Electric energy consumption prediction method and device used in EMS system, and computer equipment

Abstract

The invention relates to an electric energy consumption prediction method and device used in an EMS system, and computer equipment, and belongs to the technical field of computers. The electric energy consumption prediction method comprises the following steps: acquiring electric data; combining the three-phase unbalance degree and the power characteristics to gather a steady-state data set; respectively calculating to obtain an original value interval and a prediction transformation value interval, and further obtaining a steady-state prediction interval; and comparing actual data through a steady-state prediction interval to obtain a trend correction amount and a dynamic confidence coefficient, and obtaining prediction data in combination with a moving average method and a historical experience range. The electrical energy consumption prediction system comprises an electrical data acquisition module, a steady-state data summarization module, an original value interval acquisition module, an aggregation module, a prediction transformation value interval acquisition module, a steady-state prediction interval acquisition module, a comparison module and a prediction data generation module. According to the method, data of each hour within 48 hours in the future can be accurately predicted, and the prediction precision, the dynamic adaptability and the operation safety are balanced.

Description

Technical Field

[0001] The present invention belongs to the field of computer technology, and relates to a method, device, computer equipment, computer storage medium and computer program product for predicting electric energy consumption in an EMS system. Background Art

[0002] The EMS energy system is a system for effectively monitoring, analyzing and optimizing energy. By integrating various energy devices and data sources, it realizes the efficient utilization and refined management of energy, thereby improving energy utilization efficiency, reducing energy costs, and ensuring the stability and reliability of energy supply. Against the backdrop of the continuous adjustment of the global energy structure and the increasing demand for efficient energy management, the EMS energy system, as a key platform for realizing the optimal allocation and efficient utilization of energy, has become increasingly important.

[0003] Currently, the common methods for predicting electric energy consumption in EMS energy systems mainly include time series analysis and regression analysis methods. The prediction method of time series analysis arranges energy data in chronological order, and constructs a prediction model by analyzing the characteristics of the sequence such as trends, seasonality and periodicity. However, this method has high requirements for data stationarity, and real-world energy data is vulnerable to sudden factors such as extreme weather and policy mutations, resulting in the destruction of data stationarity and a significant decrease in prediction accuracy. The prediction method of regression analysis is to construct a multiple regression model by combining multiple variables, display the relationships between variables, clarify the influence direction and degree of each factor on energy consumption, and achieve prediction. It has low requirements for data stationarity and can handle energy data with volatility and trends. However, it is usually based on certain assumptions, such as linear relationships and the independence of error terms. However, in actual energy systems, these assumptions are often difficult to fully meet, which may lead to deviations between the model and the actual situation and affect prediction accuracy.

[0004] Therefore, developing a precise, efficient and stable method for predicting electric energy consumption has become an urgent need to improve the level of EMS systems, ensure reliable power supply and promote sustainable energy development. Summary of the Invention

[0005] In order to solve the technical problems in the above background art, the present invention provides a method, device, computer equipment, computer storage medium and computer program for predicting electric energy consumption in an EMS system.

[0006] The technical solutions of the present invention for solving the above technical problems are as follows: In the first aspect, a method for predicting electric energy consumption in an EMS system is provided. The method includes the steps of: Collecting voltage, current, active power and reactive power within a preset historical time range before the target time point; Calculate the three-phase unbalance degree of electricity and voltage, extract the power characteristics of active power and reactive power, and then summarize the steady-state data set by combining the three-phase unbalance degree and power characteristics; Calculate the first baseline value and the first range fluctuation value of the steady-state data, and then obtain the original value interval through the proportional relationship between the first baseline value and the first range fluctuation value; Aggregate the steady-state data within the target time range. The aggregation process is to take the extreme value average of all stable data within each hour to obtain the aggregated data; Calculate the second baseline value and the second range fluctuation value of the aggregated data, and then combine the second baseline value and the second range fluctuation value to obtain the predicted transformation value interval; Take the intersection of the original value interval and the predicted transformation value interval to obtain the steady-state prediction interval; Collect the actual data at least 24 hours before the target time point, compare it with the steady-state prediction interval, and obtain the trend correction amount and the dynamic confidence coefficient; Collect the steady-state data at least 24 hours before the target time point, obtain the prediction confidence interval through the moving average method, the trend correction amount and the dynamic confidence coefficient, and then output the prediction data by truncating the historical experience range.

