Fault monitoring system and method based on power collection terminal

Abstract

The present invention relates to the technical field of power fault detection, and discloses a fault monitoring system and method based on a power collection terminal, including: constructing a nonlinear relationship between power consumption and temperature by analyzing the user's historical power consumption data and weather temperature. First, the user's power consumption day is divided into multiple time periods, and a time interval threshold is set. In each time period, data is acquired through a power and temperature collection device, a scatter plot of power and temperature is constructed, and nonlinear data fitting is performed. Next, the similarity between the temperature nonlinear relationship equations of different time periods is calculated, the time periods are divided into sets, and a temperature retrieval set is established. By matching the temperature nonlinear relationship equation of the user's current time period, the corresponding power relationship set is retrieved, and based on the priority sorting and relationship similarity, it is determined whether there is power fluctuation and fault.

Description

Technical Field

[0001] The present invention relates to the technical field of power fault detection, and in particular to a fault monitoring system and method based on a power acquisition terminal. Background Art

[0002] With the popularization of smart grids and the rapid development of Internet of Things technology, power collection terminals are increasingly used in homes and industries. These devices can monitor users' power consumption in real time and transmit data to the cloud, providing strong support for the operation, maintenance and fault diagnosis of power systems. However, the complexity of the power system and the differences in users' power consumption habits still make the detection of power faults face many challenges. Traditional fault detection methods often rely on preset thresholds or rules, which are difficult to adapt to the ever-changing power consumption patterns and environmental factors, and are prone to false alarms or missed alarms. Therefore, it is of great practical significance to develop an intelligent monitoring method that can accurately and efficiently detect power faults and adapt to the power consumption characteristics of different users.

[0003] In order to solve the above problems, there is an urgent need for an intelligent monitoring method that can adapt to the power consumption characteristics of different users, take into account the influence of environmental factors, and distinguish between power fluctuations and faults. Traditional fault detection methods often find it difficult to meet these requirements at the same time, and methods based on nonlinear relationships and similarity analysis provide a new way to solve this problem. This method analyzes historical power consumption data and weather temperature data, constructs a nonlinear relationship between power consumption and weather temperature, and uses relationship similarity calculation to compare power consumption and weather temperature patterns in different time periods, thereby realizing intelligent detection of power faults.

[0004] Compared with traditional fault detection methods, the method based on nonlinear relationship and similarity analysis has obvious advantages. It is not only a simple fault detection tool, but also an intelligent system that can adapt to environmental changes, self-learn, and continuously improve monitoring accuracy and robustness. Summary of the invention

[0005] The present invention provides a fault monitoring system and method based on a power collection terminal, which promotes solving the problems mentioned in the above background technology.

[0006] The present invention provides the following technical solution: a fault monitoring method based on a power collection terminal, comprising:

[0007] Divide the user's historical electricity consumption into multiple electricity consumption days, and keep the electricity consumption days without power failure;

[0008] Set time interval threshold;

[0009] Divide each electricity consumption day into multiple time periods according to the time interval threshold;

[0010] Each time period includes a start time and an end time;

[0011] The starting time is the minimum time in the time period;

[0012] The end time is the maximum time in the time period;

[0013] For each time period, n time points are selected, and the time intervals between adjacent time points are the same;

[0014] The time points corresponding to each time period are set as 1, 2, 3, ... n in sequence, and the power consumption and weather temperature corresponding to each time point are obtained through the power collection device and the temperature collection device respectively;

[0015] For each time period, a power scatter plot is constructed through the time points and power consumption, and nonlinear data fitting is performed on the power consumption in the power scatter plot to obtain a nonlinear relationship about power consumption;

[0016] For each time period, a temperature scatter plot is constructed through the time point and weather temperature, and nonlinear data fitting is performed on the weather temperature in the temperature scatter plot to obtain a nonlinear relationship about the weather temperature;

[0017] Fault diagnosis is performed based on the nonlinear relationship between weather temperature and power consumption.

[0018] Optionally, the fault judgment according to the nonlinear relationship between weather temperature and power consumption includes:

[0019] Use the relationship similarity formula to calculate the relationship similarity between the nonlinear relationship between weather temperature in each time period;

[0020] Divide all time periods into multiple time period sets according to the relationship similarity, and establish a temperature retrieval set so that each element in the temperature retrieval set corresponds to a time period set;

[0021] After all time periods are divided, the time period in which the user is located is determined;

[0022] Obtain a nonlinear relationship between the weather temperature and the time period in which the user is located, and match the elements in the search set using the nonlinear relationship between the weather temperature and the time period in which the user is located.

[0023] The matched elements in the search set are retrieved to obtain the corresponding power relationship set, and the priority is calculated according to the start time and end time of the time period corresponding to each element in the power relationship set, as well as the weather temperature corresponding to the start time and end time. All elements in the power relationship set are sorted according to the priority for subsequent relationship similarity calculation;

[0024] Based on the similarity between the power relationship set and the nonlinear relationship expression of the power corresponding to the user in the current time period and the power consumption, it is determined whether there is power fluctuation and failure.

[0025] Optionally, for each time period, a power scatter plot is constructed by using time points and power consumption, and nonlinear data fitting is performed on the power consumption in the power scatter plot to obtain a nonlinear relationship about power consumption, including:

[0026] With each time point in the time period as the horizontal axis and the power consumption as the vertical axis, a coordinate axis showing the change of power consumption with time points is formed;

[0027] For each time period, the corresponding power consumption at each moment is marked on the coordinate axis to obtain a scatter plot of power consumption changing with time points;

[0028] The nonlinear relationship between electricity consumption and time points is determined by fitting the electricity consumption scatter plot through nonlinear regression;

[0029] The nonlinear relationship between the time points and the applied power in each time period is set as , x is the time point, y(x) is the power consumption at the time point x, a is the first parameter, b is the second parameter, and c is the third parameter;

[0030] Set the power consumption model prediction relationship as , used to calculate the residual sum of squares of applied electricity in each time period;

[0031] The residual sum of squares of the nonlinear relationship of each time period to the applied power is calculated by the following formula, as follows:

[0032] In the above formula, s is the residual sum of squares of electricity consumption, y(x) is the electricity consumption at time x on the electricity consumption scatter plot, a is the first parameter, b is the second parameter, and c is the third parameter. represents the i-th time point, i is the time point serial number, and m is the maximum value of the time point serial number;

[0033] For each time period, the time point is substituted into the above formula by the gradient descent method and the values ​​of a, b and c in each nonlinear relationship are adjusted to calculate the minimized residual sum of squares of electricity consumption in each time period, obtain the values ​​of a, b and c corresponding to the minimized residual sum of squares, and determine the nonlinear relationship in the nonlinear relationship corresponding to the electricity consumption in each time period.

