Electricity consumption prediction method and system based on intelligent electric energy meter

Abstract

The invention relates to the field of data processing, in particular to an electricity consumption prediction method and system based on an intelligent electric energy meter. The method comprises the following steps: acquiring an average temperature predicted value of a next date; obtaining the planned operation duration of the production equipment in the next date; using the intelligent electric energy meter to obtain an electricity consumption real value of continuous preset days; and inputting the average temperature prediction value of the next date and the planned operation duration into a preset electricity consumption prediction model, thereby obtaining an enterprise electricity consumption prediction value of the next date. By adopting the method, the accuracy of the enterprise electricity consumption prediction result can be improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing. More specifically, the present invention relates to a method and system for predicting electricity consumption based on an intelligent electricity meter. Background Art

[0002] A production enterprise refers to an enterprise that transforms input factors such as raw materials and components into products with higher added value through physical or chemical processing, assembly, and other production processes. When processing, assembling, and other production processes are carried out on raw materials and components, relevant production equipment is often required. With the second electricity reform, the electricity market has gradually developed towards a real-time market. In order to ensure the normal operation of production equipment within the enterprise, it is often necessary to predict the electricity consumption of the enterprise, so that the enterprise can accurately purchase an appropriate amount of electricity in the real-time market, prevent losses caused by deviation assessment due to excessive purchase of electricity, or prevent the production equipment from being unable to produce enough products due to insufficient purchase, thereby affecting the economic benefits of the enterprise.

[0003] The existing method for predicting the electricity consumption of enterprises predicts the electricity consumption of enterprises based on the order volume of the enterprises or based on the historical electricity consumption pattern of the enterprises. However, since the factors considered in the prediction are relatively single, the accuracy of the prediction result is poor. For example: The Chinese invention patent with the authorization publication number CN 111275267 B discloses a method for predicting the electricity consumption of a production enterprise. When predicting the electricity consumption of the enterprise, this method predicts the electricity consumption by fitting the functional relationship between the order volume of the enterprise and the electricity consumption of the enterprise and combining the predicted order volume. Since only the order volume is considered in the electricity consumption prediction, in fact, not only the order volume will affect the electricity consumption of the enterprise, but other factors will also affect the electricity consumption of the enterprise, resulting in a low accuracy of the electricity consumption prediction result of this method. Summary of the Invention

[0004] To solve the technical problem of the low accuracy of the prediction result of the existing method for predicting the electricity consumption of production enterprises, the present invention provides solutions in the following aspects.

[0005] In a first aspect, the present invention provides a method for predicting electricity consumption based on an intelligent electricity meter. Including: obtaining the predicted average temperature value for the next date; obtaining the planned operation duration of the enterprise's production equipment for the next date; obtaining the true value of the enterprise's electricity consumption for a continuous preset number of days using an intelligent electricity meter; inputting the predicted average temperature value for the next date and the planned operation duration into a preset electricity consumption prediction model to obtain the predicted value of the enterprise's electricity consumption for the next date; the training set of the electricity consumption prediction model consists of the true value of electricity consumption, average temperature, and operation duration for the continuous preset number of days. When training the electricity consumption prediction model, the loss function corresponding to a certain date is positively correlated with the deviation between the true value and the predicted value of electricity consumption on that date, and is positively correlated with the ratio of the planned operation duration and the actual operation duration of the production equipment on that date.

[0006] The beneficial effects are as follows: The electricity consumption prediction method based on an intelligent electricity meter in the present invention takes into account that higher or lower temperatures will increase the electricity consumption of the enterprise when predicting the enterprise's electricity consumption; if the operation duration of the production equipment is longer, the electricity consumption of the enterprise is also greater; uses a neural network model to learn the coupling relationship among the average temperature, production equipment operation duration, and enterprise electricity consumption during a day, and predicts the enterprise's electricity consumption based on data in two dimensions, thereby improving the accuracy of the prediction result; in addition, during the training process of the electricity consumption prediction model, considering the relationship between the ratio of the planned operation duration and the actual operation duration of the production equipment and equipment aging, and that equipment aging will lead to higher electricity consumption of the production equipment, by using this ratio to compensate the loss function, when the degree of equipment aging is relatively light, the penalty for prediction error is reduced, and when the degree of production equipment aging is relatively severe, the penalty for prediction error is increased, so that the electricity consumption prediction model can better adapt to the situation of higher actual electricity consumption caused by equipment aging, and effectively improve the prediction accuracy of the electricity consumption prediction model.