[0007] In the second aspect, a virtual machine configuration information conversion system is provided. The system includes: An electrical data acquisition module for acquiring voltage, current, active power, and reactive power within a preset historical time range before the target time point; A steady-state data summarization module for calculating the three-phase unbalance degree of electricity and voltage, extracting the power characteristics of active power and reactive power, and then summarizing the steady-state data set by combining the three-phase unbalance degree and power characteristics; An original value interval obtaining module for calculating the first baseline value and the first range fluctuation value of the steady-state data, and then obtaining the original value interval through the proportional relationship between the first baseline value and the first range fluctuation value; An aggregation module for aggregating the steady-state data within the target time range. The aggregation process is to take the extreme value average of all stable data within each hour to obtain the aggregated data; A predicted transformation value interval obtaining module for calculating the second baseline value and the second range fluctuation value of the aggregated data, and then combining the second baseline value and the second range fluctuation value to obtain the predicted transformation value interval; A steady-state prediction interval obtaining module for taking the intersection of the original value interval and the predicted transformation value interval to obtain the steady-state prediction interval; A comparison module for collecting the actual data at least 24 hours before the target time point, comparing it with the steady-state prediction interval, and obtaining the trend correction amount and the dynamic confidence coefficient; The predictive data generation module is used to collect the steady-state data for at least 24 hours before the target time point, obtain the prediction confidence interval through the moving average method, the trend correction amount, and the dynamic confidence coefficient, and then truncate the output prediction data through the historical experience range.

[0008] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the power consumption prediction method for the EMS system described in any one of the above are implemented.

[0009] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the power consumption prediction method for the EMS system described in any one of the above is implemented.

[0010] In a fifth aspect, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps of the power consumption prediction method for the EMS system described in any one of the above are implemented. The beneficial effects of the present invention are as follows: First, through the multi-criterion steady-state screening mechanism, the present invention extracts the real load characteristics and eliminates the instantaneous disturbance interference. Secondly, through the intersection operation of the original value interval and the predicted transformation value interval, a double verification mechanism is formed to obtain a steady-state prediction interval that conforms to the actual situation. Then, the dynamic confidence coefficient and the trend correction amount are constructed through the steady-state prediction interval to achieve the double self-adaptation of the method parameters to short-term fluctuations and long-term gradual changes. Finally, through the truncation of the historical extreme value interval, the prediction output is strictly constrained within the physical safety boundary of the device. This method can better balance the prediction accuracy, dynamic adaptability, and operation safety, and also has a significant improvement compared with the traditional scheme in terms of risk control and long-term stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0012] Figure 1 It is a schematic flowchart of the power consumption prediction method for the EMS system provided in this embodiment; Figure 2 It is a graph showing the acquisition results of the hourly energy consumption values in the past 24 hours provided in this embodiment; Figure 3 It is a graph showing the prediction results of the hourly energy consumption values in the next 48 hours provided in this embodiment; Figure 4The comparison chart of hourly energy consumption values for the past 24 hours and the next 48 hours provided by this embodiment; Figure 5 The schematic structural diagram of the electric energy consumption prediction system used in the EMS system provided by this embodiment; Figure 6 The schematic structural diagram of the summary steady-state data module provided by the embodiment of the present invention; Figure 7 The schematic structural diagram of the obtained original value interval module provided by the embodiment of the present invention; Figure 8 The schematic structural diagram of the obtained predicted transformation value interval module provided by the embodiment of the present invention; Figure 9 The schematic structural diagram of the electronic device provided by the embodiment of the present invention.

[0013] In the drawings, the list of components represented by each reference numeral is as follows: 2001, Electrical data acquisition module; 2002, Summary steady-state data module; 2003, Obtained original value interval module; 2004, Aggregation module; 2005, Obtained predicted transformation value interval module; 2006, Obtained steady-state prediction interval module; 2007, Comparison module; 2008, Generate prediction data module; 20021, Three-phase unbalance degree calculation unit; 20022, Feature extraction unit; 20023, Screening time period unit; 20024, Alignment and fusion unit; 20031, First reference line value calculation unit; 20032, First range fluctuation value calculation unit; 20033, Original value interval calculation unit; 20051, Obtained difference data unit; 20052, Second reference line value calculation unit; 20053, Second range fluctuation value calculation unit; 20054, Predicted transformation value interval calculation unit; 310, Processor; 320, Communication interface; 330, Memory; 340, Communication bus. Detailed implementation manners

