Energy-saving optimization control method and system based on load curve
By unifying the energy consumption data format at the building terminal, collecting and integrating multi-source data, and building a baseline model, the problem of difficulty in integrating different data sources is solved, the accuracy and reliability of energy consumption prediction is improved, and more comprehensive energy consumption is provided analyze.
Patent Information
- Application Number
- CN202411576832.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-06
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The energy consumption data formats and structures of different sources are different, which leads to difficulty in data integration, affects the reliability of the data, and has missing, errors or inconsistencies, affecting the quality of the fused data.
By presetting the data unit type at the building terminal, unifying the energy consumption data format, collecting multi-source data, identifying time series, integrating data fusion and Kalman filtering, extracting feature parameters, building a baseline model, and performing data cleaning and verification. Generate a load line chart.
It realizes effective integration of multi-source data, improves the accuracy and reliability of energy consumption prediction, provides a more comprehensive view of energy consumption, helps identify factors affecting energy consumption, and continuously improves prediction accuracy through model optimization.
Smart Images

Figure CN119090096A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of energy consumption data processing, and in particular to an energy-saving optimization control method and system based on a load curve. Background Art
[0002] With the development of science and technology, more and more smart devices are installed in buildings. The most common ones are meters (electricity meters, water meters, gas meters, energy meters, PM2.5 meters, etc.). Smart meters often do not work alone and need to be configured with data management software to manage and analyze the data.
[0003] Since data from different sources (such as electricity consumption, weather data, market demand, etc.) have different formats and structures, it is necessary to establish unified standards and frameworks for integration. Multi-source data may be missing, erroneous or inconsistent, affecting the reliability of the integrated data. Therefore, an effective data cleaning and verification mechanism is needed. Summary of the invention
[0004] The present invention aims to solve the problem of how to effectively integrate data from different sources to optimize energy consumption load forecasting, and to provide an energy-saving optimization control method and system based on load curves.
[0005] The present invention adopts the following technical means to solve the technical problem: The present invention provides an energy-saving optimization control method based on a load curve, comprising: Based on the data unit type preset by the building terminal, the data format of the energy consumption data is unified into the data unit type, wherein the data unit type specifically includes temperature, power and timestamp; Determining whether the energy consumption data can be presented to the building terminal in real time; If not, then collect corresponding various multi-source data from different data sources through the preset sensors of the building terminal, identify the time series of the multi-source data, apply the preset data fusion according to the time series, integrate the multi-source data to obtain fused data, extract corresponding feature parameters from the fused data, and construct a valid feature set based on the feature parameters, wherein the time series specifically includes trend changes and periodic changes, and the feature parameters specifically include weather impact index, economic activity level and historical load data; Determining whether the valid feature set requires preset data cleaning; If not necessary, a preset correlation coefficient is used to collect the linear relationship between each feature in the effective feature set and the load variable, a corresponding baseline model is established on the effective feature set, the mean square error and determination coefficient of the baseline model are calculated, and the load prediction result of the building terminal is constructed through the baseline model. The load prediction result is compared with the actual load value to generate a load line chart.
[0006] Furthermore, before the step of collecting corresponding various multi-source data from different data sources through the sensors preset in the building terminal, the step further includes: Based on the preset working range of the sensor, dividing the collection area covered by the sensor in the building through the building terminal; Determining whether the acquisition area can detect a preset overlapping monitoring edge; If possible, an overlapping time window is constructed based on the overlapping monitoring edges, and the collected measurement values of multiple sensors in the overlapping time window are identified. Data features to be fused are generated based on the collected measurement values, and a preset Kalman filter is applied to assign weights to the collected measurement values, and the weighted average of the data features is calculated, wherein the data features specifically include temperature, humidity and energy consumption.
[0007] Furthermore, the step of constructing a valid feature set based on the feature parameters further includes: Based on the preset virtual variables, the seasonal characteristics of the energy consumption of the building terminals are constructed, and the ratio characteristics and the difference characteristics between the seasonal characteristics are calculated, wherein the virtual variables specifically include spring, summer, autumn and winter, the ratio characteristics are specifically the ratio of energy consumption to temperature, and the difference characteristics are specifically the difference between the current load and the past load; Determine whether there are missing values in the ratio feature and the difference feature; If not, the ratio feature and the difference feature are standardized to ensure that they are always on the same scale, and the ratio feature and the difference feature are combined with other features to form a composite feature, and the energy consumption impact corresponding to seasonal changes is captured according to the composite feature.
[0008] Furthermore, the step of establishing a corresponding baseline model on the effective feature set further includes: Based on the preset chart of energy consumption of the building terminal, the X-axis is drawn from the chart as the energy consumption prediction value, and the Y-axis is drawn from the chart as the energy consumption residual, and the energy consumption scatter plot is obtained by combining them; Determining whether the residuals on the energy consumption scatter plot are randomly distributed; If yes, the energy consumption scatter plot is input into the baseline model, and cross-validation is performed through the baseline model to obtain fluctuation information of the baseline model on different data subsets.
[0009] Furthermore, the step of determining whether the energy consumption data can be presented to the building terminal in real time further includes: Based on the preset network bandwidth of the building terminal, detecting the network delay information when the energy consumption data is transmitted; Determining whether the network delay information exceeds a preset delay period; If so, a timestamp tag is added in the data collection link and the data display link of the energy consumption data, and the overall delay of the energy consumption data is calculated according to the timestamp tag.
[0010] Furthermore, the step of determining whether the valid feature set requires preset data cleaning further includes: Based on the pre-statistical features of the effective feature set, the proportion of feature missing values is calculated, the distribution of the feature missing values is obtained according to the proportion, and a corresponding missing value heat map is generated; Determine whether the missing value heat map matches a preset random missing; If not, then identify the association information between the feature missing value and the pre-statistical feature, and collect the corresponding missing source based on the association information.
[0011] Furthermore, the step of unifying the data format of the energy consumption data into the data unit type based on the data unit type preset by the building terminal further includes: detecting original units in the energy consumption data, and identifying different units to be converted based on the original units; Determining whether the different units can be uniformly converted into the data unit type; If not, the inconsistent items of the different units are marked, and the corresponding data of the different units are converted row by row according to the inconsistent items, and the data distribution and statistical characteristics before and after the conversion are compared.
[0012] The present invention also provides an energy-saving optimization control system based on a load curve, comprising: A unification module, used for unifying the data format of the energy consumption data into the data unit type preset by the building terminal based on the data unit type, wherein the data unit type specifically includes temperature, power and timestamp; A determination module, used to determine whether the energy consumption data can be presented to the building terminal in real time; an execution module, for, if not, collecting corresponding various multi-source data from different data sources through the preset sensors of the building terminal, identifying the time series of the multi-source data, applying preset data fusion according to the time series, integrating the multi-source data to obtain fused data, extracting corresponding feature parameters from the fused data, and constructing a valid feature set according to the feature parameters, wherein the time series specifically includes trend changes and periodic changes, and the feature parameters specifically include weather impact index, economic activity level and historical load data; A second judgment module is used to judge whether the valid feature set requires preset data cleaning; The second execution module is used to use a preset correlation coefficient if it is not needed, collect the linear relationship between each feature in the effective feature set and the load variable, establish a corresponding baseline model on the effective feature set, calculate the mean square error and determination coefficient of the baseline model, construct the load prediction result of the building terminal through the baseline model, compare the load prediction result with the actual load value, and generate a load line graph.
[0013] Furthermore, it also includes: A division module, used to divide the collection area covered by the sensor in the building through the building terminal based on the preset working range of the sensor; A third judgment module is used to judge whether the acquisition area can detect a preset overlapping monitoring edge; The third execution module is used to, if possible, construct an overlapping time window according to the overlapping monitoring edges, identify the collected measurement values of multiple sensors in the overlapping time window, generate data features to be fused according to the collected measurement values, apply a preset Kalman filter to weight the collected measurement values, and calculate the weighted average of the data features, wherein the data features specifically include temperature, humidity and energy consumption.
