A method of indoor power distribution based on self-learning strategy

Abstract

The present application provides an indoor power distribution method based on a self-learning strategy, which involves the field of intelligent power distribution technology, including: segmented statistics of user power consumption data; preliminary smoothing processing; processing the user's future power consumption by exponential smoothing method, and using the processing results as input variables of the linear regression model; using mean square error as the loss function for linear regression model training, and updating the model parameters by a parameter optimization algorithm based on gradient descent; completing accurate measurement of user power consumption; performing feature vector matching, and issuing an alarm for the matched fault type. The present application improves the intelligence of indoor weak current well power distribution, can avoid overload risks in advance, improve power utilization efficiency, reduce operating costs and extend equipment life.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent power distribution, and in particular to an indoor power distribution method based on a self-learning strategy. Background Art

[0002] During the use of indoor weak-current wells, maintenance personnel are required to collect environmental quantities on site and conduct on-site control and management of components, which results in low operation and maintenance efficiency and prominent safety hazards, affecting the reliability of the indoor distributed system and user experience. In addition, the intelligent power distribution of indoor weak-current wells is not performed according to the specific situation of user electricity consumption, which may easily lead to power distribution not meeting user needs, which may cause power waste for some users, while other users have overloaded power demand and insufficient power supply.

[0003] At present, the traditional method mainly adopts manual or human-machine combined methods to predict user electricity consumption, and its prediction efficiency and accuracy are relatively poor. In addition, due to the complex situation in the indoor weak current well, the existing fault feedback method cannot be fully adapted. Summary of the invention

[0004] The present application is made in view of the above-mentioned problems, and its purpose is to provide an indoor power distribution method based on a self-learning strategy, which can predict the user's future power consumption according to the historically collected power data and timestamp information, and automatically match the fault type for the abnormal situation of the relevant data, thereby solving the current problem that the indoor weak-current well cannot realize the intelligent remote information collection and component management, overcoming the technical defect of the low intelligence of the power distribution equipment in the related technology, improving the intelligence of the power distribution of the indoor weak-current well, avoiding the risk of overload in advance, improving the power utilization efficiency, reducing the operating cost and extending the service life of the equipment.

[0005] Specifically, the first aspect of the present application provides a method for indoor power distribution based on a self-learning strategy, comprising the following steps:

[0006] Step 1: Collect the user's real-time electricity consumption data, obtain the user's historical electricity consumption data, and calculate the user's actual electricity consumption data in segments;

[0007] The segmentation of users' actual electricity consumption data needs to be determined according to the demand and the periodic fluctuation of users' electricity consumption. For example, if it is necessary to analyze users' monthly electricity demand in order to arrange equipment maintenance and overhaul work during the time period when electricity consumption is expected to be lower, and users' electricity consumption shows weekly periodic fluctuations, then users' actual electricity consumption data should be segmented on a weekly basis.

[0008] Step 2: Perform preliminary smoothing on the obtained power consumption data, and determine the stability of the user's power consumption pattern and the key time period for power distribution;

[0009] In the preliminary smoothing of electricity consumption data, the moving average method can eliminate the noise caused by short-term fluctuations or outliers, making the trend of the data more obvious, which is helpful for subsequent data analysis and prediction.

[0010] Step 3: Based on the processed user electricity consumption data, the user's future electricity consumption is predicted by exponential smoothing method, and the prediction results are used as input variables of the linear regression model to establish a linear relationship between the predicted electricity consumption and time. The formula is as follows:

[0011] ;

[0012] ;

[0013] in: is the smoothed power consumption in the tth time period;

[0014] is the smoothing constant;

[0015] is the actual power consumption in the tth time period;

[0016] is the smoothed power consumption in the t-1th time period;

[0017] is the predicted electricity consumption in the tth time period;

[0018] is the intercept;

[0019] is the slope;

[0020] t is the time period;

[0021] is the error term of the tth time period;

[0022] Exponential smoothing is a time series analysis method that predicts future values ​​by weighting historical data. Its characteristic is that it gives a larger weight to recent data, while the weight of long-term data gradually decreases, so that it can better reflect the changing trend of data. This method is suitable for data with trends and seasonality. In the user electricity consumption forecast, the exponential smoothing method is used for medium- and short-term forecasts because it can effectively smooth random fluctuations and improve the forecast accuracy.

