

[0034] S1.2 After obtaining the user's electricity consumption data, the data is analyzed and processed. The power change during the user's electricity consumption process is decomposed through the daily load curve. Currently, during the process of mining the user's electricity consumption habit pattern, the equipment power during the user's electricity consumption process is decomposed according to the daily load curve, and then the information of the electrical equipment used by the user is obtained. Since the specific electrical equipment does not need to be known in this application, only the user's power change needs to be decomposed. That is, each power fluctuation in the load curve is recorded and a fluctuation event Dwt is generated. The fluctuation event Dwt contains three data, namely the fluctuation power Wd, the fluctuation time Td, and the fluctuation direction Df. Among them, Wd represents the magnitude of the power change. Td represents the fluctuation time, and Df represents whether the power increases or decreases, which is represented by positive and negative 1. After the power decomposition, the fluctuation events within a day are grouped into a power change sequence in the order of the fluctuation time before and after.

[0035] S1.3 Apply the PrefixSpan algorithm to mine the frequent patterns of the user and generate an electricity consumption habit model. The PrefixSpan algorithm is a relatively commonly used frequent pattern mining algorithm.

[0036] First, set the minimum support threshold as the standard for screening frequent patterns. Then, each fluctuation event in the dataset is used as a separate sequence pattern, and the number of times it appears is recorded for each sequence pattern, and each sequence pattern is sorted according to the number of times it appears. After that, the sequence pattern is grown recursively, and the recursive execution continues until no longer pattern can be generated. Finally, at the end of the recursive process, all frequent sequence patterns that meet the minimum support requirements are output to form an electricity consumption habit model.

[0037] Apply the above steps from S1.1 to S1.3 to mine the electricity consumption habits of the users served by the load switch one by one and construct a user electricity consumption habit model Yi, where i is the number of the user served by the load switch.

[0038] After a power outage, the data of the electric energy meter can be retrieved. First, use this data to determine the power outage moment Tt and obtain the load curve for a set period before the power outage moment. Similarly, the current data collected by the current transformer can also be retrieved. The length of the set period can be 30 minutes, 1 hour, or 1.5 hours. Its value range can be 0.5 to 3 hours.

[0039] After processing the retrieved load curve, the daily fluctuation events Dj are obtained, where j is 1, 2, 3,.... Among them, D1 is a daily fluctuation event before time Tt, and D2 is the fluctuation event before D1. Having the same data structure as the fluctuation events in historical data, the daily fluctuation events also include the fluctuation power Wd, the fluctuation time Td, and the fluctuation direction Df. At the same time, the time interval length Ta between the daily fluctuation event D1 and the power outage time is calculated.

[0040] Select one, two, or three daily fluctuation events and input them into the user electricity consumption habit model. Through Dj and the user electricity consumption habit model, the daily fluctuation event D0 after the daily fluctuation event D1 can be predicted. At the same time, the time interval Tb between D0 and D1 is obtained. Then calculate the predicted time difference Tc = |Ta - Tb|.

[0041] According to the above method of calculating the predicted time difference, the time differences Tci of each user are obtained in turn. The user with the smallest predicted time difference is determined as the user with leakage or the user causing a fault trip. Or, sort according to the magnitudes of Tci, and the users with smaller predicted time differences are tried to be powered on first. This ensures reducing the impact on normal users.

[0042] When selecting a single daily fluctuation event D1, input the parameters of D1 into the user electricity consumption habit model. The user's electricity consumption habit model is composed of many frequent sequence patterns, and each frequent sequence pattern represents a kind of the user's electricity consumption habit. Then compare each frequent sequence pattern one by one to obtain the frequent sequence pattern containing the data of the fluctuation power, fluctuation time, and fluctuation direction of D1, and then retrieve this sequence pattern. For example, a frequent sequence pattern is Dwt1 - Dwt2 - Dwt3 - Dwt4 - Dwt5. By comparison, it is found that the data of D1 is the same as Dwt3. At this time, it is determined that this frequent sequence pattern is the frequent sequence pattern at the power outage moment, and then extract the parameters according to this sequence pattern. At this time, D0 is Dwt4. Tb is the interval time between Dwt3 and Dwt4.

[0043] When predicting the time of Tb, there may be multiple frequent sequence patterns for starting the electrical equipment within this time period. Therefore, for the i-th user, the predicted time differences Tci may be k, namely Tci1, Tci2, Tci3,..., Tcik. At this time, when comparing, compare all the predicted time differences of all users, select the smallest predicted time difference, and determine the user with leakage or fault according to this smallest predicted time difference. This method can expand the detection range, obtain more electricity consumption habits that may cause power outages, and avoid omissions in the detection results caused by users being in relatively rare or special electricity consumption habit states.

[0044] When more than two same-day fluctuation events are selected, the same-day fluctuation events are formed into a same-day fluctuation event sequence. Taking three same-day fluctuation events as an example, the formed same-day fluctuation event sequence is D3 - D2 - D1. Then, the above same-day fluctuation event sequence is input into the user's electricity consumption habit model. Then, the frequent sequence pattern containing the same-day fluctuation event sequence is determined by comparison. After that, D0 is found according to the obtained frequent sequence pattern, and Tb is determined.