[0014] Furthermore, the execution module also includes: A calculation unit, configured to construct seasonal characteristics of energy consumption of the building terminal based on preset virtual variables, and calculate ratio characteristics and difference characteristics between the seasonal characteristics, wherein the virtual variables specifically include spring, summer, autumn and winter, the ratio characteristics specifically are the ratio of energy consumption to temperature, and the difference characteristics specifically are the difference between current load and past load; A judging unit, used for judging whether there are missing values in the ratio feature and the difference feature; an execution unit, for, if not, standardizing the ratio feature and the difference feature to ensure that the ratio feature and the difference feature are always on the same scale, combining the ratio feature and the difference feature with other features to form a composite feature, and capturing the energy consumption impact corresponding to seasonal changes according to the composite feature.
[0015] The present invention provides an energy-saving optimization control method and system based on a load curve, which has the following beneficial effects: The present invention can form a more comprehensive energy consumption view by collecting multi-source data from different data sources. Integrating these data can provide richer context, help identify various factors affecting energy consumption, and improve the accuracy of energy-saving optimization. At the same time, feature parameters are extracted so that the model can be analyzed based on multiple dimensions. These features can reveal how different factors affect energy consumption, help better understand and predict load changes, and establish a baseline model on the effective feature set, and calculate the mean square error and determination coefficient, which can quantify the performance of the model. It not only provides a benchmark for energy consumption prediction, but also can continuously optimize model parameters by comparison with actual data to improve prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 It is a flow chart of an embodiment of the energy-saving optimization control method based on load curve of the present invention; Figure 2 It is a structural block diagram of an embodiment of the energy-saving optimization control system based on load curve of the present invention. DETAILED DESCRIPTION
[0017] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. The implementation of the objectives, functional features and advantages of the present invention will be further described in conjunction with the embodiments and with reference to the accompanying drawings.
[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0019] Reference Figure 1 , is an energy-saving optimization control method based on a load curve in an embodiment of the present invention, comprising: S1: based on the data unit type preset by the building terminal, unify the data format of the energy consumption data into the data unit type, wherein the data unit type specifically includes temperature, power and timestamp; S2: Determine whether the energy consumption data can be presented to the building terminal in real time; S3: If not, then collect corresponding various multi-source data from different data sources through the preset sensors of the building terminal, identify the time series of the multi-source data, apply the preset data fusion according to the time series, integrate the multi-source data to obtain fused data, extract corresponding feature parameters from the fused data, and construct a valid feature set according to the feature parameters, wherein the time series specifically includes trend changes and periodic changes, and the feature parameters specifically include weather impact index, economic activity level and historical load data; S4: Determine whether the valid feature set requires preset data cleaning; S5: If not necessary, use the preset correlation coefficient to collect the linear relationship between each feature in the effective feature set and the load variable, establish a corresponding baseline model on the effective feature set, calculate the mean square error and determination coefficient of the baseline model, construct the load prediction result of the building terminal through the baseline model, compare the load prediction result with the actual load value, and generate a load line graph.
[0020] In this embodiment, the system unifies the data format of the energy consumption data collected in the building into the data unit type based on the data unit type pre-set by the building terminal, which specifically includes temperature, electricity and timestamp, and then the system determines whether these energy consumption data can be presented to the building terminal in real time to execute the corresponding steps; for example, when the system determines that the collected energy consumption data can be presented to the building terminal in real time, the system will consider that the data meets the requirements in terms of quality and can effectively reflect the actual energy consumption situation. The system will display the collected energy consumption data in real time on the building terminal, such as energy consumption curve graphs, pie charts or bar graphs, so that users can quickly obtain information, and at the same time, according to historical data The system automatically issues an alarm when a certain energy consumption exceeds the normal range, and records the real-time data in the database for subsequent analysis and query to ensure the integrity and traceability of the data, conduct dynamic data analysis, generate trend charts or prediction models, and help managers identify energy consumption patterns and potential abnormal situations. For example, when the system determines that the collected energy consumption data cannot be presented to the building terminal in real time, the system will consider that it cannot reflect the actual energy consumption. The system will collect the corresponding multi-source data from different energy consumption data sources through the sensors pre-set at the building terminal, identify the time series of these multi-source data, Time series specifically include trend changes and cyclical changes. According to the pre-set data fusion of time series applications, multi-source data are integrated to obtain fused data, and corresponding feature parameters are extracted from the fused data. The feature parameters specifically include weather impact index, economic activity level and historical load data. An effective feature set is constructed based on these feature parameters. The system can capture the dynamic characteristics of energy consumption data by identifying and analyzing trend changes and cyclical changes in time series. This process not only helps to understand past energy consumption patterns, but also provides an important basis for future energy consumption forecasts. Through in-depth analysis of time series, managers can identify seasonal fluctuations, economic activities and other factors. Peaks and valleys can be detected, thereby optimizing energy consumption strategies in specific periods of time. At the same time, during the fusion process, the system integrates data from different sources into a unified data set through a preset data fusion algorithm. This process helps to eliminate noise and redundancy in the data and improve data quality. Through data fusion, managers can obtain more accurate information, which in turn better supports energy consumption analysis and decision-making. Based on the extracted feature parameters, an effective feature set is constructed to lay the foundation for subsequent prediction models. Through the effective feature set, the model can capture the key driving factors of building energy consumption and improve the accuracy of prediction. Then the system determines whether the effective feature set requires pre-set data cleaning to execute the corresponding steps.For example, when the system determines that the effective feature set requires pre-set data cleaning, the system will believe that the data in the effective feature set may contain errors, outliers or noise. These problems will affect the accuracy of subsequent analysis and model building. The system will clarify the cleaning goals, such as removing duplicate data, filling missing values, processing outliers, etc., to ensure that all cleaning steps are consistent with the needs of subsequent analysis. At the same time, check whether there are missing values in the feature set, decide how to deal with these missing values, and ensure that the units, formats and types of all features are consistent. For example, the temperature unit may need to be unified as Celsius or Fahrenheit, and the power unit needs to ensure consistency. By identifying and removing duplicate data, the uniqueness of the feature set is ensured and the interference of redundancy on the analysis is reduced. For example, when the system determines that the effective feature set does not require pre-set data cleaning, the system will believe that the data in the effective feature set does not contain errors, outliers or noise. The system will use the pre-set correlation coefficient to collect the linear relationship between each feature in the effective feature set and the load variable, establish the corresponding baseline model on the effective feature set, calculate the