[0023] Linear regression models are used to determine the linear relationship between two variables, and are usually used to explore the relationship between independent variables (such as time) and dependent variables (such as electricity consumption). By establishing this relationship, future trends in electricity consumption can be more accurately predicted.

[0024] Combining the advantages of exponential smoothing and linear regression models can improve the accuracy of power load forecasting. Exponential smoothing captures trends and seasonal changes by weighting historical data, while linear regression provides a way to quantify these changes, making the forecast results more reliable and accurate.

[0025] Step 4: Use mean square error as the loss function for linear regression model training, and update the model parameters through a parameter optimization algorithm based on gradient descent until the model converges to the optimal solution. The formula is as follows:

[0026] ;

[0027] ;

[0028] Where: MSE is mean square error;

[0029] is the actual power consumption in the tth time period;

[0030] is the predicted electricity consumption in the tth time period;

[0031] N is the total number of time periods;

[0032] are model parameters;

[0033] is the learning rate;

[0034] is the gradient of the mean square error with respect to the model parameters;

[0035] In the linear regression model using mean square error (MSE) as the loss function, the model parameters are updated through a parameter optimization algorithm based on gradient descent. Its main function is to find a set of optimal model parameters so that the error between the model's predicted value and the actual value is minimized, thereby improving the accuracy of electricity consumption prediction.

[0036] When the change in the loss function is less than a preset threshold or reaches a predetermined number of iterations, the optimization process is considered to have converged. At this point, the model parameters are considered to be optimal parameters and the model's prediction performance is optimal.

[0037] Step 5: Analyze the user's power supply and distribution habits based on the user's historical power consumption data and power consumption forecast data, complete accurate power consumption measurement and achieve accurate power consumption time control;

[0038] Step 6: Combine the collected power data and environmental data, perform feature vector matching on the abnormal data, and issue an alarm for the matched fault type.

[0039] Furthermore, the segmented statistics of the actual power consumption data of the user specifically includes segmented statistics of the actual power consumption of the user in each time period as needed based on the historically collected power consumption data and timestamp information.

[0040] Furthermore, the preliminary smoothing process specifically includes: smoothing the obtained data by a moving average algorithm, and preliminarily observing the general trend of power consumption by the moving average value, and the formula is as follows:

[0041] ;

[0042] in: is the moving average of electricity consumption in the tth time period;

[0043] is the electricity consumption in the i-th time period;

[0044] is the number of periods of the moving average.

[0045] Furthermore, the determination of the stability of the user's power consumption pattern and the key period of power distribution specifically includes: determining the stability of the user's power consumption pattern by calculating the power consumption variance, and determining the key period of power distribution by calculating the peak-to-valley difference, and the formula is as follows:

[0046] ;

[0047] ;

[0048] in: is the electricity consumption variance;

[0049] is the power consumption in the tth time period;

[0050] is the average power consumption for each time period;

[0051] N is the total number of time periods;

[0052] PV is the difference between peak and valley power consumption;

[0053] is the maximum power consumption;

[0054] Minimum power consumption.

[0055] Furthermore, the step five specifically includes:

[0056] Based on the users' historical electricity consumption data and electricity consumption forecast data, we analyze the users' power supply and distribution habits, and accurately regulate the users' daily, monthly, quarterly and annual electricity consumption according to their needs.

[0057] Furthermore, the precise control specifically includes: accurately identifying peak power consumption periods and low power consumption periods. During low power consumption periods, if the user's equipment is in a non-critical operating state and power off is allowed to save energy, the system can perform automatic power off operations to reduce standby power consumption and unnecessary electricity consumption; and before the peak power consumption period arrives, start the equipment preheating or preparation program in advance to ensure that the equipment can operate efficiently during the peak period, and at the same time enter the high-precision sampling and metering mode during the peak period to accurately record the user's electricity consumption.

[0058] Furthermore, the power data and environmental data specifically include: the power data includes current data, voltage data, and power frequency data; the environmental data includes temperature data, humidity data, smoke data, and water flooding data.

[0059] Combined with the collected power data and environmental data, feature vector matching is performed on the abnormal data, and an alarm is issued for the matched fault type.