[0014] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0015] The current power consumption prediction methods in EMS energy management systems mainly rely on time series analysis and regression analysis, but face significant limitations in practical applications. Time series methods, such as ARIMA and exponential smoothing, rely on the assumption of data stationarity. However, in real energy systems, sudden or gradual factors such as extreme weather, policy regulation, and equipment aging can easily interfere, resulting in non-stationary fluctuations in data. For example, photovoltaic power output is affected by sudden changes in sunny and cloudy conditions, and industrial loads change suddenly due to production line upgrades. Traditional methods are difficult to dynamically respond to such disturbances, leading to a sharp increase in prediction errors. Although regression analysis can introduce multiple variables, such as temperature and humidity, to construct multivariate relationships, its linear assumption and requirement for error independence are difficult to fit the non-linear and time-varying coupling characteristics of actual systems. For example, the lag relationship between air conditioning load and temperature, and the non-linear decay of motor efficiency with the aging rate. The generalization ability of this method is limited. In addition, existing methods generally lack consideration of physical safety boundaries, and prediction results may exceed the equipment's bearing limit, such as the failure to intercept transformer overload predictions.

[0016] Therefore, developing an accurate, efficient, and stable power consumption prediction method has become an urgent need to improve the level of EMS systems, ensure reliable power supply, and promote sustainable energy development.

[0017] In view of the above problems, an embodiment of the present invention provides a power consumption prediction method for an EMS system. Figure 1 The following is a schematic flow diagram of the power consumption prediction method for an EMS system provided by an embodiment of the present invention. As Figure 1 shown, this method includes: Step S101, collect voltage, current, active power, and reactive power within a preset historical time range before the target time point.

[0018] It can be understood that the data source in this embodiment can be a SCADA system, a smart meter, a synchronous vector measurement device, etc., and this embodiment does not make additional limitations in this regard. The collection method can be through API interfaces or data queries, and timed scraping, or a lightweight calculation module can be pre-installed in the smart meter or gateway device to calculate and cache active / reactive power in real time. This embodiment also does not make additional limitations in this regard.

[0019] It should be noted that affected by seasonal changes, holidays, equipment aging and maintenance cycles, extreme value estimation involved in the IEEE1366 standard, and accidental events, the preset historical time range in this embodiment is preferably at least 6 months. Long-term data can smooth noise and can also cover the stable convergence time of maintenance cycles and load extreme values.

[0020] It is also worth noting that in this embodiment, four types of data are mainly collected. In terms of prediction, voltage is the core indicator of the stability of power grid operation. Voltage dips indicate overload risks and need to be warned in advance; current directly reflects the line load and the health status of equipment. A sudden increase in current may indicate a short circuit or insulation aging; active power represents the actual consumption of electrical energy and determines the electricity cost of users. It is the core target variable for energy consumption prediction; reactive power reflects the reactive power compensation demand and power factor of the system, can predict the change of capacitive / inductive load, and helps optimize the capacitor switching strategy. Among these parameters, voltage and current are the most basic monitoring parameters and are measured at almost all power nodes; active / reactive power can be directly calculated by an intelligent meter without the need to invest in dedicated equipment, which is very convenient and can also avoid parameter redundancy. Therefore, collecting these four types of data can effectively balance physical meaning, economy, and data availability.

[0021] Step S102: Calculate the three-phase unbalance degree of electricity and voltage, extract the power characteristics of active power and reactive power, and then summarize the steady-state data set by combining the three-phase unbalance degree and power characteristics.

[0022] Step S103: Calculate the first baseline value and the first range fluctuation value of the steady-state data, and then obtain the original value interval through the proportional relationship between the first baseline value and the first range fluctuation value.

[0023] Step S104: Aggregate the steady-state data within the target time range. The aggregation process is to take the extreme value average of all stable data within each hour to obtain the aggregated data.

[0024] Step S105: Calculate the second baseline value and the second range fluctuation value of the aggregated data, and then combine the second baseline value and the second range fluctuation value to obtain the predicted transformation value interval.

[0025] Step S106: Take the intersection of the original value interval and the predicted transformation value interval to obtain the steady-state prediction interval.

[0026] Step S107: Collect the actual data for the 24 hours before the target time point, compare it with the steady-state prediction interval, and obtain the trend correction amount and the dynamic confidence coefficient.