mean square error and determination coefficient of the baseline model, and then use The baseline model constructs the load forecast results of the building terminal, compares the load forecast results with the actual load value, and generates a load line chart; the system can save the time and resources required for data cleaning by directly using the effective feature set, speed up the model establishment, and enable the system to respond to changes quickly and provide load forecast results in a timely manner. At the same time, through the preset correlation coefficient, the system can analyze the linear relationship between each feature in the effective feature set and the load variable. This analysis can help managers better understand the key factors affecting building energy consumption and guide subsequent management decisions. The baseline model established based on the linear relationship usually has good interpretability. Managers can clearly see the impact of each feature on the load forecast, so as to make more targeted adjustments. By establishing a baseline model, the system can calculate the mean square error and determination coefficient of the load forecast. These indicators provide a quantitative evaluation of the accuracy of the forecast results, which helps to optimize energy consumption management. By comparing the load forecast results with the actual load value, the gap between the forecast and the actual can be quickly identified. This feedback mechanism enables the system to adjust the energy consumption management strategy in a timely manner and improve resource utilization efficiency. ;
[0021] It should be noted that the time series of the multi-source data is identified, and the preset data fusion is applied according to the time series to integrate the multi-source data to obtain fused data. The specific examples are as follows: Assuming that energy consumption forecasting is carried out in a large commercial building, the data sources include weather, economic activities and historical energy consumption data. Taking October 1, 2024 as an example, how to fuse multi-source data through time series will be explained in detail; Collect data, assuming the following information is collected from different data sources: Weather data (hourly): 08:00 - Temperature: 10°C, Humidity: 60%; 09:00 - Temperature: 12°C, Humidity: 55%; 10:00 - Temperature: 14°C, Humidity: 50%; Economic activity data (hourly): 08:00 - Passenger flow: 1,000 people, business activity index: 0.6; 09:00 - Passenger flow: 1,200 people, business activity index: 0.7; 10:00 - Passenger flow: 1,500 people, business activity index: 0.8; Historical load data (hourly): 08:00 - Energy consumption: 200kWh; 09:00 - Energy consumption: 220kWh; 10:00 - Energy consumption: 240kWh; Identify time series and align the time points of each data source to ensure that all data are in the same time range. For example, all data are timestamped at 08:00, 09:00, and 10:00 on October 1, 2024. Apply data fusion. In this phase, all data is integrated into a unified dataset. The following are the specific steps: Align timestamps: All data is at the same time point and no adjustment is required; Create a fused data table: Integrate the features of each data source into one table, where each row represents the data at a time point: Timestamp Temperature (°C) humidity(%) Passenger flow (people) Business Activity Index Energy consumption (kWh) 2024-10-01 08:00 10 60 1000 0.6 200 2024-10-01 09:00 12 55 1200 0.7 220 2024-10-01 10:00 14 50 1500 0.8 240 Weighted average: Assuming that temperature has a greater impact on energy consumption than humidity and economic activity index, weights can be set for different features, for example: Temperature: 0.5; Humidity: 0.2; Passenger flow: 0.2; Business activity index: 0.1; A weighted composite feature can be calculated using these weights, for example: Comprehensive index = 0.5*temperature+0.2*humidity+0.2*passenger flow / 1000+0.1*commercial activity index Obtaining fused data. Through the above steps, the system obtains fused data, and this fused data set will provide support for subsequent load forecasting analysis; Data analysis and application, feature extraction, based on the fused data, extract relevant features to establish a load forecasting model. These features include: temperature, passenger flow, and commercial activity index; Modeling and prediction: Use the fused data set to train a prediction model, such as linear regression or decision tree, to predict future energy consumption. For example, use the fused data from the past three hours to predict the energy consumption for the next hour. After the results are visualized, a load line chart is generated to compare the actual energy consumption with the model-predicted energy consumption to evaluate the effectiveness of the model; In summary, through the above steps, the system effectively integrates multi-source data to form a complete analysis basis. Such data fusion method not only improves the utilization efficiency of data, but also provides higher accuracy and reliability for energy consumption prediction.
[0022] Then, a preset correlation coefficient is used to collect the linear relationship between each feature in the effective feature set and the load variable. The specific example is as follows: Assume that the following energy consumption-related data is collected in a commercial building to predict its electricity consumption. The collected features include: Temperature (°C); Passenger flow (people); Business activity index (dimensionless); humidity(%); Energy consumption (kWh); Dataset, consider the following data as input dataset: Timestamp Temperature (°C) Passenger flow (people) Business Activity Index humidity(%) Energy consumption (kWh) 2024-10-01 08:00 10 1000 0.6 60 200 2024-10-01 09:00 12 1200 0.7 55 220 2024-10-01 10:00 14 1500 0.8 50 240 2024-10-01 11:00 15 1600 0.85 45 260 2024-10-01 12:00 16 1700 0.9 40 280 2024-10-01 13:00 18 1800 0.95 35 300 Calculate the correlation coefficient. Suppose you want to calculate the correlation coefficient between these features and energy consumption, you can use the Pearson correlation coefficient to calculate it. Here are the detailed steps for how to calculate it: Data preparation, first of all, you need to make sure that all data is in digital format. The following is a pseudocode example of correlation coefficient calculation using Python: import pandas as pd # Create a data frame data = { 'Temperature': [10, 12, 14, 15, 16, 18], 'Passenger flow': [1000, 1200, 1500, 1600, 1700, 1800], 'Business Activity Index': [0.6, 0.7, 0.8, 0.85, 0.9, 0.95], 'Humidity': [60, 55, 50, 45, 40, 35], 'Energy consumption': [200, 220, 240, 260, 280, 300] } df = pd.DataFrame(data) # Calculate the correlation coefficient correlation_matrix = df.corr() print(correlation_matrix) Calculation results, assuming that the calculated correlation coefficient matrix is as follows: temperature Customer flow Business Activity Index humidity Energy consumption temperature 1.00 0.98 0.95 -0.97 0.97 Customer flow 0.98 1.00 0.96 -0.95 0.98 Business Activity Index 0.95 0.96 1.00 -0.94 0.96 humidity -0.97 -0.95 -0.94 1.00 -0.96 Energy consumption 0.97 0.98 0.96 -0.96 1.00 Analyzing the correlation coefficient, the correlation coefficient between temperature and energy consumption is 0.97, indicating that there is a very strong positive correlation between temperature and energy consumption. When the temperature rises, energy consumption also increases significantly. The correlation coefficient between passenger flow and energy consumption is 0.98, indicating that an increase in passenger flow is closely related to an increase in energy consumption; The correlation coefficient between humidity and energy consumption is -0.96, indicating that there is a strong negative correlation between humidity and energy consumption. When humidity decreases, energy consumption increases; Feature selection: Based on the calculated correlation coefficient, it is possible to decide which features should be retained in the model. Features with a high correlation with energy consumption, such as temperature, passenger flow, and commercial activity index, can be selected, while humidity may be less important in the model. To build a baseline model, use the features that are more closely related to energy consumption and build a linear regression model, you can use the following steps: Model training: Use temperature, passenger flow, and business activity index as input features and energy consumption as the target variable to train a linear regression model; from sklearn.linear_model import LinearRegression # Split input and output X = df[['Temperature', 'Passenger Flow', 'Business Activity Index']] y = df['Energy consumption'] # Create and train the model model = LinearRegression() model.fit(X, y) Model evaluation, calculate the mean square error (MSE) and coefficient of determination (R²) of the model to evaluate the performance of the model; Visualize the results, draw scatter plots to compare the actual energy consumption with the model predictions, generate scatter plots and load line charts to help visualize the accuracy of the model; In summary, it is demonstrated in detail how to calculate the correlation coefficient between each feature in the effective feature set and the energy consumption variable, and how to perform feature selection and establish a baseline model based on these results. This method can effectively improve the accuracy and reliability of the energy consumption prediction model.