[0034] Optionally, for each time period, a temperature scatter plot is constructed by using time points and weather temperatures, and nonlinear data fitting is performed on the weather temperatures in the temperature scatter plot to obtain a nonlinear relationship about the weather temperature, including:

[0035] The temperature collection device is used to obtain the weather temperature corresponding to each time point in each time period;

[0036] Take each time point in the time period as the horizontal axis and the weather temperature as the vertical axis to form a coordinate axis showing the change of weather temperature with time points;

[0037] For each time period, the weather temperature corresponding to each moment is marked on the coordinate axis to obtain a temperature scatter plot of the weather temperature changing with time points;

[0038] Set the nonlinear relationship between the time point and the corresponding weather temperature in each time period as , x is the time point, is the weather temperature at time x, is the first temperature parameter, is the second temperature parameter, is the third temperature parameter;

[0039] Set the temperature model prediction relationship , used to calculate the residual sum of squares corresponding to the weather temperature in each time period;

[0040] The residual sum of squares of the nonlinear relationship of each time period to the applied power is calculated by the following formula, as follows:

[0041] In the above formula, st is the residual sum of squares of weather temperature. is the weather temperature at time point x on the temperature scatter plot, is the first temperature parameter, is the second temperature parameter, is the third temperature parameter, represents the i-th time point, i is the time point serial number, and m is the maximum value of the time point serial number;

[0042] For each time period, the time points are substituted into the above formulas by the gradient descent method and the nonlinear relationship in each period is adjusted. , and The value of is used to calculate the minimum residual sum of squares, and the corresponding value of the minimum residual sum of squares is obtained. , and The value of determines the nonlinear relationship in the nonlinear relationship corresponding to the weather temperature in each time period.

[0043] Optionally, all time periods are divided into multiple time period sets according to the relationship similarity, and a temperature search set is established so that each element in the temperature search set corresponds to a time period set, including:

[0044] S1. Enter the nonlinear relationship equations corresponding to the weather temperature in all time periods into the temperature relationship set;

[0045] S2, randomly selecting any nonlinear relational expression in the temperature relation set as the initial relational expression;

[0046] Set a similarity threshold range to compare the similarity between the initial relationship and the remaining nonlinear relationship in the temperature relationship set;

[0047] S3, taking the remaining nonlinear relational expressions in the temperature relation set as comparison relational expressions, and calculating the relationship similarity between the initial relational expression and each comparison relational expression respectively through the relationship similarity formula;

[0048] S4, recording the relationship similarity between the initial relational expression and each comparison relational expression into a similarity set;

[0049] Mark all relationship similarities in the similarity set that are within the similarity threshold range as qualified similarities;

[0050] S5, entering the time periods corresponding to the nonlinear relational expressions of all qualified similarities and the time periods corresponding to the initial relational expressions into a time period set, and entering the initial relational expression into a temperature search set;

[0051] S6, eliminating all nonlinear relational expressions and initial relational expressions marked as corresponding to qualified similarities in the temperature relationship set;

[0052] S7, reselect any nonlinear relational expression in the temperature relationship set as a new initial relational expression, and repeat operations S2 to S6 until all elements in the temperature relationship set are traversed;

[0053] Get the maximum and minimum weather temperatures in the time period corresponding to the initial relationship, recorded as and ;

[0054] Randomly obtain the maximum and minimum weather temperatures in the time period corresponding to any nonlinear relationship in the temperature relationship set, which are recorded as and ;

[0055] pass Calculate the temperature difference and enter the temperature difference set;

[0056] Reselect any nonlinear relationship in the temperature relationship set and calculate the maximum and minimum weather temperatures in the corresponding time period of the nonlinear relationship. and The temperature difference is calculated until all nonlinear relationships in the temperature relationship set are traversed;

[0057] Sort all elements in the temperature difference set according to the temperature difference, from large to small;

[0058] The weight value of each nonlinear relationship in the temperature relationship set is calculated by the following formula:

[0059] In the above formula, G is the weight of the nonlinear relationship corresponding to the i-th temperature difference in the temperature difference set. is the i-th temperature difference in the temperature difference set, and h is the number of elements in the temperature difference set.

[0060] Optionally, the method of using the remaining nonlinear relational expressions in the temperature relationship set as comparison relational expressions, and calculating the relationship similarity between the initial relational expression and each comparison relational expression respectively by using a relationship similarity formula, includes:

[0061] The relationship similarity formula is as follows:

[0062] In the above formula, Cs is the relationship similarity, n is the time point in the time period, is the initial relation, is the contrast relation, G is the weight corresponding to the contrast relation, is the i-th time point.

[0063] Optionally, obtaining a nonlinear relationship formula of the weather temperature corresponding to the time period in which the user is located, matching elements in the search set through the nonlinear relationship formula of the weather temperature, and obtaining a power relationship set according to the matching elements, includes:

[0064] Get the user's current time period and record it as the target time period;

[0065] Obtain the weather temperature corresponding to each time point in the target time period through a temperature collection device;

[0066] According to each time point in the target time period and the weather temperature corresponding to each time point, a nonlinear relationship formula of the weather temperature corresponding to the target time period is obtained, which is recorded as a target temperature relationship formula;

[0067] The relationship similarity between the target temperature relationship and each nonlinear relationship in the temperature search set is calculated by using the relationship similarity formula, and the corresponding nonlinear relationship with the largest relationship similarity is selected as the search relationship;

[0068] A time period set corresponding to the search relational expression is obtained, and a nonlinear relational expression of each time period in the time period set to the applied power is entered into the power relational set.

[0069] Optionally, the method of obtaining the corresponding power relationship set by searching the matched elements in the set, calculating the priority according to the start time and end time of the time period corresponding to each element in the power relationship set, and the weather temperature corresponding to the start time and end time, and sorting all elements in the power relationship set according to the priority for subsequent relationship similarity calculation, includes:

[0070] The priority of the elements in the power relationship set is calculated by the following formula for sorting, as follows:

[0071] In the above formula, is the target time period, and are the start and end time of the target time period, R is the time constant, To retrieve the elements in the time period set corresponding to the relation, and are the start time and end time of the jth time period in the time period set corresponding to the retrieval relation, To retrieve the temperature change between the start and end time of the jth time period in the time period set corresponding to the relationship, is the temperature change between the start time and the end time in the target time period, and for The maximum and minimum values ​​of , P is the element priority;

[0072] The nonlinear relationship expressions in the power relationship set are sorted from large to small according to the element priorities of their corresponding time periods.