[0007] Preferably, the expression of the loss function used during training is: ; In the formula, represents the loss function corresponding to the nth date, and are the weighting coefficients for underestimation error and overestimation error respectively, , and are the actual operation duration and planned operation duration of the production equipment on the nth date respectively, and are the true value and predicted value of electricity consumption on the nth date respectively.

[0008] Its beneficial effects are as follows: making the weighting coefficient of the underestimation error in the loss function greater than that of the overestimation error, so that the penalty for under-prediction is greater and the penalty for over-prediction is smaller, thereby enabling the electricity consumption prediction model to better adapt to the situation of higher actual electricity consumption caused by equipment aging and avoiding insufficient power supply for enterprises.

[0009] Preferably, it further includes: correcting the loss function, and the calculation expression of the corrected loss function is: ; In the formula, is the probability that the production equipment fails on the nth date, represents the corrected loss function.

[0010] Its beneficial effects are as follows: by increasing the penalty for under-prediction when the equipment failure probability increases and decreasing the penalty for under-prediction when the equipment failure probability decreases, the prediction model becomes more sensitive to the underestimation error (i.e., the electricity consumption prediction result is lower than the actual electricity consumption) as the probability of equipment failure increases.

[0011] Preferably, the predicted value of the average temperature for the next date is obtained using weather forecasts.

[0012] Preferably, the expression for the weighting coefficient for the underestimation error is: ; In the formula, is the penalty strength factor for the underestimation error, is the probability that the production equipment fails on the nth date.

[0013] Its beneficial effects are as follows: since when the production equipment fails, it needs to be repaired by other devices and then put into operation, and the repair time is usually not too long. After it is repaired, it can be put into operation according to the planned operation duration, which ultimately leads to an increase in its electricity consumption within a date. By setting the weighting coefficient of the underestimation error to increase with the increase in the failure probability, the electricity consumption prediction model can better adapt to the situation of higher actual electricity consumption caused by equipment failure.

[0014] Preferably, the expression for the weighting coefficient for the overestimation error is: ; In the formula, is the penalty strength factor for the overestimation error, is the probability that the production equipment fails on the nth date.

[0015] The beneficial effects are as follows: Since the power consumption on the corresponding date will increase when the production equipment fails; by making the weighted coefficient of the overestimation error negatively correlated with the equipment failure probability, the penalty for over-prediction decreases as the equipment failure probability increases, which helps to avoid insufficient power supply for enterprises.

[0016] Preferably, the continuous preset number of days is 50 days of dates.

[0017] Preferably, it further includes: processing the power consumption data obtained by using the smart electricity meter, and the processing method includes: Performing data cleaning on the power consumption data; data cleaning includes: checking the format of the power consumption data and detecting whether there are missing values, duplicate data, and error data in the power consumption data, and processing the detected missing values, duplicate data, and error data; Performing outlier detection and correction on the cleaned power consumption data; Performing data smoothing and filtering processing on the power consumption data.

[0018] Preferably, for the missing power consumption data, linear interpolation is used for filling.

[0019] In the second aspect, the present invention provides a power consumption prediction device based on a smart electricity meter, including a memory and a processor. Computer program instructions are stored in the memory, and when the computer program instructions are executed by the processor, the power consumption prediction method based on the smart electricity meter of the present invention is implemented.

[0020] In summary, the beneficial effect of the present invention is that: adopting the method of the present invention can improve the accuracy of the enterprise power consumption prediction result. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It schematically shows a flowchart of a power consumption prediction method based on a smart electricity meter according to an embodiment of the present invention; Figure 2 It schematically shows a structural diagram of a power consumption prediction system based on a smart electricity meter according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0023] Next, the specific embodiments of the present invention will be described in detail in conjunction with the drawings.

[0024] Embodiment of the power consumption prediction method based on an intelligent electricity meter: As Figure 1 shown, the power consumption prediction method based on an intelligent electricity meter of the present invention includes: S101. Obtain the predicted value of the average temperature for the next date; In this embodiment, the value of m is 10. In other embodiments, other suitable values can also be taken.

[0025] S102. Obtain the planned operating duration of the enterprise's production equipment for the next date; use the intelligent electricity meter to obtain the true values of the enterprise's power consumption for a continuous preset number of days; In this embodiment, the continuous preset number of days is 50 days. In other embodiments, other suitable numbers of days can also be taken, such as 40 days, 60 days, etc. The planned operating duration is less than or equal to the working duration of one working day of the enterprise.