[0023] A corresponding baseline model is established on the effective feature set, the mean square error and the coefficient of determination of the baseline model are calculated, and the load prediction result of the building terminal is constructed by the baseline model. The specific example is as follows: Suppose there is a data set of energy consumption of commercial buildings. The collected features include temperature, passenger flow, and business activity index. We want to use these features to predict the energy consumption of the building (in kilowatt-hours kWh). The data set is as follows: Timestamp Temperature (°C) Passenger flow (people) Business Activity Index Energy consumption (kWh) 2024-10-01 08:00 10 1000 0.6 200 2024-10-01 09:00 12 1200 0.7 220 2024-10-01 10:00 14 1500 0.8 240 2024-10-01 11:00 15 1600 0.85 260 2024-10-01 12:00 16 1700 0.9 280 2024-10-01 13:00 18 1800 0.95 300 To build a baseline model, a linear regression model will be used to predict energy consumption; First, import the necessary libraries and prepare the data: import pandas as pdfrom sklearn.linear_model importLinearRegressionfrom sklearn.metrics import mean_squared_error, r2_scoreimport matplotlib.pyplot as plt # Create a data frame data = { 'Timestamp': ['2024-10-01 08:00', '2024-10-01 09:00', '2024-10-01 10:00', '2024-10-01 11:00', '2024-10-01 12:00', '2024-10-01 13:00'], 'Temperature': [10, 12, 14, 15, 16, 18], 'Passenger flow': [1000, 1200, 1500, 1600, 1700, 1800], 'Business Activity Index': [0.6, 0.7, 0.8, 0.85, 0.9, 0.95], 'Energy consumption': [200, 220, 240, 260, 280, 300] } df = pd.DataFrame(data) df['timestamp'] = pd.to_datetime(df['timestamp']) # Convert to date and time format Then divide the features and target variables, separating the features (X) and target variables (y); X = df[['Temperature', 'Passenger Flow', 'Business Activity Index']] y = df['Energy consumption'] Then create and train the model, using a linear regression model for training: model = LinearRegression() model.fit(X, y) To predict energy consumption, you can use the trained model to predict energy consumption: predictions = model.predict(X) Calculate the mean square error and coefficient of determination. Use the prediction results to calculate the mean square error (MSE) and coefficient of determination (R²): mse = mean_squared_error(y, predictions) r2 = r2_score(y, predictions) print(f"Mean Square Error (MSE): {mse}") # Output mean square error print(f"Coefficient of Determination (R²): {r2}") # Output coefficient of determination Example results, assuming the calculation results are as follows: Mean square error (MSE): 25.0; Coefficient of determination (R²): 0.98; The mean square error (MSE) is 25.0, which represents the average square difference between the predicted value and the actual value. The smaller the value, the higher the prediction accuracy of the model; The coefficient of determination (R²) is 0.98, indicating that the model can explain 98% of the variation in energy consumption, indicating that the model fit is very good; Finally, the load forecast results are generated, the forecast results are compared with the actual energy consumption values, and a line chart of the load forecast is generated: plt.figure(figsize=(10, 5)) plt.plot(df['timestamp'], y, label='actual energy consumption', marker='o') plt.plot(df['timestamp'], predictions, label='predicted energy consumption', marker='x') plt.xlabel('time') plt.ylabel('Energy consumption (kWh)') plt.title('Energy consumption forecast') plt.legend() plt.xticks(rotation=45) plt.grid() plt.show() In summary, it is demonstrated in detail how to establish a baseline model on a valid feature set, calculate the mean square error and determination coefficient, and generate a visualization of the load forecast results. This process helps understand the model performance and provides a data basis for future energy consumption forecasting.
[0024] In this embodiment, before step S3 of collecting corresponding various multi-source data from different data sources through the preset sensor of the building terminal, the method further includes: S301: Based on the preset working range of the sensor, divide the collection area covered by the sensor in the building through the building terminal; S302: Determine whether a preset overlapping monitoring edge can be detected in the acquisition area; S303: If possible, construct an overlapping time window according to the overlapping monitoring edges, identify the collected measurement values of multiple sensors in the overlapping time window, generate data features to be fused based on the collected measurement values, apply a preset Kalman filter to assign weights to the collected measurement values, and calculate the weighted average of the data features, wherein the data features specifically include temperature, humidity and energy consumption.
[0025] In this embodiment, the system divides the collection areas covered by the sensors in the building through the building terminals based on the pre-set working range of the sensors in the building, and then the system determines whether these collection areas can detect the pre-set overlapping monitoring edges to execute the corresponding steps; for example, when the system determines that the collection areas covered by the sensors in the building do not detect the pre-set overlapping monitoring edges, the system will consider that the working range of the sensors is properly divided, and the collection areas covered by each sensor are relatively independent, avoiding repeated monitoring. The system will continue to collect energy consumption data of each area, transmit the data to the building terminal at a predetermined frequency, and monitor the data transmission. The stability of transmission ensures that the data of each sensor in the independent collection area can be transmitted to the terminal on time and completely. Even if the collection areas do not overlap, the system can also regularly check the status of the sensors, including battery life, signal strength, etc., to ensure long-term stable operation. Since the sensor coverage areas do not overlap, the system needs to regularly check whether there are any regional data omissions. If data collection in certain areas is found to be omitted or delayed, repair measures need to be taken; for example, when the system determines that the collection area covered by the sensor in the building detects the pre-set overlapping monitoring edge, the system will consider that there are multiple coverage detection areas in the working range of the sensor, and the system will According to the overlapping monitoring edge, an overlapping time window is constructed, and the collected measurement values of multiple sensors in the overlapping time window are identified. The data features to be fused are generated according to different collected measurement values. The data features specifically include temperature, humidity and energy consumption. The pre-set Kalman filter is used to assign weights to the collected measurement values and calculate the weighted average of the data features. The system can reduce the errors that may be caused by a single sensor by combining the measurement values from multiple sensors in the overlapping area. This redundant data can help confirm and correct outliers, thereby improving the accuracy of the overall data. At the same time, overlapping monitoring allows the system to use data features from multiple sources to improve the reliability of the data. Through Kalman filtering, the system can obtain more stable and reliable data features through weighted averaging in the presence of noise and uncertainty, and generate data features to be fused based on multiple measurement values in the overlapping time window, ensuring the diversity and comprehensiveness of data fusion. The measurement results of different sensors at the same time can provide richer information, making the final fused data features more representative. The Kalman filter automatically adjusts the weights of different sensors according to the changes in real-time measurements, so that the system can dynamically optimize data processing according to actual conditions. This adaptive ability is particularly important in a changing environment and can respond to changes in sensor status in real time.
[0026] It should be noted that an overlapping time window is constructed according to the overlapping monitoring edges, the collected measurement values of multiple sensors in the overlapping time window are identified, the data features to be fused are generated according to the collected measurement values, a preset Kalman filter is applied to weight the collected measurement values, and the weighted average value of the data features is calculated. The specific examples are as follows: Suppose there are two temperature sensors in an office building, sensor A and sensor B, and their overlapping time window is from 10:00 to 10:05. During this time period, the two sensors recorded temperature data respectively. The following is the specific data record: Sensor A's records: 10:00 - 22°C; 10:01 - 23°C; 10:02 - 24°C; 10:03 - 22°C; 10:04 - 23°C; 10:05 - 24°C; Sensor B records: 10:00 - 21°C; 10:01 - 22°C; 10:02 - 23°C; 10:03 - 22°C; 10:04 - 21°C; 10:05 - 23°C; Construct an overlapping time window and identify the overlapping time period between sensors A and B from 10:00 to 10:05. This time window will be used to integrate the data of these two sensors. Mark the collected measurement values. In the overlapping time window, the system will extract the temperature measurement values of the two sensors and obtain the following table: time Sensor A (°C) Sensor B (°C) 10:00 22 21 10:01 23 22 10:02 24 23 10:03 22 22 10:04 23 21 10:05 24 23 Generate data features to be fused, and extract temperature features to be fused from the table. For example, for the time period of 10:00, the features to be fused are [22, 21]; Apply Kalman filtering and set the filtering parameters. Assume that the measurement noise covariance of sensor A is 0.1 and the measurement noise covariance of sensor B is 0.3. The system will dynamically adjust the weights according to the different noises.
[0027] Weight calculation: Weight of sensor A
[0028] Weight of sensor B
[0029] Total weight =
[0030] Normalized weights ,
[0031] Taking 10:00 as an example, the weighted average of the data features is calculated:
[0032] Repeat the above steps for other time points to obtain: time Weighted mean temperature (°C) 10:00 21.75 10:01 (0.75×23)+(0.25×22)=22.75 10:02 (0.75×24)+(0.25×23)=23.75 10:03 (0.75×22)+(0.25×22)=22.00 10:04 (0.75×23)+(0.25×21)=22.75 10:05 (0.75×24)+(0.25×23)=23.75 In summary, the final weighted average temperature data will be more accurate and can better reflect the actual environmental conditions. This data fusion method based on overlapping monitoring edges and Kalman filtering provides more accurate and reliable environmental monitoring data for intelligent building management, ensuring the effectiveness and scientificity of decision-making.