[0060] Furthermore, the step six specifically includes:

[0061] Step 8.1: Set the threshold value according to the data range of normal operation of the equipment;

[0062] Step 8.2: Based on experience and thresholds, determine the abnormality of the equipment according to power data and environmental data;

[0063] Step 8.3: Combine the abnormal data at the same location into a feature vector and match it with the feature vector of the known fault type in the fault model library;

[0064] Step 8.4: Output the matched fault type. For abnormal data that cannot be matched, it is manually determined whether it is a new fault type.

[0065] Furthermore, the step 8.2 specifically includes: if the data exceeds the set threshold, it is preliminarily judged as an abnormal situation; if the rate of change of the data exceeds the normal situation, that is, the ratio of the calculated data change per unit time to the original data exceeds the normal ratio, it is judged as an abnormal situation.

[0066] Furthermore, the step 8.3 specifically includes: for an abnormal situation, combining the detected power data and environmental data into a feature vector, such as [temperature data change, humidity data change, voltage data change, current data change, smoke signal, water immersion signal], and matching it with the feature vectors of known fault types in the fault model library.

[0067] In the second aspect, the present application also provides a computing device, which has the function of implementing the method described in the first aspect above. The beneficial effects can be found in the description of the first aspect, which will not be repeated here. The function can be implemented by hardware, or by executing the corresponding software implementation through hardware. The hardware or software includes one or more modules corresponding to the above functions. In one possible design, the structure of the device includes an acquisition module, a training module, and optionally, a construction module. These modules can implement the function of the training node in the method example of the first aspect above. Please refer to the detailed description in the method example for details, which will not be repeated here.

[0068] In the third aspect, the present application also provides a computing device, which is used to implement the functions of the method described in the first aspect above. The beneficial effects can be found in the description of the first aspect and will not be repeated here. The structure of the computing device includes a processor and a memory, and the memory is used to store instructions and / or data. The memory is coupled to the processor, and when the processor executes the program instructions stored in the memory, the function of the training node in the example of the first aspect above can be implemented. The structure of the computing device also includes a communication interface for communicating with other devices.

[0069] In a fourth aspect, the present application also provides a computer-readable storage medium, in which instructions are stored, and when the computer-readable storage medium is run on a computer, the computer executes the method in the above-mentioned first aspect and various possible designs of the first aspect.

[0070] In a fifth aspect, the present application also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the method in the above-mentioned first aspect and various possible designs of the first aspect.

[0071] In a sixth aspect, the present application also provides a computing chip, which is connected to a memory, and the chip is used to read and execute a software program stored in the memory, and to execute the methods in the above-mentioned first aspect and various possible implementation methods of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the embodiments of the present drawings or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present drawings. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0073] Figure 1 is a flow chart of the steps of the present invention;

[0074] The purpose, features and advantages of this figure will be further described in conjunction with the embodiments with reference to the accompanying drawings. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0076] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0077] Unless otherwise specified, all embodiments and optional embodiments of the present application can be combined with each other to form a new technical solution.

[0078] Unless otherwise specified, all technical features and optional technical features of this application can be combined with each other to form a new technical solution.

[0079] If there is no special explanation, all steps of the present application can be performed sequentially or randomly, preferably sequentially. For example, the method includes steps (a) and (b), which means that the method may include steps (a) and (b) performed sequentially, or may include steps (b) and (a) performed sequentially. For example, the method may further include step (c), which means that step (c) may be added to the method in any order. For example, the method may include steps (a), (b) and (c), or may include steps (a), (c) and (b), or may include steps (c), (a) and (b), etc.

[0080] If there is no special explanation, the "include" and "comprising" mentioned in this application are open-ended or closed-ended. For example, the "include" and "comprising" may mean that other components not listed may also be included or only the listed components may be included or only the listed components may be included.

[0081] If not specifically stated, in this application, the term "or" is inclusive. For example, the phrase "A or B" means "A, B, or both A and B". More specifically, any of the following conditions satisfies the condition "A or B": A is true (or exists) and B is false (or does not exist); A is false (or does not exist) and B is true (or exists); or both A and B are true (or exist).

[0082] In order to better understand the solutions of the embodiments of the present application, some relevant terms and concepts that may be involved in the embodiments of the present application are first introduced below.