[0026] In this embodiment, when collecting the enterprise's power consumption, it is necessary to ensure that the used intelligent electricity meter supports long-term data storage and remote data access. The adopted intelligent electricity meter needs to be connected to the data acquisition platform through wired or wireless communication to upload data. After installing the intelligent electricity meter, the basic parameters of the intelligent electricity meter need to be set. For example, the data recording interval is set to once every 15 minutes, 20 minutes, or 1 hour. The duration of storing historical data is set to 60 days.

[0027] S103. Obtain the predicted value of the enterprise's power consumption for the next date, specifically: input the predicted value of the average temperature for the next date and the planned operating duration into a preset power consumption prediction model, so as to obtain the predicted value of the enterprise's power consumption for the next date; the training set of the power consumption prediction model is composed of the true values of the power consumption, the average temperature, and the operating duration of the production equipment for the continuous preset number of days. When training the power consumption prediction model, the loss function corresponding to a certain date is positively correlated with the deviation between the true value and the predicted value of the power consumption on that date, and is positively correlated with the ratio of the planned operating duration to the actual operating duration of the production equipment on that date.

[0028] The production enterprises in this embodiment refer to the production enterprises in the machining industry, and the operating mode of their production equipment is an intermittent production mode. During a day, the equipment will be in four states, namely, maintenance state, repair state, operating state, and shutdown state. The operating duration of the equipment in a day is related to the order volume. When the order volume is sufficient, the operating duration is longer. When the order volume is insufficient or in the production off-season, the operating duration of the equipment in a day is shorter. If the ratio of the planned operating duration to the actual operating duration of the production equipment increases, it indicates that the equipment has aged. The aging of the production equipment will lead to a relatively high power consumption of the production equipment. By using this ratio to compensate the loss function, when the planned operating duration of the production equipment is close to the actual operating duration (i.e., the equipment is less aged), the penalty for the prediction error is reduced, and when the deviation between the planned operating duration and the actual operating duration of the production equipment is large (i.e., the equipment is more severely aged), the penalty for the prediction error is increased, so that the power consumption prediction model can better adapt to the situation of relatively high actual power consumption caused by equipment aging, and effectively improve the prediction accuracy of the power consumption prediction model.

[0029] If the temperature is relatively high or low during a day, in order to ensure that the staff are in a comfortable working environment, the usage duration of the enterprise's air conditioning equipment will increase, and the air conditioner will operate at high power consumption, thus increasing the power consumption of the enterprise. For example, when the temperature is too high, the air conditioner will turn on the cooling mode, and the air conditioner temperature will be set relatively low and the wind speed will be set relatively high, resulting in an increase in the power of the air conditioner; when the temperature is too low, the air conditioner will turn on the heating mode, the air conditioner temperature will be set relatively high and the wind speed will be set relatively high, which will also lead to an increase in the power of the air conditioner and an increase in the power consumption per unit duration.

[0030] If the operating duration of the production equipment is relatively long, it will also increase the power consumption of the enterprise. The power consumption prediction method based on the intelligent electricity meter of the present invention comprehensively considers the influence of the average temperature during a day on the enterprise's power consumption and the influence of the operating duration of the production equipment on the enterprise's power consumption when predicting the enterprise's power consumption. Since the neural network model can learn the coupling relationship between the average temperature during a day, the operating duration of the production equipment and the enterprise's power consumption, two-dimensional data is used in combination with the neural network model to predict the enterprise's power consumption, so that the power consumption prediction result is more accurate.

[0031] In this embodiment, the production equipment may be a numerically controlled machine tool. In other embodiments, it may also be other types of production equipment.

[0032] In one embodiment, the expression of the loss function used when training the power consumption prediction model is: ; In the formula, represents the loss function corresponding to the nth date, and are the weighting coefficients for the underestimated error and the overestimated error respectively, , and are the actual operating duration and the planned operating duration of the production equipment on the nth date respectively, and are the true value and the predicted value of the power consumption on the nth date respectively.