[0033] In this embodiment, the step S3 of constructing a valid feature set according to the feature parameters further includes: S31: Based on preset virtual variables, construct seasonal characteristics of energy consumption of the building terminals, and calculate ratio characteristics and difference characteristics between the seasonal characteristics, wherein the virtual variables specifically include spring, summer, autumn and winter, the ratio characteristics specifically are the ratio of energy consumption to temperature, and the difference characteristics specifically are the difference between the current load and the past load; S32: Determine whether there are missing values in the ratio feature and the difference feature; S33: If not, standardize the ratio feature and the difference feature to ensure that the ratio feature and the difference feature are always on the same scale, combine the ratio feature and the difference feature with other features to form a composite feature, and capture the energy consumption impact corresponding to seasonal changes according to the composite feature.
[0034] In this embodiment, the system constructs the seasonal characteristics of energy consumption of building terminals based on pre-set virtual variables, which specifically include spring, summer, autumn and winter, and calculates the ratio characteristics and difference characteristics between seasonal characteristics. The ratio characteristics are specifically the ratio of energy consumption to temperature, and the difference characteristics are specifically the difference between current load and past load. Then the system determines whether there are missing values in these ratio characteristics and difference characteristics to execute corresponding steps; for example, when the system determines that there are missing values in the ratio characteristics and difference characteristics between seasonal characteristics, the system will believe that these characteristics may not accurately reflect the seasonal changes in energy consumption, which may affect subsequent analysis and prediction. The system will record the specific location (time point and characteristics) of the missing values, calculate the proportion of missing values to evaluate data integrity, and resample the data at the same time, select representative data within the time period for calculation, to ensure the timeliness and continuity of the data, and for missing values that cannot be reasonably filled, you can choose to mark them as "missing" or "abnormal", and eliminate them in subsequent analysis or process them separately. After processing the missing values, Recalculate ratio features and difference features to evaluate their impact on subsequent models; for example, when the system determines that there are no missing values in the ratio features and difference features between seasonal features, the system will believe that these features can accurately reflect the seasonal changes in energy consumption. The system will standardize these ratio features and difference features to ensure that the ratio features and difference features are always on the same scale, combine these ratio features and difference features to form composite features, and capture the energy consumption impact corresponding to seasonal changes based on the composite features; the system can ensure that all features are compared on the same scale by standardizing the ratio features and difference features. This standardization can prevent certain features from dominating other features due to their large numerical range, making the model more balanced and improving prediction accuracy. For example, the ratio of energy consumption to temperature may be relatively small, while the energy consumption difference may be large. After standardization, they can be calculated within the same range to avoid imbalance between features. At the same time, composite features can capture more complex energy consumption patterns by combining different ratio features and difference features. This combination of features can reveal a deeper connection between seasonal changes and energy consumption fluctuations in the model, so that the model does not rely solely on a single feature, but can comprehensively analyze the driving factors of energy consumption changes from multiple dimensions. This will help improve the prediction accuracy of the model. The combination of ratio features and difference features can accurately capture the changing trends of energy consumption in different seasons. For example, load changes in winter may have different energy consumption drivers than those in spring and summer. By analyzing these composite features, the system can identify specific energy consumption patterns in winter or summer, thereby optimizing load forecasts. Composite features can provide richer information for the prediction model, making the prediction results more interpretable. Decision makers can clarify the specific influencing factors of energy consumption in different seasons based on composite features, so as to better understand the source of energy consumption fluctuations in actual operations and take corresponding measures.
[0035] It should be noted that the ratio feature and the difference feature are standardized to ensure that the ratio feature and the difference feature are always on the same scale, and the ratio feature and the difference feature are combined with other features to form a composite feature. The energy consumption impact corresponding to seasonal changes is captured according to the composite feature. Specific examples are as follows: Assume that the system manages a large office building and collects the following data: Temperature (T): daily outside temperature (unit: °C); Energy consumption (E): energy consumption of the building (unit: kWh), recorded every hour; Historical energy consumption (E_history): energy consumption records for the past 24 hours (unit: kWh); Season (S): four dummy variables indicating the current season: spring (1), summer (2), autumn (3), and winter (4); Sample data: Assume that on a summer afternoon, the system collects the following data: Current temperature (T): 30°C; Current energy consumption (E): 300 kWh; Energy consumption in the past 24 hours (E_history): 250 kWh; Season (S): Summer (indicated by 2); First, calculate the ratio feature and the difference feature. Ratio feature (R): Calculate the ratio of energy consumption to temperature:
[0036] Difference feature (D): Calculate the difference between current energy consumption and past energy consumption:
[0037] Assuming that the energy consumption data is for the entire summer, the following are the statistics of the ratio feature and the difference feature: Mean of ratio feature (μr): 8 Standard deviation of ratio feature (σr): 2 Mean value of difference feature (μd): 30 Standard deviation of difference features (σd): 10 Then the standardization of the ratio feature is:
[0038] Standardization of difference features:
[0039] The standardized ratio feature, difference feature and seasonal feature are combined into a composite feature; Assuming that the seasonal feature S (summer) is represented as 2, the following composite feature can be formed:
[0040]
[0041] Capturing the impact of seasonal changes on energy consumption, the system now has composite features C1 and C2, which can be used to analyze the seasonal impact of energy consumption; Composite feature C1 = 2: indicates that in summer, energy consumption is relatively more sensitive to temperature; Composite feature C2 = 4: indicates that current energy consumption has increased significantly by 50 kWh compared to historical energy consumption, and this change is common in summer; In summary, through these composite features, the system can use machine learning models to predict energy consumption in the next few hours, taking into account the impact of seasonality and temperature changes. At the same time, if the system predicts that the temperature will continue to rise in the next few hours, the composite features show that energy consumption will increase significantly. Managers can initiate energy-saving measures in advance, such as adjusting the temperature setting of the air conditioner or starting backup energy. The system can also generate detailed reports on energy consumption and climate impacts to help building managers make more scientific decisions.
[0042] In this embodiment, the step S5 of establishing a corresponding baseline model on the effective feature set further includes: S51: Based on the preset chart of energy consumption of the building terminal, draw an X-axis from the chart as the energy consumption prediction value, and draw a Y-axis from the chart as the energy consumption residual, and combine them to obtain an energy consumption scatter plot; S52: Determine whether the residuals on the energy consumption scatter plot are randomly distributed; S53: If yes, the energy consumption scatter plot is input into the baseline model, and cross-validation is performed through the baseline model to obtain fluctuation information of the baseline model on different data subsets.
[0043] In this embodiment, the system draws the energy consumption forecast value on the X-axis and the energy consumption residual on the Y-axis based on the pre-set chart of energy consumption of the building terminal, and combines them to obtain an energy consumption scatter plot. Then the system determines whether the residuals on the energy consumption scatter plot are randomly distributed to execute the corresponding steps. For example, when the system determines that the residuals on the energy consumption scatter plot are not randomly distributed, the system will believe that the baseline model may have systematic deviations, and the model does not accurately capture the true relationship between energy consumption and related features. The system will check whether the features used in the model cover all key factors that may affect energy consumption, especially whether the external factors such as weather, economic activities, and the operating status of building equipment are fully reflected. At the same time, the model's ability to recognize energy consumption patterns is enhanced by introducing more virtual variables, adjusting the construction method of features, or considering interactive features, and generating an energy consumption scatter plot again to check whether the residual distribution is random to ensure that the model can better reflect actual energy consumption fluctuations. For example, when the system determines that The residuals on the energy consumption scatter plot belong to random distribution. At this time, the system will think that the baseline model can accurately capture the true relationship between energy consumption and related features. The system will input the energy consumption scatter plot into the baseline model, and cross-validate the baseline model to obtain the fluctuation information of the baseline model on different data subsets. The system can evaluate the stability and consistency of the model by cross-validating on different data subsets to ensure that the model can maintain good prediction performance in different scenarios. At the same time, by cross-validating on different data subsets, multiple groups of model performance indicators (such as mean square error, determination coefficient, etc.) can be obtained to fully understand the prediction ability of the model. In addition, the hyperparameters of the model can be adjusted during the cross-validation process to find the best model configuration and improve the prediction accuracy and robustness of the model. In the cross-validation process, the influence of each feature on the model prediction results can be analyzed to help identify which features still play an important role under different conditions, thereby providing a basis for subsequent feature selection and model improvement.