[0027] Step S108: Collect the steady-state data for the 24 hours before the target time point, obtain the prediction confidence interval through the moving average method, the trend correction amount, and the dynamic confidence coefficient, and then output the prediction data by truncating the historical experience range.

[0028] In the embodiments of the present invention, first, a multi-criterion steady-state screening mechanism is used to extract real load characteristics and eliminate instantaneous disturbance interference; second, a dual verification mechanism is formed through the intersection operation of the original value interval and the predicted transformation value interval to obtain a steady-state prediction interval that conforms to the actual situation; then, a dynamic confidence coefficient and a trend correction amount are constructed through the steady-state prediction interval to achieve the dual adaptability of the method parameters to short-term fluctuations and long-term gradual changes; finally, the prediction output is strictly constrained within the physical safety boundary of the device through historical extreme value interval truncation. This method can better balance prediction accuracy, dynamic adaptability, and operation safety, and also has a significant improvement compared with traditional solutions in terms of risk control and long-term stability.

[0029] Based on the above embodiments, in this method, step S102 further includes: Step S1021, calculating the three-phase unbalance degree of voltage by using the negative sequence component method or the maximum value deviation method, and the three-phase unbalance degree of current; Step S1022, extracting features of active power and reactive power, and statistically calculating the power peak-valley average value and the reactive power range within each hour; Step S1023, screening out the time period data with stable operation through a preset steady-state criterion; Step S1024, aligning and fusing the time period data, the three-phase unbalance degree, and the power characteristics in the time dimension to obtain a steady-state data set.

[0030] It can be understood that calculating the three-phase voltage can be the ratio of the negative sequence component to the positive sequence component, and setting a threshold value, which can exclude the voltage imbalance time period caused by grid faults or load mutations; calculating the three-phase current can be the deviation percentage of the maximum value to the average value, which can identify the situation of equipment aging or uneven three-phase load distribution.

[0031] It can also be understood that by statistically calculating the relative deviation between the maximum active power and the minimum value within the time window, the active power volatility and the reactive power volatility are obtained, which can filter out the instantaneous power fluctuations caused by equipment start-stop or impact loads, and at the same time allow the normal switching operation of reactive power compensation equipment, and exclude abnormal over-compensation / under-compensation.

[0032] In addition, further, for the screening operation through the preset steady-state criterion in step S1023, specifically, the historical data can be segmented by hour and analyzed segment by segment; calculate the three-phase voltage unbalance degree V unbalance 、the three-phase current unbalance degree I unbalance 、the active power peak-valley average value ΔP%, and the reactive power range ΔQ% for the data within each hour. If all parameters are lower than the threshold value, it is marked as a steady-state time period; then delete all the time period data that fails the verification.

[0033] Regarding the alignment and fusion in step S1024, it can specifically be as follows: For each steady state period, extract the following features and align them according to the time stamp: the moving window mean of the voltage / current unbalance degree, the peak-valley mean of the active power, and the range of the reactive power. Then summarize the above features on an hourly basis to form a structured steady state data set. The example format is as follows: By excluding transient interference data, the steady state data set can better reflect the true steady state law. At the same time, features such as the peak-valley mean of the active power and the range of the reactive power in the steady state data set are directly related to the equipment energy efficiency and compensation strategy, making the prediction results easier for operation and maintenance personnel to understand and apply.

[0034] Therefore, the steady state data set in this embodiment separates the "healthy steady state" from the "noise transient" in the original data through the three-phase unbalance criterion and power feature extraction, providing an input with a high signal-to-noise ratio for subsequent prediction.

[0035] Based on the above embodiment, in this method, step S103 further includes: Step S1031, calculate the average value of all steady state data to obtain the first baseline value; Step S1032, traverse the steady state data to obtain the peak value and the valley value respectively, and then calculate the range fluctuation value through the peak value and the valley value; Step S1033, based on the proportionality coefficient k, combine the first baseline value and the first range fluctuation value to obtain the original value interval.

[0036] It can be understood that as mentioned above, the transient interference is excluded from the data by setting a certain time window and filtering conditions for the steady state data. For the baseline value and the range fluctuation value, since the transient interference has been filtered out from the steady state data, the influence of the transient interference on the calculation can be avoided. The baseline value can more accurately reflect the true load level of the system, and the range fluctuation value only captures the reasonable fluctuations during normal operation, so it can also more accurately reflect the fluctuation range of the system under normal operation. At the same time, the time continuity of the steady state data ensures that the changes in the baseline value and the range fluctuation value can be strictly aligned with the equipment maintenance log, and will also provide a reliable causal relationship for equipment life prediction.