[0083] (1) Moving average method is a smoothing forecasting technology. Its basic idea is to calculate the average value of a certain number of items in sequence according to the time series data, item by item, to reflect the long-term trend. Therefore, when the value of the time series fluctuates greatly due to the influence of periodic changes and random fluctuations, and it is difficult to show the development trend of the event, the moving average method can eliminate the influence of these factors, show the development direction and trend of the event (i.e., trend line), and then analyze and predict the long-term trend of the series based on the trend line.

[0084] (2) Exponential smoothing is actually a special weighted moving average method. Its characteristics are: First, the exponential smoothing method further strengthens the role of recent observations in the observation period on the predicted value. The weights assigned to observations at different times are different, thereby increasing the weights of recent observations, so that the predicted value can quickly reflect the actual changes in the market. The weights are reduced in a geometric series, the first term of which is the smoothing constant a, and the common ratio is (1-a). Second, the exponential smoothing method is flexible in the weights assigned to the observations, and different a values ​​can be used to change the rate of change of the weights. If a takes a small value, the weight changes more quickly, and the recent change trend of the observations can be more quickly reflected in the exponential moving average. Therefore, using the exponential smoothing method, different a values ​​can be selected to adjust the uniformity of the time series observations (that is, the smoothness of the trend change).

[0085] (3) Linear Regression is a type of regression analysis that models the relationship between one or more independent variables and a dependent variable using a least squares function called the linear regression equation. This function is a linear combination of one or more model parameters called regression coefficients. The case with only one independent variable is called simple regression, and the case with more than one independent variable is called multiple regression. (This in turn should be distinguished from multiple linear regression, where multiple related dependent variables are predicted, rather than a single scalar variable.)

[0086] In linear regression, the data is modeled using a linear prediction function, and the unknown model parameters are estimated from the data. These models are called linear models. The most common linear regression model is that the conditional mean of y given the value of X is an affine function of X.

[0087] (4) Linear regression model is a statistical analysis method used to predict the relationship between a dependent variable (or outcome variable) and one or more independent variables (or predictor variables). Linear regression models can be divided into simple linear regression and multiple linear regression. Simple linear regression involves one independent variable and one dependent variable, while multiple linear regression involves multiple independent variables. The goal is to estimate these regression coefficients by minimizing the sum of squared errors between the predicted values ​​and the actual observed values.

[0088] (5) Mean Squared Error (MSE) is an indicator used to measure the degree of difference between an estimate and the true value. Specifically, it is the expected value of the square of the difference between the parameter estimate and the true value, which can effectively reflect the gap between the model prediction value and the true value. The smaller its value, the more accurate the model's description of the data. The mean squared error not only takes into account the bias (the difference between the estimate and the true value), but also the variance (the volatility of the estimate itself), so it is a comprehensive error measurement indicator.

[0089] (6) Gradient descent algorithm is an optimization algorithm widely used in machine learning and deep learning. Its core goal is to minimize the loss function or objective function by iteratively updating the model parameters. The basic idea of ​​gradient descent is to gradually adjust the parameters along the negative gradient direction of the loss function until a local minimum or convergence is found. Choosing a suitable learning rate is crucial to the performance of the gradient descent algorithm. A learning rate that is too high may cause the model to deviate from the optimal solution, while a learning rate that is too low may result in slow convergence.

[0090] In this embodiment, Figure 1 As shown, a method for indoor power distribution based on a self-learning strategy comprises the following steps:

[0091] Step 1: Collect the user's real-time electricity consumption data, obtain the user's historical electricity consumption data, and calculate the user's actual electricity consumption data in segments;

[0092] The segmentation of users' actual electricity consumption data needs to be determined according to the demand and the periodic fluctuation of users' electricity consumption. For example, if it is necessary to analyze users' monthly electricity demand in order to arrange equipment maintenance and overhaul work during the time period when electricity consumption is expected to be lower, and users' electricity consumption shows weekly periodic fluctuations, then users' actual electricity consumption data should be segmented on a weekly basis.

[0093] Step 2: Perform preliminary smoothing on the obtained power consumption data, and determine the stability of the user's power consumption pattern and the key time period for power distribution;

[0094] In the preliminary smoothing of electricity consumption data, the moving average method can eliminate the noise caused by short-term fluctuations or outliers, making the trend of the data more obvious, which is helpful for subsequent data analysis and prediction.