[0044] It should be noted that the energy consumption scatter plot is input into the baseline model, and cross-validation is performed through the baseline model to obtain fluctuation information of the baseline model on different data subsets. The specific examples are as follows: Suppose there is an energy consumption dataset of an office building, and the data includes the following features: Temperature (°C); humidity(%); level of economic activity (index); Historical energy consumption (kWh); Current energy consumption (kWh) (target variable); Data preparation, Collect data: Collect hourly energy consumption data for one year, assuming there are 8,760 records in total; Data cleaning: remove missing values and outliers to ensure data quality; Divide the data set: randomly divide the data set into 10 subsets (K=10), that is, each subset contains 876 records; Cross-validation process, training and testing cycle: Round 1: Use subset 1 as the test set and subsets 2 to 10 as the training set; Train the model, record the model parameters, predict the energy consumption using the input features of subset 1, and calculate the mean square error (MSE); Round 2: Use subset 2 as the test set, and subset 1 and subsets 3 to 10 as the training set; Continue this process until all subsets have been used as test sets once; Calculate the performance indicators, assuming that after 10 rounds of cross-validation, the recorded mean square error is as follows: Subset 1: MSE = 5.2; Subset 2: MSE = 4.8; Subset 3: MSE = 5.0; Subset 4: MSE = 6.1; Subset 5: MSE = 5.4; Subset 6: MSE = 4.9; Subset 7: MSE = 5.3; Subset 8: MSE = 5.1; Subset 9: MSE = 4.7; Subset 10: MSE = 5.5; Summary: Calculate the mean MSE and standard deviation: Average MSE = (5.2 + 4.8 + 5.0 + 6.1 + 5.4 + 4.9 + 5.3 + 5.1 + 4.7 + 5.5) / 10 = 5.2; Standard deviation = 0.26 (assuming calculated); Visualization and analysis, Draw a chart: Generate a chart with the horizontal axis showing actual energy consumption and the vertical axis showing predicted energy consumption. Add a trend line to intuitively display the model's prediction capability. Generate residual plot: Plot the residual plot to see if the residuals are randomly distributed and whether there is a pattern; The final model optimization analysis results show that the MSE of the model is higher in certain time periods (such as during the high temperature period in summer), which may be due to the complex relationship between temperature and energy consumption; Model adjustment: Consider adding more features (such as solar radiation intensity, wind speed, etc.); In summary, through this detailed cross-validation process, we can comprehensively evaluate the performance of the model on different data sets and ensure that the model has good generalization capabilities. If it is found that the model performs poorly in certain specific situations, targeted adjustments can be made based on the results of cross-validation to improve the accuracy of building energy consumption prediction. Such a detailed analysis and optimization process will help achieve more effective energy consumption management, improve energy efficiency, and reduce operating costs.
[0045] In this embodiment, the step S2 of determining whether the energy consumption data can be presented to the building terminal in real time further includes: S21: Based on the preset network bandwidth of the building terminal, detecting network delay information when the energy consumption data is transmitted; S22: Determine whether the network delay information exceeds a preset delay period; S23: If yes, then add a timestamp tag in the data collection phase and the data display phase of the energy consumption data, and calculate the overall delay of the energy consumption data according to the timestamp tag.
[0046] In this embodiment, the system detects the network delay information when the energy consumption data is transmitted based on the network bandwidth preset by the building terminal, and then the system determines whether the network delay information exceeds the preset delay period to execute the corresponding steps; for example, when the system determines that the network delay information when the energy consumption data is transmitted does not exceed the preset delay period, the system will consider that the network connection is normal, the data transmission is smooth, and the real-time energy consumption situation in the building can be reflected in time. The system will maintain the current data transmission status to ensure that the energy consumption data is continuously uploaded to the building terminal, record the current delay information, and update the system monitoring panel to ensure real-time monitoring of the data transmission status, and perform data integrity checks to ensure that no data is lost or damaged during the transmission process, generate real-time feedback reports, and show the current energy consumption data status and network transmission status to the management personnel to ensure that all relevant personnel understand the system operation status; for example, when the system determines that the network delay information when the energy consumption data is transmitted If the pre-set delay period is exceeded, the system will consider that the data transmission is abnormal and cannot reflect the real-time energy consumption in the building in a timely manner. The system will add timestamp tags in the data collection and data display links of the energy consumption data, and adaptively calculate the overall delay of the energy consumption data according to different timestamp tags. The system can accurately record the collection and transmission time of each data point through timestamp tags, helping the system to more accurately reflect the energy consumption situation and reduce misjudgments due to delays. At the same time, through the analysis of different timestamp tags, the system can identify the patterns and trends of delays, so as to understand the reasons for the increase in delays under specific conditions, which is convenient for subsequent optimization of network configuration. In addition, when network delays occur, the system can quickly locate the specific delay link (such as data collection or data transmission), which is convenient for technical personnel to perform targeted troubleshooting and repairs. The adaptive calculation of the overall delay can help the system adjust the data processing strategy. For example, when the network conditions are poor, reduce the frequency of data transmission or choose a more efficient data compression method.
[0047] In this embodiment, the step S4 of determining whether the valid feature set requires preset data cleaning further includes: S41: Calculate the proportion of feature missing values based on the pre-statistical features of the effective feature set, obtain the distribution of the feature missing values according to the proportion, and generate a corresponding missing value heat map; S42: Determine whether the missing value heat map matches the preset random missing; S43: If not, identifying the association information between the feature missing value and the pre-statistical feature, and collecting the corresponding missing source based on the association information.
[0048] In this embodiment, the system calculates the proportion of missing values of features based on the pre-statistical features of the effective feature set, obtains the distribution of missing values of features according to different proportions, generates a corresponding missing value heat map, and then the system determines whether the missing value heat map matches the pre-set random missingness to execute the corresponding steps; for example, when the system determines that the missing value heat map can match the pre-set random missingness, the system will believe that the missing value distribution of these features is not affected by systematic factors, and the missing values are randomly generated rather than caused by some specific reasons. The system will record the results of the missing value heat map and its matching status, and generate an analysis report so that the team can understand the status of the missing data and its impact on the analysis. At the same time, because the missing values are random, the system can continue to evaluate and verify the stability and accuracy of the model without making major adjustments to the data, and although the current missing values are random, the system still needs to monitor the missingness of subsequent data to ensure that the data quality continues to meet expectations. Standard, and by analyzing the pattern of missing values, the system can improve the data collection strategy to reduce the occurrence of missing values in the future, thereby improving the overall quality of the data; for example, when the system determines that the missing value heat map cannot match the pre-set random missing, the system will consider that the distribution of missing values of these features is affected, and the system will identify the association information between the missing values of the features and the pre-statistical features, and collect the corresponding missing sources based on different associated information; by analyzing the association between missing values and other features, the system can identify potential causes of missing values, such as sensor failure, data collection errors or external environmental factors, which helps to solve the problem in a targeted manner. At the same time, identifying non-random missing values can help model designers understand the limitations of the data set, consider these factors when building the model, and enhance the robustness of the model to missing data. Identifying the source of feature missingness can provide important information for building managers, helping them consider the potential impact of these data missingness on energy consumption analysis and prediction when making decisions.