[0037] Regarding the proportionality coefficient k, first, it is necessary to ensure that the adjusted interval of the proportionality coefficient k does not exceed the rated parameters of the equipment to avoid the mathematical prediction deviating from the physical constraints. In terms of specific values, it can be dynamically adjusted according to the standard deviation or range of the historical steady state data. When the fluctuation is large, k is appropriately increased to cover the uncertainty, and when the operation is stable, k is reduced to improve the accuracy. It can also be adjusted according to the requirements of safety classification. In high-reliability scenarios, such as nuclear power plants, k = 2.5 - 3.0 is taken to expand the safety margin. In conventional scenarios, such as commercial buildings, k = 1.5 - 2.0 is taken to balance the accuracy and risk. In addition, the proportionality coefficient k can be optimized by combining the recent prediction hit rate. The hit rate formula is: knew = k old × target hit rate / actual hit rate. This embodiment does not make specific limitations on this.

[0038] This embodiment can more accurately reflect the true load level of the system and the reasonable fluctuations during normal operation, and based on the method of the ratio of the baseline to the fluctuation value, by adjusting the proportionality coefficient k, it can dynamically adapt to the fluctuation characteristics of the data.

[0039] Based on the above embodiment, in this method, the step S104 further includes: Step S1041: Arrange the aggregated data in a time series, and calculate the difference values between adjacent aggregated data in sequence to obtain difference data; Step S1042: Calculate the average value of all the difference data to obtain the second baseline value; Step S1043: Traverse the difference data to obtain the peak value and the valley value respectively, and then calculate the second range fluctuation value through the peak value and the valley value; Step S1044: Based on the proportionality coefficient p, combine the second baseline value and the second range fluctuation value to obtain the predicted transformation value interval.

[0040] It can be understood that the aggregated data takes the average of the maximum and minimum values per hour as the representative data points, which can compress the fluctuation information within the time window while retaining the trend characteristics, realizing data dimensionality reduction and smoothing.

[0041] It can also be understood that calculating the average value of all the difference values in step S1042 is to reflect the overall trend direction and rate of system changes. Extracting the peak value and the valley value of the difference values and calculating the range in step S1043 is to quantify the maximum fluctuation ability of the system within adjacent time periods. The average value reflects the overall trend, and the range fluctuation value reflects the short-term perturbation. The combination of the two makes the prediction interval more in line with the actual working conditions. In addition, only basic statistics such as the mean and the range are required, without complex calculations, which is suitable for deployment on edge devices such as smart meters. Updating the difference values and intervals every hour also meets the high-frequency prediction requirements.

[0042] It should also be noted that the range fluctuation interval is calculated through real-time difference data, and it can also lay a foundation for automatically responding to system state changes.

[0043] This embodiment calculates the trend and the fluctuation separately, captures the long-term gradual change and the short-term perturbation respectively, and then combines the two in the prediction stage, which can generate a dynamic predicted transformation value interval, improve the prediction accuracy, and have the advantage of lightweight calculation.

[0044] Based on the above embodiment, in this method, the step S108 further includes: Step S1081: Collect historical steady-state data for at least 24 hours before the target time point, and use the moving average method to calculate the preliminary prediction value for the next hour. Step S1082: Superimpose a trend correction amount on the preliminary prediction value to obtain a corrected prediction value. Step S1083: Use a dynamic confidence coefficient to adjust the interval of the corrected prediction value to obtain a prediction confidence interval. Step S1084: Truncate the confidence prediction interval based on the historical experience range to obtain the final prediction data.

[0045] It can be understood that in this embodiment, the moving average method generates a preliminary prediction value through steady-state data. Based on the characteristic that steady-state data can reflect the true operating state of the system, the moving average method further smooths the remaining minor fluctuations and highlights the long-term trend.

[0046] It can also be understood that in this embodiment, the trend correction amount is to calculate the change trend of adjacent moving average values and perform a second moving average on the trend values, which can correct the lag of the prediction value.

[0047] It can also be understood that in this embodiment, the dynamic confidence coefficient is a confidence that is dynamically adjusted by continuously updating in combination with real-time data, and can adaptively adjust the tightness of the interval. The formula for the dynamic confidence coefficient is: CI = Yt ± n × σ, where n changes with the historical volatility and σ is the standard deviation. Centered on the prediction value after moving average and trend correction, the dynamic confidence coefficient can determine the interval elasticity and quantify the uncertainty to form a confidence interval.