[0073] Optionally, judging whether there is power fluctuation and failure according to the similarity between the power relationship set and the nonlinear relationship expression of the power corresponding to the user in the current time period and the power consumption includes:

[0074] Select the minimum power consumption and the maximum power consumption among all nonlinear relationship expressions in the power relationship set to form a reasonable power consumption range;

[0075] Obtain the user's electricity consumption at each time point in the target time period in real time;

[0076] If the power consumption corresponding to each time point in the target time period is not within the reasonable power range, the user is marked as having a power failure in the target time period. Otherwise, the nonlinear relationship between the target time period and the applied power is obtained based on each time point in the target time period and the power consumption corresponding to each time point, which is recorded as the target power relationship.

[0077] Use the relationship similarity formula to calculate the relationship similarity between the target power relationship formula and each nonlinear relationship formula in the power relationship set, and enter the similarity set;

[0078] If there is a relationship similarity within the similarity threshold range in the similarity set, the power consumption in the adjacent time period behind the target time period is detected in real time;

[0079] Otherwise, the user is marked as having power fluctuations during the target time period.

[0080] Optionally, a system for implementing the fault monitoring method based on a power collection terminal includes:

[0081] The power collection terminal includes a power collection device and a temperature collection device;

[0082] An electricity collection device, used to collect the electricity consumption of users;

[0083] Temperature collection device, used to collect daily temperature changes;

[0084] A relationship building module, which builds nonlinear relationships between power consumption and temperature according to the user's power consumption and the corresponding temperature changes;

[0085] The data processing module calculates the relationship similarity through the relationship similarity formula, and groups the nonlinear relationship of the user's power consumption according to the nonlinear relationship corresponding to the temperature change according to the relationship similarity;

[0086] The data retrieval module retrieves the corresponding power quantity relationship set by calculating the temperature nonlinear relationship of the user's current time period and the temperature nonlinear relationship corresponding to each time period;

[0087] The similarity judgment module uses the relationship similarity formula to traverse and calculate the relationship similarity between the nonlinear relationship of the power of the user's current time period and each element in the corresponding power relationship set.

[0088] The present invention has the following beneficial effects:

[0089] 1. Draw a scatter plot of electricity consumption by taking each time point in the time period as the horizontal axis and the corresponding electricity consumption as the vertical axis. Then, fit these scatter points through nonlinear regression to obtain a nonlinear relationship, in which electricity consumption y(x) is a function of time point x, and the function contains three parameters a, b and c. In order to determine the best parameters a, b and c, the gradient descent method is used to evaluate the degree of model fitting by calculating the residual sum of squares s, and the parameter values ​​are continuously adjusted until the minimum value of the residual sum of squares is found. At this time, the values ​​of a, b and c are the best parameters for the time period, and the nonlinear relationship corresponding to the electricity consumption in the time period is determined; the advantage of this is that the nonlinear pattern of electricity consumption changing over time can be accurately modeled. Through nonlinear regression, the trend of electricity consumption can be better captured, rather than simply assuming a linear relationship, so that the model can more accurately reflect the actual electricity consumption. The use of the gradient descent method can effectively find the optimal nonlinear model parameters, minimize the model fitting error, and thus improve the accuracy of the model. At the same time, fitting each time period separately instead of using a unified model can better adapt to the differences in electricity consumption habits in different time periods, making the model more adaptive. This approach provides a more accurate baseline model of power consumption for subsequent power fault detection.

[0090] 2. Obtain the maximum and minimum weather temperatures for the time period corresponding to the initial relationship, recorded as and Then, for each comparative relationship in the temperature relationship set, obtain the maximum and minimum weather temperatures in the corresponding time period, recorded as and . Next, calculate the sum of the difference between the maximum and minimum weather temperatures of the initial relation and the maximum and minimum weather temperatures of each comparison relation, and enter these sums into a temperature difference set. This process will traverse all the comparison relations in the temperature relationship set. Then, sort the elements in the temperature difference set from large to small according to the temperature difference. Finally, according to the sorted temperature difference, use the formula to calculate the weight value of each comparison relation. The advantage of this is that when calculating the relationship similarity, the influence of the temperature change amplitude in different time periods on the similarity can be taken into account, making the similarity calculation more accurate and reasonable. By introducing weights, time periods with larger temperature changes can have higher weights when calculating similarity, so that more attention can be paid to those time periods with more drastic temperature changes. These time periods are usually more representative and can better reflect the overall temperature change law. At the same time, sorting the temperature difference can make it easier to calculate the weight value, making the calculation of the weight value more efficient.

[0091] 3. By calculating the relationship similarity between the initial relationship and each comparison relationship through the relationship similarity formula, the temperature changes in the temperature relationship set that are similar to the initial relationship can be accurately obtained, which reduces the amount of data for subsequent retrieval operations and improves the calculation rate. In the above formula, by introducing the weight G factor, the accuracy of data screening can be improved, making data screening more in line with actual conditions, and at the same time able to adapt to the impact of other weather on temperature. For example, when the weather temperature reaches the maximum value in a period of time, followed by a short period of rainfall, the weather temperature changes dramatically. After the rain stops, the weather temperature returns to the lowest temperature. Therefore, if the calculation is performed through the previous relationship similarity, this situation will be excluded, resulting in incomplete subsequent statistics on the data. After the weight G factor is introduced, the data corresponding to this situation can be taken into account without affecting other data under normal weather conditions.

[0092] 4. Obtain the time period in which the user is currently located and record it as the target time period. Then, the weather temperature at each time point in the target time period is obtained through the temperature collection device, and the nonlinear relationship of the weather temperature corresponding to the target time period is fitted using these temperature data, which is recorded as the target temperature relationship. Next, the similarity between the target temperature relationship and each relationship in the temperature retrieval set is calculated, and the relationship with the highest similarity is selected as the retrieval relationship. Finally, the time period set corresponding to the retrieval relationship is obtained, and the nonlinear relationship of the power consumption corresponding to all time periods in the set is entered into the power relationship set; the advantage of this is that according to the temperature change law of the current time period, the time period with similar temperature change law in history can be accurately found, and the power consumption data of these historical time periods can be used for subsequent power fault detection. By comparing the similarity of the current target temperature relationship with all the relationships in the temperature retrieval set, the most suitable match can be found, ensuring the accuracy and reliability of the subsequent power consumption data selection. The information in the temperature retrieval set can be used to quickly locate the relevant time period set, thereby avoiding the inefficient search of traversing all historical data. Furthermore, by acquiring the power consumption data of the time period set corresponding to the retrieval relation, a group of representative reference data can be provided for subsequent power fault detection.