[0095] Step 3: Based on the processed user electricity consumption data, the user's future electricity consumption is predicted by exponential smoothing method, and the prediction results are used as input variables of the linear regression model to establish a linear relationship between the predicted electricity consumption and time. The formula is as follows:

[0096] ;

[0097] ;

[0098] In this embodiment, the smoothing parameter is 0.85.

[0099] Exponential smoothing is a time series analysis method that predicts future values ​​by weighting historical data. Its characteristic is that it gives a larger weight to recent data, while the weight of long-term data gradually decreases, so that it can better reflect the changing trend of data. This method is suitable for data with trends and seasonality. In the user electricity consumption forecast, the exponential smoothing method is used for medium- and short-term forecasts because it can effectively smooth random fluctuations and improve the forecast accuracy.

[0100] Linear regression models are used to determine the linear relationship between two variables, and are usually used to explore the relationship between independent variables (such as time) and dependent variables (such as electricity consumption). By establishing this relationship, future trends in electricity consumption can be more accurately predicted.

[0101] Combining the advantages of exponential smoothing and linear regression models can improve the accuracy of power load forecasting. Exponential smoothing captures trends and seasonal changes by weighting historical data, while linear regression provides a way to quantify these changes, making the forecast results more reliable and accurate.

[0102] Step 4: Use mean square error as the loss function for linear regression model training, and update the model parameters through a parameter optimization algorithm based on gradient descent until the model converges to the optimal solution. The formula is as follows:

[0103] ;

[0104] ;

[0105] In this embodiment, the learning rate is initially set to 0.1, and as the number of iterations increases, the learning rate is gradually reduced, which can speed up the convergence speed in the early iterations and make subtle adjustments in the later stages.

[0106] In the linear regression model using mean square error (MSE) as the loss function, the model parameters are updated through a parameter optimization algorithm based on gradient descent. Its main function is to find a set of optimal model parameters so that the error between the model's predicted value and the actual value is minimized, thereby improving the accuracy of electricity consumption prediction.

[0107] When the change in the loss function is less than a preset threshold or reaches a predetermined number of iterations, the optimization process is considered to have converged. At this point, the model parameters are considered to be optimal parameters and the model's prediction performance is optimal.

[0108] Step 5: Analyze the user's power supply and distribution habits based on the user's historical power consumption data and power consumption forecast data, complete accurate power consumption measurement and achieve accurate power consumption time control;

[0109] Step 6: Combine the collected power data and environmental data, perform feature vector matching on the abnormal data, and issue an alarm for the matched fault type.

[0110] Furthermore, the actual power consumption data of the user is segmented and counted, specifically, the actual power consumption data of the user in each time period is segmented and counted as needed according to the historically collected power consumption data and timestamp information.

[0111] Furthermore, preliminary smoothing processing specifically includes: smoothing the obtained data through a moving average algorithm, and preliminarily observing the general trend of power consumption through the moving average value. The formula is as follows:

[0112] ;

[0113] Furthermore, the stability of the user's power consumption pattern and the key period of power distribution are determined, specifically including: determining the stability of the user's power consumption pattern by calculating the power consumption variance, and determining the key period of power distribution by calculating the peak-to-valley difference, the formula is as follows:

[0114] ;

[0115] ;

[0116] Furthermore, step five specifically includes:

[0117] Based on the users' historical electricity consumption data and electricity consumption forecast data, we analyze the users' power supply and distribution habits, and accurately regulate the users' daily, monthly, quarterly and annual electricity consumption according to their needs.

[0118] 1) Daily electricity usage habit analysis and strategy

[0119] Within a single day, the system continuously monitors the user's electricity consumption data, such as changes in electricity consumption every hour or shorter time intervals. By collecting and analyzing electricity consumption data for a whole day, it identifies peak electricity consumption periods (such as a surge in electricity consumption of production equipment in enterprises during daytime office hours, and concentrated use of household appliances during residents' dinner time, etc.) and low electricity consumption periods (such as most equipment in standby or off mode in the early morning).