[0048] It can further be understood that in this embodiment, the historical experience range sets a truncation threshold according to the distribution characteristics of historical data, eliminates prediction values that exceed the reasonable range, and ensures a safety margin. When making a specific selection, factors such as time span, data magnitude, and scenario matching can be combined. For example, data under the same model and similar load conditions are preferably selected. This embodiment does not make specific limitations in this regard.

[0049] Based on at least 24 hours of steady-state data, this embodiment generates an initial prediction value through the moving average method, eliminates the lag bias by combining the trend correction amount, then introduces a dynamic confidence coefficient to quantify the uncertainty to form a confidence interval, and finally truncates and corrects the prediction value through the historical experience range to ensure the rationality and reliability of the output result.

[0050] Next, we provide an experimental example to further illustrate this embodiment, which is specifically as follows: Table 1. Experimental environment configuration parameters In the experiment, as Figure 2 shown, collect the actual hourly energy consumption values for the past 24 hours, ranging from 1886 kWh to 3101 kWh, including all peak hours from 12 noon to 23:00 in the evening.

[0051] The experimental examples were implemented according to the method steps of the above-mentioned embodiments, and the same running code was written in Java and Vue and run on a computer. As Figure 3 and Figure 4 shown, the predicted hourly energy consumption values for the next 48 hours were obtained, ranging from 1986 kWh to 3369 kWh. After that, the predicted values were compared with the actual values, the absolute value of the percentage error was calculated hour by hour, and then the average value over 48 hours was taken. The error calculation formula is: |(actual value - predicted value) / actual value| × 100%. The final average error over 48 hours was 0.93%, meeting the requirement of ≤ 1%.

[0052] Thus, it can be verified that the method of this embodiment can accurately predict the data for each hour within the next 48 hours, and the deviation between the prediction result and the actual value does not exceed 1% on average.

[0053] As Figure 5 shown, in one embodiment, an electric energy consumption prediction system for an EMS system is provided. The system includes: An electrical data acquisition module 2001 for acquiring voltage, current, active power, and reactive power within a preset historical time range before the target time point; A steady-state data summary module 2002 for calculating the three-phase unbalance degree of electricity and voltage, extracting the power characteristics of active power and reactive power, and then summarizing a steady-state data set by combining the three-phase unbalance degree and power characteristics; An original value interval obtaining module 2003 for calculating the first baseline value and the first range fluctuation value of the steady-state data, and then obtaining the original value interval through the proportional relationship between the first baseline value and the first range fluctuation value; An aggregation module 2004 for aggregating the steady-state data within the target time range. The aggregation process is to take the extreme value average of all stable data within each hour to obtain the aggregated data; A predicted transformation value interval obtaining module 2005 for calculating the second baseline value and the second range fluctuation value of the aggregated data, and then obtaining the predicted transformation value interval by combining the second baseline value and the second range fluctuation value; A steady-state prediction interval obtaining module 2006 for taking the intersection of the original value interval and the predicted transformation value interval to obtain the steady-state prediction interval; A comparison module 2007 for collecting the actual data at least 24 hours before the target time point, comparing it with the steady-state prediction interval, and obtaining a trend correction amount and a dynamic confidence coefficient; A predicted data generation module 2008 for collecting the steady-state data at least 24 hours before the target time point, obtaining a prediction confidence interval through the moving average method, the trend correction amount, and the dynamic confidence coefficient, and then outputting the predicted data by truncating the historical experience range.

[0054] As Figure 6 shown, the summarized steady-state data module 2002 includes: A three-phase unbalance degree calculation unit 20021 for calculating the three-phase unbalance degree of voltage and the three-phase unbalance degree of current by using the negative-sequence component method or the maximum deviation method; A feature extraction unit 20022 for extracting features from active power and reactive power, and statistically obtaining the peak-valley mean value of power and the reactive power range within each hour; A screening time period unit 20023 for screening out the time period data with stable operation through a preset steady-state criterion; An alignment and fusion unit 20024 for aligning and fusing the time period data, the three-phase unbalance degree, and the power features in the time dimension to obtain a steady-state data set.