[0093] 5. By using a formula to calculate the priority P of each element, the formula takes into account the start and end times of the target time period, as well as the temperature changes at these time points, and sorts the elements in the power relationship set from large to small according to the priority P of each element; the advantage of this is that the elements in the power relationship set can be effectively processed according to the temperature change characteristics of the target time period, so that in the subsequent similarity calculation, those elements that are more similar to the temperature change pattern of the target time period are given priority. By introducing the temperature change factor as the basis for calculating the priority, the sorting can be made more reasonable, thereby avoiding the deviation caused by simply sorting according to time sequence or other subjective factors. This pre-sorting operation makes the subsequent similarity calculation more efficient and more accurate, because it gives priority to the temperature changes corresponding to the start and end times, so that power fluctuations and faults can be detected more effectively.

[0094] 6. A reasonable power interval is defined by the minimum and maximum power consumption of all nonlinear relationships in the power relationship set. Then, the power consumption of the user in the target time period is monitored in real time. If the power consumption at all time points in the target time period is not within the reasonable interval, it is directly determined to be a power failure. Otherwise, the target power relationship is fitted according to the power consumption data of the target time period. Then, the similarity between the target power relationship and each relationship in the power relationship set is calculated and recorded in a similarity set. If there is a similarity value within the similarity threshold range in the similarity set, it is considered that the current power consumption mode is similar to the history, and the system will continue to detect the power consumption in subsequent time periods; if all similarities are not within the threshold range, it is considered that the current power consumption mode does not match the history, and it is marked as power fluctuation; the advantage of this is that a multi-level power fluctuation and fault detection mechanism is established, which takes into account both the absolute value of power consumption and the relative change of power consumption mode. By setting a reasonable power interval, it is possible to quickly determine whether there is an obvious power consumption anomaly, thereby quickly eliminating some obvious fault conditions. By calculating the similarity between the target power relationship and the power relationship set, it is possible to determine whether the current power consumption pattern conforms to historical rules, thereby more accurately detecting power fluctuations. Using relationship similarity analysis, subtle changes in power consumption patterns can be better discovered, rather than relying solely on the absolute size of power consumption for judgment. This hierarchical judgment strategy improves the efficiency and accuracy of fault detection, and can more effectively identify power fluctuations and faults, rather than generalizing, providing a more comprehensive power system monitoring capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 It is a schematic diagram of the process of the present invention.

[0096] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0097] 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.

[0098] Example 1, see Figure 1 to Figure 2 , a fault monitoring method based on a power collection terminal includes:

[0099] Divide the user's historical electricity consumption into multiple electricity consumption days, and keep the electricity consumption days without power failure;

[0100] Set time interval threshold;

[0101] Divide each electricity consumption day into multiple time periods according to the time interval threshold;

[0102] Each time period includes a start time and an end time;

[0103] The starting time is the minimum time in the time period;

[0104] The end time is the maximum time in the time period;

[0105] For each time period, n time points are selected, and the time intervals between adjacent time points are the same;

[0106] The time points corresponding to each time period are set as 1, 2, 3, ... n in sequence, and the power consumption and weather temperature corresponding to each time point are obtained through the power collection device and the temperature collection device respectively;

[0107] For each time period, a power scatter plot is constructed through the time points and power consumption, and nonlinear data fitting is performed on the power consumption in the power scatter plot to obtain a nonlinear relationship about power consumption;

[0108] For each time period, a temperature scatter plot is constructed through the time point and weather temperature, and nonlinear data fitting is performed on the weather temperature in the temperature scatter plot to obtain a nonlinear relationship about the weather temperature;

[0109] Fault diagnosis is performed based on the nonlinear relationship between weather temperature and power consumption.

[0110] The fault judgment according to the nonlinear relationship between weather temperature and power consumption includes:

[0111] Use the relationship similarity formula to calculate the relationship similarity between the nonlinear relationship between weather temperature in each time period;

[0112] Divide all time periods into multiple time period sets according to the relationship similarity, and establish a temperature retrieval set so that each element in the temperature retrieval set corresponds to a time period set;

[0113] After all time periods are divided, the time period in which the user is located is determined;

[0114] Obtain a nonlinear relationship between the weather temperature and the time period in which the user is located, and match the elements in the search set using the nonlinear relationship between the weather temperature and the time period in which the user is located.

[0115] The matched elements in the search set are retrieved to obtain the corresponding power relationship set, and the priority is calculated according to the start time and end time of the time period corresponding to each element in the power relationship set, as well as the weather temperature corresponding to the start time and end time. All elements in the power relationship set are sorted according to the priority for subsequent relationship similarity calculation;

[0116] Based on the similarity between the power relationship set and the nonlinear relationship expression of the power corresponding to the user in the current time period and the power consumption, it is determined whether there is power fluctuation and failure.

[0117] For each time period, a power scatter plot is constructed through time points and power consumption, and nonlinear data fitting is performed on the power consumption in the power scatter plot to obtain a nonlinear relationship about power consumption, including:

[0118] With each time point in the time period as the horizontal axis and the power consumption as the vertical axis, a coordinate axis showing the change of power consumption with time points is formed;

[0119] For each time period, the corresponding power consumption at each moment is marked on the coordinate axis to obtain a scatter plot of power consumption changing with time points;

[0120] The nonlinear relationship between electricity consumption and time points is determined by fitting the electricity consumption scatter plot through nonlinear regression;

[0121] The nonlinear relationship between the time points and the applied power in each time period is set as , x is the time point, y(x) is the power consumption at the time point x, a is the first parameter, b is the second parameter, and c is the third parameter;

[0122] Set the power consumption model prediction relationship as , used to calculate the residual sum of squares of applied electricity in each time period;

[0123] The residual sum of squares of the nonlinear relationship of each time period to the applied power is calculated by the following formula, as follows:

[0124] In the above formula, s is the residual sum of squares of electricity consumption, y(x) is the electricity consumption at time x on the electricity consumption scatter plot, a is the first parameter, b is the second parameter, and c is the third parameter. represents the i-th time point, i is the time point serial number, and m is the maximum value of the time point serial number;

[0125] For each time period, the time points are substituted into the above formula by the gradient descent method and the values ​​of a, b and c in each nonlinear relationship are adjusted to calculate the minimized residual sum of squares of electricity consumption in each time period, and the values ​​of a, b and c corresponding to the minimized residual sum of squares are obtained to determine the nonlinear relationship in the nonlinear relationship corresponding to the electricity consumption in each time period;

[0126] By first taking each time point in the time period as the horizontal axis and the corresponding power consumption as the vertical axis, a scatter plot of power consumption is drawn. Then, these scatter points are fitted by nonlinear regression to obtain a nonlinear relationship, in which the power consumption y(x) is a function of the time point x, and the function contains three parameters a, b and c. In order to determine the best parameters a, b and c, the gradient descent method is used to evaluate the degree of model fitting by calculating the residual sum of squares s, and the parameter values ​​are continuously adjusted until the minimum value of the residual sum of squares is found. At this time, the values ​​of a, b and c are the best parameters for the time period, and the nonlinear relationship corresponding to the power consumption of the time period is determined; the advantage of this is that the nonlinear pattern of power consumption changing over time can be accurately modeled. Through nonlinear regression, the trend of power consumption can be better captured, rather than simply assuming a linear relationship, so that the model can more accurately reflect the actual power consumption. The use of the gradient descent method can effectively find the optimal nonlinear model parameters, minimize the model fitting error, and thus improve the accuracy of the model. At the same time, fitting each time period separately instead of using a unified model can better adapt to the differences in power consumption habits in different time periods, making the model more adaptive. This approach provides a more accurate baseline model of power consumption for subsequent power fault detection.