[0120] During the off-peak period, if the user's equipment is in a non-critical operating state and power-off is allowed, the system can perform automatic power-off operations to reduce standby power consumption and unnecessary power consumption. Before the peak period of electricity consumption, the equipment preheating or preparation program is started in advance to ensure that the equipment can operate efficiently during the peak period. At the same time, during the peak period, the high-precision sampling and metering mode is entered to accurately record the user's electricity consumption for subsequent accurate billing and electricity consumption analysis.

[0121] 2) Monthly electricity usage habit analysis and strategy

[0122] Collect users' daily electricity consumption data every month and analyze the changing trend of electricity consumption throughout the month. It may be found that in some months, the overall electricity consumption increases due to heavier production tasks and seasonal factors (such as frequent use of air conditioners in summer), and it can also determine which days or weeks of the month have more concentrated electricity consumption.

[0123] According to the monthly electricity consumption trend, equipment maintenance and overhaul work is arranged during the period when electricity consumption is expected to be low, because the impact of equipment downtime on overall production or use is relatively small at this time. For months with high electricity consumption, communicate with users in advance and provide energy-saving suggestions, such as adjusting production plans, optimizing equipment operating parameters, etc., to reduce overall electricity consumption. At the same time, according to the distribution of peak electricity consumption days within the month, reasonably adjust power allocation resources to ensure sufficient power supply on peak days, and strengthen high-precision metering and monitoring on these peak days.

[0124] 3) Quarterly electricity usage habit analysis and strategy

[0125] Analyze user electricity consumption data on a quarterly basis, taking into account the impact of factors such as climate characteristics in different seasons and industry production rules on electricity consumption. For example, the manufacturing industry may consume a lot of electricity in the first and fourth quarters due to busy production during the peak order season, and commercial places may consume more electricity in summer and winter due to air conditioning cooling or heating needs.

[0126] Develop long-term power resource planning based on the electricity consumption characteristics of different quarters. In the off-season (such as the second quarter of some industries), you can negotiate with the power supply department to adjust the power contract capacity and reduce basic electricity expenses. For the peak season, reserve backup power equipment in advance or coordinate with distributed energy systems to ensure the stability of power supply. And within each quarter, further refine the time-of-use metering and energy-saving management and control strategies based on the monthly and daily electricity consumption habits within the quarter. For example, in the months and days with peak electricity consumption in the quarter, strengthen the precise power collection of user equipment during peak electricity consumption and the implementation of power-off energy-saving strategies during low electricity consumption.

[0127] 4) Annual electricity usage habit analysis and strategy

[0128] Comprehensively analyze the electricity consumption data throughout the year to fully understand the users' annual electricity consumption patterns, including fluctuations in the total annual electricity consumption, the proportion of electricity consumption in different seasons and months, the distribution and changing trends of peak and trough electricity consumption within the year, etc.

[0129] Based on the annual electricity consumption habits, we formulate annual electricity budgets and energy-saving targets for users, and provide corresponding energy-saving plans and technical support. In terms of electricity supply contracts, we negotiate with power supply companies for more favorable electricity price packages or electricity procurement plans based on the annual total electricity consumption forecast. At the same time, we conduct annual energy efficiency evaluations on user equipment to determine whether equipment needs to be upgraded or replaced to improve energy efficiency. During the peak electricity consumption period each year (such as specific seasons or production peak seasons), we ensure the effective implementation of high-precision metering and energy-saving control strategies, and continuously optimize this year's strategies based on the lessons learned from the previous year to achieve year-on-year energy savings and power management optimization.

[0130] Furthermore, precise control specifically includes: accurately identifying peak and low power consumption periods. During low power consumption periods, if the user's equipment is in a non-critical operating state and power off is allowed, the system can perform automatic power off operations to reduce standby power consumption and unnecessary electricity consumption; before the peak power consumption period arrives, start the equipment preheating or preparation program in advance to ensure that the equipment can operate efficiently during the peak period. At the same time, enter the high-precision sampling and metering mode during the peak period to accurately record the user's electricity consumption.

[0131] Furthermore, the power data and environmental data specifically include: the power data includes current data, voltage data, and power frequency data; the environmental data includes temperature data, humidity data, smoke data, and water flooding data.