[0049] It should be noted that based on the pre-statistical features of the effective feature set, the proportion of feature missing values is calculated, the distribution of the feature missing values is obtained according to the proportion, and the corresponding missing value heat map is generated. The specific examples are as follows: Suppose there is a data set for building energy consumption monitoring, including the following features: temperature; humidity; Energy consumption; the level of economic activity; A sample record of the dataset is as follows: Temperature(℃) humidity(%) Energy consumption (kWh) Economic activity level 22 60 300 0.8 23 NaN 320 0.9 NaN 65 290 NaN 24 70 NaN 0.7 25 NaN 310 0.85 NaN 68 330 0.9 Calculate the proportion of missing values. For each feature, calculate the proportion of missing values.
[0050] Temperature: number of missing values = 3, total number of records = 6, missing value ratio = 3 / 6 = 0.5 (50%); Humidity: number of missing values = 3, proportion of missing values = 3 / 6 = 0.5 (50%); Energy consumption: number of missing values = 2, missing value ratio = 2 / 6 = 0.33 (33.33%); Economic activity level: number of missing values = 1, proportion of missing values = 1 / 6 = 0.17 (16.67%); Get the missing value distribution and organize the calculation results into a table: feature Missing value ratio temperature 0.50 humidity 0.50 Energy consumption 0.33 Economic activity level 0.17 Generate a missing value heat map. The sample code for generating a heat map is as follows: import seaborn as snsimport matplotlib.pyplot as pltimport pandas aspd # Create sample data data = { 'Features': ['Temperature', 'Humidity', 'Energy Consumption', 'Economic Activity Level'], 'Missing value ratio': [0.50, 0.50, 0.33, 0.17] } df = pd.DataFrame(data) # Create heatmap plt.figure(figsize=(8, 4)) heatmap_data = df.set_index('feature') sns.heatmap(heatmap_data, annot=True, cmap='YlGnBu', cbar=True) plt.title('Missing Value Heat Map') plt.xlabel('Missing value ratio') plt.ylabel('Feature') plt.show() From the heat map, we can see that the missing value ratio of "temperature" and "humidity" has reached 50%, which indicates that the data quality of these two features is poor and may need further cleaning and filling. Based on the analysis of the heat map, the team can decide to use interpolation or other filling methods to deal with missing values, or re-examine the data collection process to reduce future missing values. In summary, by generating a heat map, we can intuitively display the missing data, which makes it easier to quickly identify data problems. At the same time, based on the heat map results, we can formulate more reasonable data cleaning and processing strategies to improve the integrity and accuracy of the data. Regularly generating heat maps will help monitor changes in missing data and ensure the continued effectiveness of data collection.
[0051] In this embodiment, based on the data unit type preset by the building terminal, the step S1 of unifying the data format of the energy consumption data into the data unit type further includes: S11: Detecting original units in the energy consumption data, and identifying different units to be converted based on the original units; S12: Determine whether the different units can be uniformly converted into the data unit type; S13: If not, mark the inconsistent items of the different units, perform unit conversion on the corresponding data of the different units row by row according to the inconsistent items, and compare the data distribution and statistical characteristics before and after the conversion.
[0052] In this embodiment, the system detects the original units in the energy consumption data, identifies the different units to be converted based on the original units, and then determines whether these different units can be uniformly converted into data unit types to execute the corresponding steps; for example, when the system determines that the different units to be converted can be uniformly converted into data unit types, the system will consider that these units have the same dimensions and can be converted through standard conversion formulas or ratios. The system will clarify the conversion relationship between each original unit and the data unit type, and at the same time, according to the defined conversion rules, convert the data to be converted item by item, and update the converted data to the data set, and record the original units and the converted units for traceability and auditing. Once the data is unified into a standard unit type, further data analysis, modeling and visualization can be performed; for example, when the system determines that the different units to be converted cannot be uniformly converted into data unit types, the system will consider that these units have the same dimensions and can be converted through standard conversion formulas or ratios. If the units do not have the same dimensions and cannot be converted uniformly, the system will mark the inconsistent items of different units, convert the corresponding data of different units line by line based on these inconsistent items, and compare the data distribution and statistical characteristics before and after the conversion; the system can avoid errors or deviations that may be introduced during global processing by converting units line by line, especially in complex situations, such as large differences in dimensions between units or unclear conversion rules. The system processes each piece of data separately according to its characteristics to ensure the accuracy of data conversion. At the same time, when encountering units that cannot be converted, the system will not simply discard the data, but will compare the data distribution and statistical characteristics before and after the conversion to ensure that meaningful information is retained. This method can maintain a high level of data integrity in a complex data environment and mark the inconsistent items of different units to ensure that these inconsistencies can be traced back in the subsequent analysis process to prevent confusion or misuse of data in the wrong units, so that problems can be quickly located and corrected when they occur.
[0053] Reference Figure 2 , is an energy-saving optimization control system based on a load curve in one embodiment of the present invention, comprising: A unification module 10, configured to unify the data format of the energy consumption data into the data unit type preset by the building terminal, wherein the data unit type specifically includes temperature, power and timestamp; A determination module 20, used to determine whether the energy consumption data can be presented to the building terminal in real time; The execution module 30 is used for, if not, collecting corresponding various multi-source data from different data sources through the preset sensors of the building terminal, identifying the time series of the multi-source data, applying the preset data fusion according to the time series, integrating the multi-source data to obtain fused data, extracting corresponding feature parameters from the fused data, and constructing a valid feature set according to the feature parameters, wherein the time series specifically includes trend changes and periodic changes, and the feature parameters specifically include weather impact index, economic activity level and historical load data; A second judgment module 40 is used to judge whether the valid feature set requires preset data cleaning; The second execution module 50 is used to use a preset correlation coefficient if it is not needed, collect the linear relationship between each feature in the effective feature set and the load variable, establish a corresponding baseline model on the effective feature set, calculate the mean square error and determination coefficient of the baseline model, construct the load prediction result of the building terminal through the baseline model, compare the load prediction result with the actual load value, and generate a load line graph.