[0127] For each time period, a temperature scatter plot is constructed by using the time point and the weather temperature, and nonlinear data fitting is performed on the weather temperature in the temperature scatter plot to obtain a nonlinear relationship about the weather temperature, including:

[0128] The temperature collection device is used to obtain the weather temperature corresponding to each time point in each time period;

[0129] Take each time point in the time period as the horizontal axis and the weather temperature as the vertical axis to form a coordinate axis showing the change of weather temperature with time points;

[0130] For each time period, the weather temperature corresponding to each moment is marked on the coordinate axis to obtain a temperature scatter plot of the weather temperature changing with time points;

[0131] Set the nonlinear relationship between the time point and the corresponding weather temperature in each time period as , x is the time point, is the weather temperature at time x, is the first temperature parameter, is the second temperature parameter, is the third temperature parameter;

[0132] Set the temperature model prediction relationship , used to calculate the residual sum of squares corresponding to the weather temperature in each time period;

[0133] The residual sum of squares of the nonlinear relationship of each time period to the applied power is calculated by the following formula, as follows:

[0134] In the above formula, st is the residual sum of squares of weather temperature. is the weather temperature at time point x on the temperature scatter plot, is the first temperature parameter, is the second temperature parameter, is the third temperature parameter, represents the i-th time point, i is the time point serial number, and m is the maximum value of the time point serial number;

[0135] For each time period, the time points are substituted into the above formulas by the gradient descent method and the nonlinear relationship in each period is adjusted. , and The value of is used to calculate the minimum residual sum of squares, and the corresponding value of the minimum residual sum of squares is obtained. , and The value of determines the nonlinear relationship in the nonlinear relationship corresponding to the weather temperature in each time period.

[0136] The method divides all time periods into multiple time period sets according to the relationship similarity, and establishes a temperature search set so that each element in the temperature search set corresponds to a time period set, including:

[0137] S1. Enter the nonlinear relationship equations corresponding to the weather temperature in all time periods into the temperature relationship set;

[0138] S2, randomly selecting any nonlinear relational expression in the temperature relation set as the initial relational expression;

[0139] Set a similarity threshold range to compare the similarity between the initial relationship and the remaining nonlinear relationship in the temperature relationship set;

[0140] S3, taking the remaining nonlinear relational expressions in the temperature relation set as comparison relational expressions, and calculating the relationship similarity between the initial relational expression and each comparison relational expression respectively through the relationship similarity formula;

[0141] S4, recording the relationship similarity between the initial relational expression and each comparison relational expression into a similarity set;

[0142] Mark all relationship similarities in the similarity set that are within the similarity threshold range as qualified similarities;

[0143] S5, entering the time periods corresponding to the nonlinear relational expressions of all qualified similarities and the time periods corresponding to the initial relational expressions into a time period set, and entering the initial relational expression into a temperature search set;

[0144] S6, eliminating all nonlinear relational expressions and initial relational expressions marked as corresponding to qualified similarities in the temperature relationship set;

[0145] S7, reselect any nonlinear relational expression in the temperature relationship set as a new initial relational expression, and repeat operations S2 to S6 until all elements in the temperature relationship set are traversed;

[0146] First, the nonlinear weather temperature relations corresponding to all time periods are placed in the temperature relation set. Then, a relation is randomly selected as the initial relation, and a similarity threshold range is set. Next, the remaining relation in the temperature relation set are used as comparison relations, and their relationship similarities with the initial relation are calculated one by one. If the calculated similarity value is within the preset threshold range, the time periods corresponding to these comparison relations and the initial relations are classified into the same time period set, and the initial relations are placed in the temperature retrieval set as the index of the time period set. After that, the already clustered relations are removed from the temperature relation set. Repeat the above steps until all relations in the temperature relation set have been processed; the advantage of doing so is that time periods with similar temperature change laws can be classified together, so that the electricity consumption pattern can be predicted more accurately. Through the relationship similarity calculation, the similarity of temperature changes between different time periods can be quantified, making the clustering process more objective and scientific, and avoiding errors that may be caused by subjective division. The establishment of the temperature retrieval set enables us to quickly retrieve similar historical time periods based on the temperature change pattern of the current time period, and then use the power consumption data of these historical time periods to detect power failures, improving the efficiency and accuracy of retrieval. In addition, this clustering method also enables us to learn and use historical experience, so as to better adapt to the impact of ambient temperature changes on user power consumption patterns.

[0147] Get the maximum and minimum weather temperatures in the time period corresponding to the initial relationship, recorded as and ;

[0148] Randomly obtain the maximum and minimum weather temperatures in the time period corresponding to any nonlinear relationship in the temperature relationship set, which are recorded as and ;

[0149] pass Calculate the temperature difference and enter the temperature difference set;

[0150] Reselect any nonlinear relationship in the temperature relationship set and calculate the maximum and minimum weather temperatures in the corresponding time period of the nonlinear relationship. and The temperature difference is calculated until all nonlinear relationships in the temperature relationship set are traversed;

[0151] Sort all elements in the temperature difference set according to the temperature difference, from large to small;

[0152] The weight value of each nonlinear relationship in the temperature relationship set is calculated by the following formula:

[0153] In the above formula, G is the weight of the nonlinear relationship corresponding to the i-th temperature difference in the temperature difference set. is the i-th temperature difference in the temperature difference set, and h is the number of elements in the temperature difference set.

[0154] By obtaining the maximum and minimum weather temperatures for the time period corresponding to the initial relationship, recorded as and Then, for each comparative relationship in the temperature relationship set, obtain the maximum and minimum weather temperatures in the corresponding time period, recorded as and . Next, calculate the sum of the difference between the maximum and minimum weather temperatures of the initial relation and the maximum and minimum weather temperatures of each comparison relation, and enter these sums into a temperature difference set. This process will traverse all the comparison relations in the temperature relationship set. Then, sort the elements in the temperature difference set from large to small according to the temperature difference. Finally, according to the sorted temperature difference, use the formula to calculate the weight value of each comparison relation. The advantage of this is that when calculating the relationship similarity, the influence of the temperature change amplitude in different time periods on the similarity can be taken into account, making the similarity calculation more accurate and reasonable. By introducing weights, time periods with larger temperature changes can have higher weights when calculating similarity, so that more attention can be paid to those time periods with more drastic temperature changes. These time periods are usually more representative and can better reflect the overall temperature change law. At the same time, sorting the temperature difference can make it easier to calculate the weight value, making the calculation of the weight value more efficient.