[0033] Since the aging of production equipment will lead to a relatively high power consumption of the production equipment, it may cause insufficient power supply for the enterprise and affect the production efficiency of products. Therefore, the consequence of underestimating the power consumption is more serious. By making the weighting coefficient of the underestimation error in the loss function greater than that of the overestimation error, the penalty for under-prediction is increased and the penalty for over-prediction is reduced, so that the power consumption prediction model can better adapt to the situation of relatively high actual power consumption caused by equipment aging, further improve the accuracy of the power consumption prediction result, and avoid the power consumption prediction being lower than the actual power consumption, thus avoiding insufficient power supply for the enterprise.

[0034] In one embodiment, it further includes: correcting the loss function, and the calculation expression of the corrected loss function is: ; In the formula, is the probability that the production equipment fails on the nth date, represents the corrected loss function. is a weighting factor based on the equipment failure probability. By using to correct the loss function, a greater penalty for under-prediction is achieved when the equipment failure probability is relatively high, so that as the probability of equipment failure increases, the prediction model is more sensitive to the underestimation error (i.e., the power consumption prediction result is lower than the actual power consumption), and thus the power consumption prediction model can better adapt to the situation of relatively high actual power consumption caused by equipment failure.

[0035] Under normal circumstances, the power consumption of production equipment is determined by the operation duration. Since after the production equipment fails, other electrical devices need to be used to repair the production equipment, additional electric energy will be consumed. Therefore, the penalty for under-prediction increases as the probability of equipment failure increases, and the prediction model is more sensitive to the underestimation error, so that the power consumption prediction model can better adapt to the situation of relatively high actual power consumption caused by equipment failure. Training the power consumption prediction model with the corrected loss function can further improve the prediction accuracy of the power consumption prediction model.

[0036] In this embodiment, there are various methods for obtaining the probability that the production equipment fails on the nth date. For example: an LSTM neural network model can be used to obtain the failure probability of the production equipment on each date. It is also possible to use the equipment failure prediction method disclosed in the invention patent with the authorization announcement number CN112395178B to predict the probability that the production equipment will fail on the next date.

[0037] LSTM neural network, that is, long short-term memory network, is a type of time-recurrent neural network. The long short-term memory network can solve the long-term dependence problem existing in general RNN (recurrent neural network).

[0038] If the LSTM neural network model is used to obtain the failure probability of production equipment for each date, the training process of the LSTM neural network model is as follows: S201. Construct a training set and a test set: Obtain the equipment operation parameters, maintenance records, and environmental parameters of the production equipment in the historical time period, and perform missing value processing and outlier detection on the obtained data.

[0039] In this embodiment, if the production equipment is a machine tool, the equipment operation parameters include current, voltage, spindle speed, spindle temperature, and spindle vibration signal. In other embodiments, other types of operation parameters may also be included.

[0040] S202. Extract the time window features of the preprocessed data, and convert the failure occurrence time into a binary label (1 represents failure, 0 represents normal) for supervised learning.

[0041] The equipment operation parameters, maintenance records, and environmental parameters of the production equipment in the historical time period can be set as time series data, and a sliding window is set to traverse the time series data. The time window features may include: mean, maximum value, minimum value, etc. to reflect the short-term trend. The time window can be set to 24 hours or other appropriate durations. The data of one time window is one sample.

[0042] S203. Define the LSTM model and initialize the parameters of the LSTM model. The constructed LSTM network includes an input layer, an LSTM layer, a fully connected layer, and an output layer. The input layer is used to receive the time window features. The LSTM layer stacks 1 - 3 layers of LSTM cells, and each layer contains 64 - 256 neurons to capture the time dependence relationship. The fully connected layer maps the LSTM output to the failure probability, and the fully connected layer can use the Sigmoid activation function. The output layer is a single neuron used to output the failure probability value, and the value range of the failure probability value is 0 - 1.

[0043] The weights can be initialized using Xavier, and the biases are initialized to 0. Xavier (also known as Glorot initialization) is a neural network weight initialization method proposed by Xavier Glorot and Yoshua Bengio. Its core goal is to keep the variance of the activation values between network layers consistent, thereby accelerating the training convergence and avoiding gradient vanishing or explosion.

[0044] S204. Define the loss function and the optimizer.

[0045] The binary cross-entropy loss function can be selected as the loss function, and Adam as the optimizer. The binary cross-entropy loss function (BCE) is a core loss function used for binary classification tasks in deep learning, which measures the difference between the probability distribution predicted by the model and the true labels.