[0055] As Figure 7 shown, the obtained original value interval module 2003 includes: A first baseline value calculation unit 20031 for calculating the average value of all steady-state data to obtain a first baseline value; A first range fluctuation value calculation unit 20032 for traversing the steady-state data to obtain peak values and valley values respectively, and then calculating the range fluctuation value through the peak values and valley values; An original value interval calculation unit 20033 for obtaining an original value interval based on a proportionality coefficient, in combination with the first baseline value and the first range fluctuation value.

[0056] As Figure 8 shown, the obtained predicted transformation value interval module 2005 includes An obtained difference data unit 20051 for arranging the aggregated data in a time series and sequentially calculating the difference values between adjacent aggregated data to obtain difference data; A second baseline value calculation unit 20052 for calculating the average value of all difference data to obtain a second baseline value; A second range fluctuation value calculation unit 20053 for traversing the difference data to obtain peak values and valley values respectively, and then calculating the second range fluctuation value through the peak values and valley values; A predicted transformation value interval calculation unit 20054 for obtaining a predicted transformation value interval based on a proportionality coefficient, in combination with the second baseline value and the second range fluctuation value.

[0057] In Figure 9 This is a schematic structural diagram of the electronic device provided by the embodiment of the present invention. As Figure 9As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 complete communication with each other through the communication bus 340. The processor 310 may call the logical instructions in the memory 330 to execute the following method: a method for predicting power consumption in the EMS system.

[0058] In addition, when the logical instructions in the above-mentioned memory 330 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, may be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0059] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, it implements the method for predicting power consumption in the EMS system as described in the embodiment.

[0060] In this embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, it implements the steps of the method for predicting power consumption in the EMS system as described in the embodiment.

[0061] The computer storage medium of the embodiments of the present invention may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the computer-readable storage medium include: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0062] The computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries the computer-readable program code. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0063] The program code contained on the computer-readable medium may be transmitted by any appropriate medium, including but not limited to wireless, wire, optical fiber cable, RF, etc., or any suitable combination of the above.

[0064] The computer program code for performing the operations of the present invention may be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, Ruby, Go, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0065] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the scope of the patent for the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention shall be subject to the appended claims.

Claims

1. A method for predicting power consumption in an EMS system, characterized in that, The method includes the steps of: Collecting voltage, current, active power, and reactive power within a preset historical time range before the target time point; Calculating the three-phase unbalance degree of electricity and voltage, extracting the power characteristics of active power and reactive power, and summarizing a steady-state data set by combining the three-phase unbalance degree and power characteristics; Calculating the first baseline value and the first range fluctuation value of the steady-state data, and obtaining the original value interval through the proportional relationship between the first baseline value and the first range fluctuation value; Aggregating the steady-state data within the target time range. The aggregation process is to take the extreme value average of all stable data within each hour to obtain the aggregated data; Calculating the second baseline value and the second range fluctuation value of the aggregated data, and obtaining the predicted transformation value interval by combining the second baseline value and the second range fluctuation value; Taking the intersection of the original value interval and the predicted transformation value interval to obtain the steady-state prediction interval; Collecting the actual data at least 24 hours before the target time point, comparing it with the steady-state prediction interval to obtain the trend correction amount and the dynamic confidence coefficient; Collecting the steady-state data at least 24 hours before the target time point, obtaining the prediction confidence interval through the moving average method, the trend correction amount, and the dynamic confidence coefficient, and then truncating the output prediction data through the historical experience range.

2. The method for predicting power consumption in the EMS system according to claim 1, wherein, The step of calculating the three-phase unbalance degree of electricity and voltage, extracting the power characteristics of active power and reactive power, and summarizing a steady-state data set by combining the three-phase unbalance degree and power characteristics further includes: Calculating the three-phase unbalance degree of voltage and the three-phase unbalance degree of current by using the negative sequence component method or the maximum deviation method; Performing feature extraction on active power and reactive power, and statistically calculating the power peak-valley average value and the reactive power range within each hour; Filtering out the data of the stable operation period through a preset steady-state criterion; Aligning and fusing the period data, the three-phase unbalance degree, and the power characteristics in the time dimension to obtain the steady-state data set.

3. The power consumption prediction method for an EMS system according to claim 1 or 2, characterized in that The step of calculating the first baseline value and the first range fluctuation value of the steady-state data, and obtaining the original value interval through the proportional relationship between the first baseline value and the first range fluctuation value further includes: Calculating the average value of all steady-state data to obtain the first baseline value; Traversing the steady-state data to obtain the peak value and the valley value, and then calculating the range fluctuation value through the peak value and the valley value; Based on the proportional coefficient, combining the first baseline value and the first range fluctuation value to obtain the original value interval.