[0155] The method of using the remaining nonlinear relational expressions in the temperature relation set as comparison relational expressions and calculating the relationship similarity between the initial relational expression and each comparison relational expression by using a relationship similarity formula includes:

[0156] The relationship similarity formula is as follows:

[0157] In the above formula, Cs is the relationship similarity, n is the time point in the time period, is the initial relation, is the contrast relation, G is the weight corresponding to the contrast relation, is the i-th time point.

[0158] By calculating the relationship similarity between the initial relationship and each comparison relationship through the relationship similarity formula, the temperature changes in the temperature relationship set that are similar to the initial relationship can be accurately obtained, which reduces the amount of data for subsequent retrieval operations and improves the calculation rate. In the above formula, by introducing the weight G factor, the accuracy of data screening can be improved, making data screening more in line with actual conditions, and at the same time able to adapt to the impact of other weather on temperature. For example, when the weather temperature reaches the maximum value in a period of time, followed by a short period of rainfall, the weather temperature changes dramatically, and after the rain stops, the weather temperature returns to the lowest temperature. Therefore, if the previous relationship similarity is used for calculation, this situation will be excluded, resulting in incomplete subsequent statistics on the data. After the weight G factor is introduced, the data corresponding to this situation can be taken into account without affecting other data under normal weather conditions.

[0159] The step of obtaining a nonlinear relationship between the weather temperature and the time period in which the user is located, matching elements in a search set by using the nonlinear relationship between the weather temperature, and obtaining a power relationship set according to the matching elements includes:

[0160] Get the user's current time period and record it as the target time period;

[0161] Obtain the weather temperature corresponding to each time point in the target time period through a temperature collection device;

[0162] According to each time point in the target time period and the weather temperature corresponding to each time point, a nonlinear relationship formula of the weather temperature corresponding to the target time period is obtained, which is recorded as a target temperature relationship formula;

[0163] The relationship similarity between the target temperature relationship and each nonlinear relationship in the temperature search set is calculated by using the relationship similarity formula, and the corresponding nonlinear relationship with the largest relationship similarity is selected as the search relationship;

[0164] A time period set corresponding to the search relational expression is obtained, and a nonlinear relational expression of each time period in the time period set to the applied power is entered into the power relational set.

[0165] The method obtains the corresponding power relationship set by searching the matched elements in the set, calculates the priority according to the start time and end time of the time period corresponding to each element in the power relationship set, and the weather temperature corresponding to the start time and the end time, and sorts all elements in the power relationship set according to the priority for subsequent relationship similarity calculation, including:

[0166] The priority of the elements in the power relationship set is calculated by the following formula for sorting, as follows:

[0167] In the above formula, is the target time period, and are the start and end time of the target time period, R is the time constant, To retrieve the elements in the time period set corresponding to the relation, and are the start time and end time of the jth time period in the time period set corresponding to the retrieval relation, To retrieve the temperature change between the start and end time of the jth time period in the time period set corresponding to the relationship, is the temperature change between the start time and the end time in the target time period, and for The maximum and minimum values ​​of , P is the element priority;

[0168] The nonlinear relationship expressions in the power relationship set are sorted from large to small according to the element priorities of their corresponding time periods.

[0169] By using a formula to calculate the priority P of each element, the formula takes into account the start and end times of the target time period, as well as the temperature changes at these time points, and sorts the elements in the power relationship set from large to small according to the priority P of each element; the advantage of this is that the elements in the power relationship set can be effectively processed according to the temperature change characteristics of the target time period, so that in the subsequent similarity calculation, those elements that are more similar to the temperature change pattern of the target time period are given priority. By introducing the temperature change factor as the basis for calculating the priority, the sorting can be made more reasonable, thereby avoiding the deviation caused by simply sorting according to time sequence or other subjective factors. This pre-sorting operation makes the subsequent similarity calculation more efficient and more accurate, because it gives priority to the temperature changes corresponding to the start and end times, so that the nonlinear relationship of power corresponding to the relationship similarity in the reasonable power range can be detected more effectively and timely, thereby improving the efficiency of detecting power fluctuations and faults.

[0170] The determining whether there is power fluctuation and failure according to the power consumption and the power consumption of the power relationship set and the nonlinear power relationship of the user in the current time period, includes:

[0171] Select the minimum power consumption and the maximum power consumption among all nonlinear relationship expressions in the power relationship set to form a reasonable power consumption range;

[0172] Obtain the user's electricity consumption at each time point in the target time period in real time;

[0173] If the power consumption corresponding to each time point in the target time period is not within the reasonable power range, the user is marked as having a power failure in the target time period. Otherwise, the nonlinear relationship between the target time period and the applied power is obtained based on each time point in the target time period and the power consumption corresponding to each time point, which is recorded as the target power relationship.

[0174] Use the relationship similarity formula to calculate the relationship similarity between the target power relationship formula and each nonlinear relationship formula in the power relationship set, and enter the similarity set;

[0175] If there is a relationship similarity within the similarity threshold range in the similarity set, the power consumption in the adjacent time period behind the target time period is detected in real time;

[0176] Otherwise, the user is marked as having power fluctuations during the target time period.

[0177] Through the minimum and maximum power consumption of all nonlinear relational expressions in the power relationship set, a reasonable power interval is delineated. Then, the power consumption of the user in the target time period is monitored in real time. If the power consumption at all time points in the target time period is not within the reasonable interval, it is directly determined as a power failure. Otherwise, the target power relationship is fitted according to the power consumption data of the target time period. Then, the similarity between the target power relationship and each relationship in the power relationship set is calculated and recorded in a similarity set. If there is a similarity value within the similarity threshold range in the similarity set, it is considered that the current power consumption mode is similar to the history, and the system will continue to detect the power consumption in subsequent time periods; if all similarities are not within the threshold range, it is considered that the current power consumption mode does not match the history, and it is marked as power fluctuation; the advantage of this is that a multi-level power fluctuation and fault detection mechanism is established, which takes into account both the absolute value of power consumption and the relative change of power consumption mode. By setting a reasonable power interval, it is possible to quickly determine whether there is an obvious power consumption anomaly, thereby quickly eliminating some obvious fault conditions. By calculating the similarity between the target power relationship and the power relationship set, it is possible to determine whether the current power consumption pattern conforms to historical rules, thereby more accurately detecting power fluctuations. Using relationship similarity analysis, subtle changes in power consumption patterns can be better discovered, rather than relying solely on the absolute size of power consumption for judgment. This hierarchical judgment strategy improves the efficiency and accuracy of fault detection, and can more effectively identify power fluctuations and faults, rather than generalizing, providing a more comprehensive power system monitoring capability.