[0045] In this way, it can be more accurately judged which electricity consumption habit mode the current user is in, more accurate prediction can be achieved, some computing time can be saved, the computing efficiency can be improved, and the power outage time can be shortened.

[0046] In addition to being able to detect the tripping faults caused by leakage, this system can also detect some tripping faults caused by short-circuit faults.

[0047] Based on the above principle, the present invention also provides a method for restoring power supply. It specifically includes the following steps.

[0048] Step 1: Extract the historical data before the power outage moment, and construct the user's electricity consumption habit model according to the obtained historical data. The detailed solution of this step can refer to the specific method for constructing the electricity consumption habit model for each user carried by the load switch in the above text.

[0049] Step 2: Obtain the power outage moment Tt and the same-day fluctuation event Dj before this moment.

[0050] Step 3: Input the same-day fluctuation event Dj into the user's electricity consumption habit model to obtain the predicted time difference Tci of this user.

[0051] Step 4: Sort the predicted time differences from small to large, and restore power supply to the users in the order from small to large.

[0052] The above method can be applied to a low-voltage distribution outgoing line leakage monitoring system, which includes a data acquisition system, a model construction system, and a prediction and analysis system; the data acquisition system obtains the user's electricity consumption load data; the model construction system constructs the user's electricity consumption habit model by using the acquired data; the prediction and analysis system obtains the power outage moment and the same-day fluctuation event according to the acquired data and calculates the predicted time difference according to the electricity consumption habit model; the prediction and analysis system selects the user with the smallest predicted time difference as the user causing the tripping fault based on the predicted time difference of each user.

[0053] In summary, the above are only the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Through the above description, relevant staff can make various changes and modifications without departing from the technical idea of the present invention. The technical scope of the present invention is not limited to the content in the specification. All equivalent changes and modifications in terms of the shape, structure, features and spirit described in the scope of the claims of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for identifying leakage users, characterized in that The following steps are involved: Step 1: Build a power usage habit model for each user served by the load switch; Step 2: Determine the load curve at the time of power outage and the set time period before the power outage, and obtain the daily fluctuation event Dj based on the load curve of the set time period obtained; Step 3: Input the daily fluctuation event Dj of the corresponding user into the power consumption habit model of the user to predict the interval time Tb of power fluctuation after the daily fluctuation event, and calculate the predicted time difference Tc=|Ta-Tb|, where Ta is the interval time from the daily fluctuation event Dj to the power outage time; Step 4: Repeat step 3 until the predicted time difference Tc of all users is obtained, and the user with the shortest predicted time difference is determined to be the leakage user; In step 1: S1.1 Collect users' daily load change data to form a data set; S1.2 performs power decomposition on the collected daily load change data and generates a fluctuation event Dwt for each power fluctuation; S1.3 Apply the PrefixSpan algorithm to all fluctuation events Dwt to mine frequent patterns of users and generate a power consumption habit model; The fluctuation event Dwt includes three data: fluctuation power Wd, fluctuation time Td and fluctuation direction Df; Select the first daily fluctuation event D1 and input it into the user electricity usage habit model; Obtain all frequent sequence patterns of the daily volatility event D1, and obtain a predicted time difference according to each frequent sequence pattern; Select two or more consecutive daily fluctuation events to form a daily fluctuation event sequence and input it into the user's electricity consumption habit model, and then calculate the predicted time difference after obtaining the corresponding frequent sequence pattern; The length of the set time period is 0.5 to 3 hours.

2. A method for restoring power supply after a fault trip, using a method for identifying leakage users as claimed in claim 1, characterized in that: Step 1: Build a power usage habit model for each user served by the load switch; Step 2: Obtain the power outage time Tt and the daily fluctuation event Dj before this time; Step 3: Input the daily fluctuation event Dj into the user's electricity usage habit model to obtain the user's predicted time difference Tc; Step 4: Sort the predicted time differences from small to large, and restore power supply to users in this order.

3. The method for restoring power supply after a fault trip according to claim 2, characterized in that: In step 1: S1.1 Collect users' daily load change data to form a data set; S1.2 performs power decomposition on the collected daily load change data and generates a fluctuation event Dwt for each power fluctuation; S1.3 applies the PrefixSpan algorithm to all fluctuation events Dwt to mine frequent patterns of users and generate a power consumption habit model.

4. A power distribution low-voltage outgoing line leakage monitoring system adopts a leakage user identification method as claimed in claim 1, characterized in that: It includes a data acquisition system, a model building system and a prediction and analysis system; the data acquisition system obtains the user's power load data; the model building system uses the collected data to build the user's power consumption habit model; the prediction and analysis system obtains the power outage time and the fluctuation event of the day according to the collected data and calculates the predicted time difference according to the power consumption habit model; the prediction and analysis system selects the user with the smallest predicted time difference according to the predicted time difference of each user and determines it as the user who caused the tripping fault.

Images