4. The power consumption prediction method for an EMS system according to claim 1 or 2, characterized in that The step of calculating the second baseline value and the second range fluctuation value of the aggregated data, and obtaining the predicted transformation value interval by combining the second baseline value and the second range fluctuation value further includes: Arranging the aggregated data in a time series, and sequentially calculating the difference values between adjacent aggregated data to obtain the difference data; Calculating the average value of all difference data to obtain the second baseline value; Traversing the difference data to obtain the peak value and the valley value, and then calculating the second range fluctuation value through the peak value and the valley value; Based on the proportional coefficient, combining the second baseline value and the second range fluctuation value to obtain the predicted transformation value interval.

5. The method for predicting power consumption in an EMS system according to claim 1 or 2, characterized in that, The step of collecting the steady-state data at least 24 hours before the target time point, obtaining the prediction confidence interval through the moving average method, the trend correction amount, and the dynamic confidence coefficient, and then truncating the output prediction data through the historical experience range further includes: Collect historical steady-state data for at least 24 hours before the target time point, and calculate the preliminary prediction value for the next hour using the moving average method; Superimpose the trend correction amount on the preliminary prediction value to obtain the corrected prediction value; Use the dynamic confidence coefficient to adjust the interval of the corrected prediction value to obtain the prediction confidence interval; Based on the historical experience range, truncate the confidence prediction interval to obtain the final prediction data.

6. An electric energy consumption prediction system for an EMS system, characterized in that, The system includes: An electrical data collection module for collecting voltage, current, active power, and reactive power within a preset historical time range before the target time point; A steady-state data summary module for calculating the three-phase unbalance degree of electricity and voltage, extracting the power characteristics of active power and reactive power, and then summarizing the steady-state data set by combining the three-phase unbalance degree and power characteristics; An original value interval obtaining module for calculating the first baseline value and the first range fluctuation value of the steady-state data, and then obtaining the original value interval through the proportional relationship between the first baseline value and the first range fluctuation value; An aggregation module for aggregating the steady-state data within the target time range. The aggregation process is to take the extreme value average of all stable data within each hour to obtain the aggregated data; A predicted transformation value interval obtaining module for calculating the second baseline value and the second range fluctuation value of the aggregated data, and then combining the second baseline value and the second range fluctuation value to obtain the predicted transformation value interval; A steady-state prediction interval obtaining module for taking the intersection of the original value interval and the predicted transformation value interval to obtain the steady-state prediction interval; A comparison module for collecting the actual data for at least 24 hours before the target time point, comparing it with the steady-state prediction interval, and obtaining the trend correction amount and the dynamic confidence coefficient; A predicted data generation module for collecting the steady-state data for at least 24 hours before the target time point, obtaining the prediction confidence interval through the moving average method, the trend correction amount, and the dynamic confidence coefficient, and then truncating the output prediction data through the historical experience range.

7. The power consumption prediction system for an EMS system according to claim 6, characterized in that, The original value interval obtaining module includes: A first baseline value calculation unit for calculating the average value of all steady-state data to obtain the first baseline value; A first range fluctuation value calculation unit for traversing the steady-state data to obtain the peak value and the valley value, and then calculating the range fluctuation value through the peak value and the valley value; An original value interval calculation unit for obtaining the original value interval by combining the first baseline value and the first range fluctuation value based on the proportional coefficient; Among them, the predicted transformation value interval obtaining module includes A difference data obtaining unit for arranging the aggregated data in time series and sequentially calculating the difference values between adjacent aggregated data to obtain the difference data; A second baseline value calculation unit for calculating the average value of all difference data to obtain the second baseline value; A second range fluctuation value calculation unit for traversing the difference data to obtain the peak value and the valley value, and then calculating the second range fluctuation value through the peak value and the valley value; A predicted transformation value interval calculation unit for obtaining the predicted transformation value interval by combining the second baseline value and the second range fluctuation value based on the proportional coefficient.

8. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the power consumption prediction method for the EMS system according to any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the method for predicting power consumption in an EMS system according to any one of claims 1 to 5.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method for predicting power consumption in an EMS system according to any one of claims 1 to 5.

Images