[0178] Example 2, see Figure 1 to Figure 2 , a system for implementing the fault monitoring method based on a power collection terminal, comprising:

[0179] The power collection terminal includes a power collection device and a temperature collection device;

[0180] An electricity collection device, used to collect the electricity consumption of users;

[0181] Temperature collection device, used to collect daily temperature changes;

[0182] A relationship building module, which builds nonlinear relationships between power consumption and temperature according to the user's power consumption and the corresponding temperature changes;

[0183] The data processing module calculates the relationship similarity through the relationship similarity formula, and groups the nonlinear relationship of the user's power consumption according to the nonlinear relationship corresponding to the temperature change according to the relationship similarity;

[0184] The data retrieval module retrieves the corresponding power quantity relationship set by calculating the temperature nonlinear relationship of the user's current time period and the temperature nonlinear relationship corresponding to each time period;

[0185] The similarity judgment module uses the relationship similarity formula to traverse and calculate the relationship similarity between the nonlinear relationship of the power in the user's current time period and each element in the corresponding power relationship set.

[0186] 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.

[0187] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A fault monitoring method based on a power collection terminal, characterized in that: include: Divide the user's historical electricity consumption into multiple electricity consumption days, and keep the electricity consumption days without power failure; Set time interval threshold; Divide each electricity consumption day into multiple time periods according to the time interval threshold; Each time period includes a start time and an end time; The starting time is the minimum time in the time period; The end time is the maximum time in the time period; For each time period, n time points are selected, and the time intervals between adjacent time points are the same; The time points corresponding to each time period are set as 1, 2, 3, ... n in sequence, and the power consumption and weather temperature corresponding to each time point are obtained through the power collection device and the temperature collection device respectively; For each time period, a power scatter plot is constructed through the time points and power consumption, and nonlinear data fitting is performed on the power consumption in the power scatter plot to obtain a nonlinear relationship about power consumption; For each time period, a temperature scatter plot is constructed through the time point and weather temperature, and nonlinear data fitting is performed on the weather temperature in the temperature scatter plot to obtain a nonlinear relationship about the weather temperature; Fault diagnosis is performed based on the nonlinear relationship between weather temperature and power consumption, including: Use the relationship similarity formula to calculate the relationship similarity between the nonlinear relationship between weather temperature in each time period; Divide all time periods into multiple time period sets according to the relationship similarity, and establish a temperature retrieval set so that each element in the temperature retrieval set corresponds to a time period set; After all time periods are divided, the time period in which the user is located is determined; Obtain a nonlinear relationship expression for the weather temperature corresponding to the time period in which the user is located, match the elements in the search set through the nonlinear relationship expression for the weather temperature, and obtain a power relationship set based on the matching elements; The matched elements in the search set are retrieved to obtain the corresponding power relationship set, and the priority is calculated according to the start time and end time of the time period corresponding to each element in the power relationship set, as well as the weather temperature corresponding to the start time and end time. All elements in the power relationship set are sorted according to the priority for subsequent relationship similarity calculation; Based on the similarity between the power relationship set and the nonlinear relationship expression of the power corresponding to the user in the current time period and the power consumption, it is determined whether there is power fluctuation and failure.

2. The fault monitoring method based on the power collection terminal according to claim 1 is characterized in that: For each time period, a power scatter plot is constructed through time points and power consumption, and nonlinear data fitting is performed on the power consumption in the power scatter plot to obtain a nonlinear relationship about power consumption, including: With each time point in the time period as the horizontal axis and the power consumption as the vertical axis, a coordinate axis showing the change of power consumption with time points is formed; For each time period, the corresponding power consumption at each moment is marked on the coordinate axis to obtain a scatter plot of power consumption changing with time points; The nonlinear relationship between electricity consumption and time points is determined by fitting the electricity consumption scatter plot through nonlinear regression; The nonlinear relationship between the time points and the applied power in each time period is set as , x is the time point, y(x) is the power consumption at the time point x, a is the first parameter, b is the second parameter, and c is the third parameter; Set the power consumption model prediction relationship as , used to calculate the residual sum of squares of applied electricity in each time period; The residual sum of squares of the nonlinear relationship of each time period to the applied power is calculated by the following formula, as follows: In the above formula, s is the residual sum of squares of electricity consumption, y(x) is the electricity consumption at time x on the electricity consumption scatter plot, a is the first parameter, b is the second parameter, and c is the third parameter. represents the i-th time point, i is the time point serial number, and m is the maximum value of the time point serial number; For each time period, the time point is substituted into the above formula by the gradient descent method and the values ​​of a, b and c in each nonlinear relationship are adjusted to calculate the minimized residual sum of squares of electricity consumption in each time period, obtain the values ​​of a, b and c corresponding to the minimized residual sum of squares, and determine the nonlinear relationship in the nonlinear relationship corresponding to the electricity consumption in each time period.

3. The fault monitoring method based on the power collection terminal according to claim 1 is characterized in that: For each time period, a temperature scatter plot is constructed by using the time point and the weather temperature, and nonlinear data fitting is performed on the weather temperature in the temperature scatter plot to obtain a nonlinear relationship about the weather temperature, including: The temperature collection device is used to obtain the weather temperature corresponding to each time point in each time period; Take each time point in the time period as the horizontal axis and the weather temperature as the vertical axis to form a coordinate axis showing the change of weather temperature with time points; For each time period, the weather temperature corresponding to each moment is marked on the coordinate axis to obtain a temperature scatter plot of the weather temperature changing with time points; Set the nonlinear relationship between the time point and the corresponding weather temperature in each time period as , x is the time point, is the weather temperature at time x, is the first temperature parameter, is the second temperature parameter, is the third temperature parameter; Set the temperature model prediction relationship , used to calculate the residual sum of squares corresponding to the weather temperature in each time period; The residual sum of squares of the nonlinear relationship of each time period to the applied power is calculated by the following formula, as follows: In the above formula, st is the residual sum of squares of weather temperature. is the weather temperature at time point x on the temperature scatter plot, is the first temperature parameter, is the second temperature parameter, is the third temperature parameter, represents the i-th time point, i is the time point serial number, and m is the maximum value of the time point serial number; For each time period, the time points are substituted into the above formulas by the gradient descent method and the nonlinear relationship in each period is adjusted. , and The value of is used to calculate the minimum residual sum of squares, and the corresponding value of the minimum residual sum of squares is obtained. , and The value of determines the nonlinear relationship in the nonlinear relationship corresponding to the weather temperature in each time period.