[0046] The Adam optimizer is an adaptive learning rate optimization algorithm that combines the ideas of the Momentum method and RMSProp, and normalizes the parameter updates to ensure that the magnitudes of each parameter update are similar, effectively improving the training effect.

[0047] S205. Forward propagation: Input the time window features into the LSTM model to obtain the fault probability and calculate the loss.

[0048] S206. Backward propagation: Calculate the gradient through the chain rule.

[0049] S207. Parameter update: The optimizer adjusts the weights according to the gradient.

[0050] S208. Iteratively perform forward propagation, backward propagation, and parameter update until the preset number of times is reached.

[0051] In one embodiment, to improve the quality of the collected true electricity consumption values, it further includes: processing the electricity consumption data obtained by the smart electricity meter, and the processing method includes: S301. Perform data cleaning on the electricity consumption data to handle format errors, missing values, duplicate records, and unreasonable data.

[0052] Data cleaning includes: checking the format of the electricity consumption data and detecting whether there are missing values, duplicate data, and error data in the electricity consumption data, and processing the detected missing values, duplicate data, and error data.

[0053] For missing electricity consumption data, linear interpolation, mean filling, spline interpolation, or time-weighted weights can be used to fill in the missing values to ensure the integrity of the data.

[0054] For duplicate data, delete the data with duplicate records at the same time point to ensure the uniqueness of the data at each time point.

[0055] For error data, these error data can be deleted or replaced with reasonable data values.

[0056] The reasonable data value can be the mean of adjacent data.

[0057] By performing output cleaning on the electricity consumption data, the quality of the data is ensured, and format errors, missing values, duplicate records, and unreasonable data are processed.

[0058] S302. Detect and correct outliers in the cleaned electricity consumption data.

[0059] The Z-score method or the box plot method can be used to detect outliers. For the corresponding outliers, they can be selected for removal or repaired using the mean of the previous and subsequent data points.

[0060] The process of detecting outliers using the Z-score method is as follows: Calculate the Z-score of each data point. If it exceeds the set threshold (such as 3 or -3), this point is regarded as an outlier.

[0061] The process of detecting outliers using the box plot method is as follows: Use the interquartile range (IQR) method to identify outliers.

[0062] By detecting and correcting outliers in the cleaned electricity consumption data, outliers caused by factors such as sensor failures and grid fluctuations can be detected and corrected.

[0063] S303. Perform data smoothing and filtering on the electricity consumption data.

[0064] The methods for data smoothing of the electricity consumption data can use simple moving average (SMA) or weighted moving average (WMA).

[0065] The process of performing data smoothing on the electricity consumption data using simple moving average is as follows: Apply simple moving average to the data and take the mean value within the window for smoothing. For example, a window of 7 days or 30 days can be used to smooth the electricity consumption data to remove short-term fluctuations.

[0066] The process of performing data smoothing on the electricity consumption data using weighted moving average is as follows: Assign different weights to each data point. Usually, the data points closer to the current time have larger weights, thereby reducing the interference of historical data on the current trend.

[0067] The method for filtering the electricity consumption data can use the median filtering method. The process is as follows: For isolated outliers, use median filtering to replace them.

[0068] Median filtering is very effective in eliminating sudden power fluctuations and can smooth short-term anomalies in the data.

[0069] By performing data smoothing and filtering on the electricity consumption data, short-term noise in the data can be removed and a smoother long-term trend can be extracted.

[0070] In one embodiment, the predicted value of the average temperature for the next date is obtained using weather forecasts.

[0071] In one embodiment, the weighted coefficient expression for the underestimation error is: ; In the formula, is the penalty strength factor for the underestimation error, is the probability that the production equipment fails on the nth date, The specific value of can be obtained based on experiments.

[0072] Since when the production equipment fails, it needs to be repaired and then put back into operation, its power consumption will increase within a day (that is, it will cause the enterprise's power consumption to increase on this day). By setting the weighting coefficient of the underestimation error to increase as the failure probability increases, it is thus realized that as the equipment failure probability increases, the prediction model is more sensitive to the underestimation error, making the power consumption prediction model more adaptable to the actual situation of higher power consumption caused by equipment failures.

[0073] In one embodiment, the expression for the weighting coefficient for the overestimation error is: ; In the formula, is the penalty strength factor for the overestimation error, is the probability that the production equipment fails on the nth date, The specific value of can be obtained based on experiments.