[0132] Combined with the collected power data and environmental data, feature vector matching is performed on the abnormal data, and an alarm is issued for the matched fault type.

[0133] Furthermore, step six specifically includes:

[0134] Step 8.1: Set the threshold value according to the data range of normal operation of the equipment;

[0135] For environmental quantities and power data, you can first set thresholds based on the experience range of normal equipment operation. For example, for temperature, set a normal operating temperature range, such as -10°C - 40°C, based on the equipment specifications and actual operating experience. When the collected temperature data exceeds this range, it is judged to be abnormal. Similarly, for humidity, 30% - 60% can be set as the normal range; for voltage, in a three-phase power system, the normal range of phase voltage is ±10% of the rated voltage (such as a rated 380V phase voltage, the normal range is 342V - 418V).

[0136] For frequency, in a power system with a power frequency of 50Hz, an acceptable fluctuation range is set, such as 49.5Hz-50.5Hz as the normal range. When the data exceeds these pre-set thresholds, it is initially determined to be abnormal data.

[0137] Step 8.2: Based on experience and thresholds, determine the abnormality of the equipment according to power data and environmental data;

[0138] Step 8.3: Combine the abnormal data at the same location into a feature vector and match it with the feature vector of the known fault type in the fault model library;

[0139] Step 8.4: Output the matched fault type. For abnormal data that cannot be matched, it is manually determined whether it is a new fault type.

[0140] Furthermore, step 8.2 specifically includes: if the data exceeds the set threshold, it is preliminarily judged as an abnormal situation; if the rate of change of the data exceeds the normal situation, that is, the ratio of the calculated data change per unit time to the original data exceeds the normal ratio, it is judged as an abnormal situation.

[0141] Calculate the ratio of the change in data per unit time to the original data. For example, for current data, if the current change rate exceeds a certain value, such as 50%, in a very short time (such as a few seconds), this may mean that there is an abnormal situation such as a short circuit in the circuit or a high-power device is suddenly connected.

[0142] A similar approach can be used for environmental data such as temperature and humidity. If the temperature rises or falls faster than the normal heat exchange or environmental adjustment speed in a short period of time (such as a few minutes), it is considered abnormal. For example, under normal circumstances, the temperature in the computer room does not change by more than 5°C per hour. If the temperature rises by 10°C within 10 minutes, it is considered abnormal.

[0143] Furthermore, step 8.3 specifically includes: for an abnormal situation, combining the detected power data and environmental data into a feature vector, such as [temperature data change, humidity data change, voltage data change, current data change, smoke signal, water immersion signal], and matching them with the feature vectors of known fault types in the fault model library.

[0144] In this embodiment, in the fault model library, the characteristic vector of the known electrical short circuit fault is [rapid temperature rise, normal humidity, sudden voltage drop, sharp current increase, no smoke, no water immersion]. When the collected characteristic vector is similar to this pattern, it is determined that an electrical short circuit fault may have occurred.

[0145] It should be noted that the present application is not limited to the above-mentioned embodiments. The above-mentioned embodiments are only examples, and the embodiments having the same structure as the technical idea and exerting the same effect within the scope of the technical solution of the present application are all included in the technical scope of the present application. In addition, without departing from the scope of the main purpose of the present application, various modifications that can be thought of by those skilled in the art to the embodiments and other methods of combining some of the constituent elements in the embodiments are also included in the scope of the present application.

Claims

1. A method for indoor power distribution based on a self-learning strategy, characterized in that: The following steps are involved: Step 1: Collect the user's real-time electricity consumption data, obtain the user's historical electricity consumption data, and calculate the user's actual electricity consumption data in segments; Step 2: Perform preliminary smoothing on the obtained power consumption data, and determine the stability of the user's power consumption pattern and the key time period for power distribution; Step 3: Based on the processed user power consumption data, the user's future power consumption is smoothed by exponential smoothing method, and the smoothed power consumption is used as the input variable of the linear regression model to establish a linear relationship between the predicted power consumption and time. The formula is as follows: ； ； in: is the smoothed power consumption in the tth time period; is the smoothing constant; is the actual power consumption in the tth time period; is the smoothed power consumption in the t-1th time period; is the predicted electricity consumption in the tth time period; is the intercept; is the slope; t is the time period; is the error term of the tth time period; Step 4: Use mean square error as the loss function for linear regression model training, and update the model parameters through a parameter optimization algorithm based on gradient descent until the model converges to the optimal solution. The formula is as follows: ； ； Where: MSE is the mean square error; is the actual power consumption in the tth time period; is the predicted electricity consumption in the tth time period; N is the total number of time periods; are model parameters; is the learning rate; is the gradient of the mean square error with respect to the model parameters; Step 5: Analyze the user's power supply and distribution habits based on the user's historical power consumption data and power consumption forecast data, complete accurate power consumption measurement and achieve accurate power consumption time control; Step 6: Combine the collected power data and environmental data, perform feature vector matching on the abnormal data, and issue an alarm for the matched fault type.