4. The fault monitoring method based on the power collection terminal according to claim 1 is characterized in that: The method divides all time periods into multiple time period sets according to the relationship similarity, and establishes a temperature search set so that each element in the temperature search set corresponds to a time period set, including: S1. Enter the nonlinear relationship equations corresponding to the weather temperature in all time periods into the temperature relationship set; S2, randomly selecting any nonlinear relational expression in the temperature relation set as the initial relational expression; Set a similarity threshold range to compare the similarity between the initial relationship and the remaining nonlinear relationship in the temperature relationship set; S3, taking the remaining nonlinear relational expressions in the temperature relation set as comparison relational expressions, and calculating the relationship similarity between the initial relational expression and each comparison relational expression respectively through the relationship similarity formula; S4, recording the relationship similarity between the initial relational expression and each comparison relational expression into a similarity set; Mark all relationship similarities in the similarity set that are within the similarity threshold range as qualified similarities; S5, entering the time periods corresponding to the nonlinear relational expressions of all qualified similarities and the time periods corresponding to the initial relational expressions into a time period set, and entering the initial relational expression into a temperature search set; S6, eliminating all nonlinear relational expressions and initial relational expressions marked as corresponding to qualified similarities in the temperature relationship set; S7, reselect any nonlinear relational expression in the temperature relation set as a new initial relational expression, and repeat S2 to S6 operations until all elements in the temperature relation set are traversed; obtain the maximum weather temperature and the minimum weather temperature in the time period corresponding to the initial relational expression, which are recorded as and ; Randomly obtain the maximum and minimum weather temperatures in the time period corresponding to any nonlinear relationship in the temperature relationship set, which are recorded as and ; pass Calculate the temperature difference and enter the temperature difference set; Reselect any nonlinear relationship in the temperature relationship set and calculate the maximum and minimum weather temperatures in the corresponding time period of the nonlinear relationship. and The temperature difference is calculated until all nonlinear relationships in the temperature relationship set are traversed; Sort all elements in the temperature difference set according to the temperature difference, from large to small; The weight value of each nonlinear relationship in the temperature relationship set is calculated by the following formula: In the above formula, G is the weight of the nonlinear relationship corresponding to the i-th temperature difference in the temperature difference set. is the i-th temperature difference in the temperature difference set, and h is the number of elements in the temperature difference set.

5. The fault monitoring method based on the power collection terminal according to claim 4 is characterized in that: The method of using the remaining nonlinear relational expressions in the temperature relation set as comparison relational expressions and calculating the relationship similarity between the initial relational expression and each comparison relational expression by using a relationship similarity formula includes: The relationship similarity formula is as follows: In the above formula, Cs is the relationship similarity, n is the time point in the time period, is the initial relation, is the contrast relation, G is the weight corresponding to the contrast relation, is the i-th time point.

6. The fault monitoring method based on the power collection terminal according to claim 1 is characterized in that: The step of obtaining a nonlinear relationship between the weather temperature and the time period in which the user is located, matching elements in a search set by using the nonlinear relationship between the weather temperature, and obtaining a power relationship set according to the matching elements includes: Get the user's current time period, which is recorded as the target time period; Obtain the weather temperature corresponding to each time point in the target time period through a temperature collection device; According to each time point in the target time period and the weather temperature corresponding to each time point, a nonlinear relationship formula of the weather temperature corresponding to the target time period is obtained, which is recorded as a target temperature relationship formula; The relationship similarity between the target temperature relationship and each nonlinear relationship in the temperature search set is calculated by using the relationship similarity formula, and the corresponding nonlinear relationship with the largest relationship similarity is selected as the search relationship; A time period set corresponding to the search relational expression is obtained, and a nonlinear relational expression of each time period in the time period set to the applied power is entered into the power relational set.

7. The fault monitoring method based on the power collection terminal according to claim 1 is characterized in that: The method obtains the corresponding power relationship set by searching the matched elements in the set, calculates the priority according to the start time and end time of the time period corresponding to each element in the power relationship set, and the weather temperature corresponding to the start time and the end time, and sorts all elements in the power relationship set according to the priority for subsequent relationship similarity calculation, including: The priority of the elements in the power relationship set is calculated by the following formula for sorting, as follows: In the above formula, is the target time period, and are the start and end time of the target time period, R is the time constant, To retrieve the elements in the time period set corresponding to the relation, and are the start time and end time of the jth time period in the time period set corresponding to the retrieval relation, To retrieve the temperature change between the start and end time of the jth time period in the time period set corresponding to the relationship, is the temperature change between the start time and the end time in the target time period, and for The maximum and minimum values ​​of , P is the element priority; The nonlinear relationship expressions in the power relationship set are sorted from large to small according to the element priorities of their corresponding time periods.

8. The fault monitoring method based on the power collection terminal according to claim 1 is characterized in that: The determining whether there is power fluctuation and failure according to the power consumption and the power consumption of the power relationship set and the nonlinear power relationship of the user in the current time period, includes: Select the minimum power consumption and the maximum power consumption among all nonlinear relationship expressions in the power relationship set to form a reasonable power consumption range; Obtain the user's electricity consumption at each time point in the target time period in real time; If the power consumption corresponding to each time point in the target time period is not within the reasonable power range, the user is marked as having a power failure in the target time period. Otherwise, the nonlinear relationship between the target time period and the applied power is obtained based on each time point in the target time period and the power consumption corresponding to each time point, which is recorded as the target power relationship. Use the relationship similarity formula to calculate the relationship similarity between the target power relationship formula and each nonlinear relationship formula in the power relationship set, and enter the similarity set; If there is a relationship similarity within the similarity threshold range in the similarity set, the power consumption in the adjacent time period behind the target time period is detected in real time; Otherwise, the user is marked as having power fluctuations during the target time period.

9. A system using the fault monitoring method based on the power collection terminal according to claim 1, characterized in that: include: The power collection terminal includes a power collection device and a temperature collection device; An electricity collection device, used to collect the electricity consumption of users; A temperature acquisition device is used to acquire daily temperature changes; a relationship building module is used to build nonlinear relationships between power consumption and temperature according to the user's power consumption and the corresponding temperature changes; The data processing module calculates the relationship similarity through the relationship similarity formula, and groups the nonlinear relationship of the user's power consumption according to the nonlinear relationship corresponding to the temperature change according to the relationship similarity; The data retrieval module retrieves the corresponding power quantity relationship set by calculating the temperature nonlinear relationship of the user's current time period and the temperature nonlinear relationship corresponding to each time period; The similarity judgment module uses the relationship similarity formula to traverse and calculate the relationship similarity between the nonlinear relationship of the power of the user's current time period and each element in the corresponding power relationship set.

Images