[0074] By making the weighting coefficient for the overestimation error negatively correlated with the equipment failure probability, when the equipment failure probability increases, the penalty for over-prediction is reduced, thus helping to avoid insufficient power supply for the enterprise.

[0075] Embodiment of the power consumption prediction device based on an intelligent electricity meter: The present invention also provides a power consumption prediction device based on an intelligent electricity meter. As Figure 2 shown, the power consumption prediction device based on an intelligent electricity meter includes a processor and a memory, and the memory stores computer program instructions. When the computer program instructions are executed by the processor, it realizes a power consumption prediction method based on an intelligent electricity meter according to an embodiment of the present invention above.

[0076] The power consumption prediction device based on an intelligent electricity meter further includes other components well-known to those skilled in the art such as a communication bus and a communication interface, and their settings and functions are known in the art, so they will not be elaborated here.

[0077] In the description of this specification, "a plurality" means at least two, for example, two, three or more, etc., unless otherwise specifically defined.

[0078] Although this specification has shown and described several embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many variations, changes and alternative means will occur to those skilled in the art without departing from the spirit and scope of the present invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in practicing the invention.

Claims

1. A method for predicting power consumption based on an intelligent electricity meter, characterized in that, Including: Obtaining the predicted average temperature value for the next date; Obtaining the planned operating duration of the enterprise's production equipment for the next date; Using an intelligent electricity meter to obtain the true value of the enterprise's electricity consumption for a continuous preset number of days; Inputting the predicted average temperature value for the next date and the planned operating duration into a preset electricity consumption prediction model to obtain the predicted value of the enterprise's electricity consumption for the next date; The training set of the electricity consumption prediction model consists of the true values of electricity consumption, average temperature, and operating duration for the continuous preset number of days. When training the electricity consumption prediction model, the loss function corresponding to a certain date is positively correlated with the deviation between the true value and the predicted value of electricity consumption on that date, and is positively correlated with the ratio of the planned operating duration and the actual operating duration of the production equipment on that date.

2. The electricity consumption prediction method based on an intelligent electricity meter according to claim 1, characterized in that The expression of the loss function used during training is: ； In the formula, represents the loss function corresponding to the nth date, and are the weighting coefficients for the underestimated error and the overestimated error respectively, , and are the actual operating duration and the planned operating duration of the production equipment on the nth date respectively, and are the true value and the predicted value of the power consumption on the nth date respectively.

3. The power consumption prediction method based on an intelligent electricity meter according to claim 2, wherein, Also including: Modifying the loss function, and the calculation expression of the modified loss function is: ； Wherein, is the probability that the production equipment fails on the nth date, represents the corrected loss function.

4. The power consumption prediction method based on an intelligent electricity meter according to claim 1, wherein The predicted average temperature value for the next date is obtained using weather forecasts.

5. The electricity consumption prediction method based on an intelligent electricity meter according to claim 2, wherein The weighted coefficient expression for the underestimation error is as follows: ; In the formula, is the penalty strength factor for underestimation error, is the probability that the production equipment fails on the nth date.

6. The power consumption prediction method based on an intelligent electricity meter according to claim 2, wherein The weighted coefficient expression for the overestimation error is as follows: ; In the formula, is the penalty strength factor for the overestimation error, is the probability that the production equipment fails on the nth date.

7. The power consumption prediction method based on an intelligent electricity meter according to claim 1, wherein The continuous preset number of days is 50 days.

8. The electricity consumption prediction method based on an intelligent electricity meter according to any one of claims 1 to 7, characterized in that, Also including: Processing the electricity consumption data obtained using an intelligent electricity meter, and the processing method includes: Performing data cleaning on the electricity consumption data; data cleaning includes: checking the format of the electricity consumption data and detecting whether there are missing values, duplicate data, and error data in the electricity consumption data, and processing the detected missing values, duplicate data, and error data; Detecting and correcting outliers in the cleaned electricity consumption data; Performing data smoothing and filtering processing on the electricity consumption data.

9. The electricity consumption prediction method based on an intelligent electricity meter according to claim 8, wherein, For the missing electricity consumption data, linear interpolation is used to fill it.

10. An electricity consumption prediction device based on an intelligent electricity meter, comprising a memory and a processor, wherein computer program instructions are stored in the memory, and characterized in that, When the computer program instructions are executed by a processor, the electricity consumption prediction method based on an intelligent electricity meter described in any one of claims 1 to 9 is implemented.

Images