2. According to claim 1, a method for indoor power distribution based on a self-learning strategy is characterized in that: The segmented statistics of the actual power consumption data of the user specifically includes segmented statistics of the actual power consumption of the user in each time period as required based on the historically collected power consumption data and timestamp information.

3. According to the method of indoor power distribution based on self-learning strategy in claim 1, it is characterized in that: The preliminary smoothing process specifically includes: smoothing the obtained data by a moving average algorithm, and preliminarily observing the general trend of power consumption by the moving average value, and the formula is as follows: ； in: is the moving average of electricity consumption in the tth time period; is the electricity consumption in the i-th time period; is the number of periods of the moving average.

4. According to claim 1, a method for indoor power distribution based on a self-learning strategy is characterized in that: The determination of the stability of the user's power consumption pattern and the key period of power distribution specifically includes: determining the stability of the user's power consumption pattern by calculating the power consumption variance, and determining the key period of power distribution by calculating the peak-to-valley difference, and the formula is as follows: ； ； in: is the electricity consumption variance; is the power consumption in the tth time period; is the average power consumption for each time period; N is the total number of time periods; PV is the difference between peak and valley power consumption; is the maximum power consumption; Minimum power consumption.

5. The indoor power distribution method based on self-learning strategy according to claim 1 is characterized in that: The step five specifically includes: Based on the users' historical electricity consumption data and electricity consumption forecast data, we analyze the users' power supply and distribution habits, and accurately regulate the users' daily, monthly, quarterly and annual electricity consumption according to their needs.

6. A method for indoor power distribution based on a self-learning strategy according to claim 5, characterized in that: The precise control specifically includes: accurately identifying peak power consumption periods and off-peak power consumption periods. During off-peak power consumption periods, if the user's equipment is in a non-critical operating state and power-off for energy saving is allowed, the system can perform an automatic power-off operation to reduce standby power consumption and unnecessary electricity consumption; before the peak power consumption period arrives, start the equipment preheating or preparation program in advance to ensure that the equipment can operate efficiently during the peak period, and at the same time enter a high-precision sampling and metering mode during the peak period to accurately record the user's electricity consumption.

7. The indoor power distribution method based on self-learning strategy according to claim 1 is characterized in that: The power data and environmental data specifically include: the power data includes current data, voltage data, and power frequency data; the environmental data includes temperature data, humidity data, smoke data, and water flooding data.

8. The indoor power distribution method based on self-learning strategy according to claim 1 is characterized in that: The step six specifically includes: Step 8.1: Set the threshold value according to the data range of normal operation of the equipment; Step 8.2: Based on experience and thresholds, determine the abnormality of the equipment according to power data and environmental data; Step 8.3: Combine the abnormal data at the same location into a feature vector and match it with the feature vector of the known fault type in the fault model library; Step 8.4: Output the matched fault type. For abnormal data that cannot be matched, it is manually determined whether it is a new fault type.

9. The indoor power distribution method based on the self-learning strategy according to claim 8 is characterized in that: The step 8.2 specifically includes: if the data exceeds the set threshold, it is preliminarily judged as an abnormal situation; if the rate of change of the data exceeds the normal situation, that is, the ratio of the calculated data change per unit time to the original data exceeds the normal ratio, it is judged as an abnormal situation.

10. The indoor power distribution method based on self-learning strategy according to claim 8, characterized in that: The step 8.3 specifically includes: for an abnormal situation, combining the detected power data and environmental data into a feature vector, and matching it with the feature vectors of known fault types in the fault model library.

Images