[0054] In this embodiment, the unification module 10 unifies the data format of the energy consumption data collected in the building into the data unit type based on the data unit type pre-set by the building terminal, the data unit type specifically includes temperature, power and timestamp, and then the judgment module 20 judges whether these energy consumption data can be presented to the building terminal in real time to execute the corresponding steps; for example, when the system determines that the collected energy consumption data can be presented to the building terminal in real time, the system will consider that the data meets the requirements in terms of quality and can effectively reflect the actual energy consumption situation. The system will display the collected energy consumption data in real time on the building terminal, such as an energy consumption curve chart, a pie chart or a bar chart, so that users can quickly obtain information, and at the same time according to The alarm threshold is set based on historical data and demand. When a certain energy consumption exceeds the normal range, the system automatically issues an alarm so that timely measures can be taken, and the real-time data is recorded in the database for subsequent analysis and query to ensure the integrity and traceability of the data, conduct dynamic data analysis, generate trend charts or prediction models, and help managers identify energy consumption patterns and potential abnormal situations. For example, when the system determines that the collected energy consumption data cannot be presented to the building terminal in real time, the execution module 30 will consider that it cannot reflect the actual energy consumption situation. The system will collect the corresponding various multi-source data from different energy consumption data sources through the sensors pre-set at the building terminal, and identify the time of these multi-source data. Sequence, time series specifically includes trend changes and cyclical changes. According to the pre-set data fusion of time series application, multi-source data are integrated to obtain fused data, and corresponding feature parameters are extracted from the fused data. The feature parameters specifically include weather impact index, economic activity level and historical load data. An effective feature set is constructed based on these feature parameters; the system can capture the dynamic characteristics of energy consumption data by identifying and analyzing trend changes and cyclical changes in time series. This process not only helps to understand past energy consumption patterns, but also provides an important basis for future energy consumption forecasts. Through in-depth analysis of time series, managers can identify seasonal fluctuations, peaks of economic activities, and other factors that may affect energy consumption. and valleys, thereby optimizing energy consumption strategies in specific periods of time. At the same time, during the fusion process, the system integrates data from different sources into a unified data set through a preset data fusion algorithm. This process helps to eliminate noise and redundancy in the data and improve data quality. Through data fusion, managers can obtain more accurate information, thereby better supporting energy consumption analysis and decision-making. Based on the extracted feature parameters, an effective feature set is constructed to lay the foundation for subsequent prediction models. Through the effective feature set, the model can capture the key driving factors of building energy consumption and improve the accuracy of prediction. Then the second judgment module 40 determines whether the effective feature set requires pre-set data cleaning to execute the corresponding steps;For example, when the system determines that the effective feature set requires pre-set data cleaning, the system will believe that the data in the effective feature set may contain errors, outliers or noise, which will affect the accuracy of subsequent analysis and model building. The system will clarify the cleaning goals, such as removing duplicate data, filling missing values, processing outliers, etc., to ensure that all cleaning steps are consistent with the needs of subsequent analysis. At the same time, check whether there are missing values in the feature set, decide how to deal with these missing values, and ensure that the units, formats and types of all features are consistent. For example, the temperature unit may need to be unified as Celsius or Fahrenheit, and the power unit needs to ensure consistency. By identifying and removing duplicate data, the uniqueness of the feature set is ensured, and the interference of redundancy on the analysis is reduced; for example, when the system determines that the effective feature set does not require pre-set data cleaning, the second execution module 50 will believe that the data in the effective feature set does not contain errors, outliers or noise. The system will use a pre-set correlation coefficient to collect the linear relationship between each feature in the effective feature set and the load variable, establish a corresponding baseline model on the effective feature set, and calculate the mean square error and determination coefficient of the baseline model. , the load forecast results of the building terminal are constructed through the baseline model, the load forecast results are compared with the actual load value, and a load line chart is generated; the system can save the time and resources required for data cleaning by directly using the effective feature set, speed up the model establishment, enable the system to respond to changes quickly, and provide load forecast results in a timely manner. At the same time, through the preset correlation coefficient, the system can analyze the linear relationship between each feature in the effective feature set and the load variable. This analysis can help managers better understand the key factors affecting building energy consumption and guide subsequent management decisions. The baseline model established based on the linear relationship usually has good interpretability. Managers can clearly see the impact of each feature on the load forecast, so as to make more targeted adjustments. By establishing a baseline model, the system can calculate the mean square error and determination coefficient of the load forecast. These indicators provide a quantitative evaluation of the accuracy of the forecast results, which is helpful to optimize energy consumption management. By comparing the load forecast results with the actual load value, the gap between the forecast and the actual can be quickly identified. This feedback mechanism enables the system to adjust the energy consumption management strategy in a timely manner and improve resource utilization efficiency. ;
[0055] In this embodiment, it also includes: A division module, used to divide the collection area covered by the sensor in the building through the building terminal based on the preset working range of the sensor; A third judgment module is used to judge whether the acquisition area can detect a preset overlapping monitoring edge; The third execution module is used to, if possible, construct an overlapping time window according to the overlapping monitoring edges, identify the collected measurement values of multiple sensors in the overlapping time window, generate data features to be fused according to the collected measurement values, apply a preset Kalman filter to weight the collected measurement values, and calculate the weighted average of the data features, wherein the data features specifically include temperature, humidity and energy consumption.
[0056] In this embodiment, the system divides the collection areas covered by the sensors in the building through the building terminals based on the pre-set working range of the sensors in the building, and then the system determines whether these collection areas can detect the pre-set overlapping monitoring edges to execute the corresponding steps; for example, when the system determines that the collection areas covered by the sensors in the building do not detect the pre-set overlapping monitoring edges, the system will consider that the working range of the sensors is properly divided, and the collection areas covered by each sensor are relatively independent, avoiding repeated monitoring. The system will continue to collect energy consumption data of each area, transmit the data to the building terminal at a predetermined frequency, and monitor the data transmission. The stability of transmission ensures that the data of each sensor in the independent collection area can be transmitted to the terminal on time and completely. Even if the collection areas do not overlap, the system can also regularly check the status of the sensors, including battery life, signal strength, etc., to ensure long-term stable operation. Since the sensor coverage areas do not overlap, the system needs to regularly check whether there are any regional data omissions. If data collection in certain areas is found to be omitted or delayed, repair measures need to be taken; for example, when the system determines that the collection area covered by the sensor in the building detects the pre-set overlapping monitoring edge, the system will consider that there are multiple coverage detection areas in the working range of the sensor, and the system will According to the overlapping monitoring edge, an overlapping time window is constructed, and the collected measurement values of multiple sensors in the overlapping time window are identified. The data features to be fused are generated according to different collected measurement values. The data features specifically include temperature, humidity and energy consumption. The pre-set Kalman filter is used to assign weights to the collected measurement values and calculate the weighted average of the data features. The system can reduce the errors that may be caused by a single sensor by combining the measurement values from multiple sensors in the overlapping area. This redundant data can help confirm and correct outliers, thereby improving the accuracy of the overall data. At the same time, overlapping monitoring allows the system to use data features from multiple sources to improve the reliability of the data. Through Kalman filtering, the system can obtain more stable and reliable data features through weighted averaging in the presence of noise and uncertainty, and generate data features to be fused based on multiple measurement values in the overlapping time window, ensuring the diversity and comprehensiveness of data fusion. The measurement results of different sensors at the same time can provide richer information, making the final fused data features more representative. The Kalman filter automatically adjusts the weights of different sensors according to the changes in real-time measurements, so that the system can dynamically optimize data processing according to actual conditions. This adaptive ability is particularly important in a changing environment and can respond to changes in sensor status in real time.
[0057] In this embodiment, the execution module further includes: A calculation unit, configured to construct seasonal characteristics of energy consumption of the building terminal based on preset virtual variables, and calculate ratio characteristics and difference characteristics between the seasonal characteristics, wherein the virtual variables specifically include spring, summer, autumn and winter, the ratio characteristics specifically are the ratio of energy consumption to temperature, and the difference characteristics specifically are the difference between current load and past load; A judging unit, used for judging whether there are missing values in the ratio feature and the difference feature; an execution unit, for, if not, standardizing the ratio feature and the difference feature to ensure that the ratio feature and the difference feature are always on the same scale, combining the ratio feature and the difference feature with other features to form a composite feature, and capturing the energy consumption impact corresponding to seasonal changes according to the composite feature.
[0058] In this embodiment, the system constructs the seasonal characteristics of energy consumption of building terminals based on pre-set virtual variables, which specifically include spring, summer, autumn and winter, and calculates the ratio characteristics and difference characteristics between seasonal characteristics. The ratio characteristics are specifically the ratio of energy consumption to temperature, and the difference characteristics are specifically the difference between current load and past load. Then the system determines whether there are missing values in these ratio characteristics and difference characteristics to execute corresponding steps; for example, when the system determines that there are missing values in the ratio characteristics and difference characteristics between seasonal characteristics, the system will believe that these characteristics may not accurately reflect the seasonal changes in energy consumption, which may affect subsequent analysis and prediction. The system will record the specific location (time point and characteristics) of the missing values, calculate the proportion of missing values to evaluate data integrity, and resample the data at the same time, select representative data within the time period for calculation, to ensure the timeliness and continuity of the data, and for missing values that cannot be reasonably filled, you can choose to mark them as "missing" or "abnormal", and eliminate them in subsequent analysis or process them separately. After processing the missing values, Recalculate ratio features and difference features to evaluate their impact on subsequent models; for example, when the system determines that there are no missing values in the ratio features and difference features between seasonal features, the system will believe that these features can accurately reflect the seasonal changes in energy consumption. The system will standardize these ratio features and difference features to ensure that the ratio features and difference features are always on the same scale, combine these ratio features and difference features to form composite features, and capture the energy consumption impact corresponding to seasonal changes based on the composite features; the system can ensure that all features are compared on the same scale by standardizing the ratio features and difference features. This standardization can prevent certain features from dominating other features due to their large numerical range, making the model more balanced and improving prediction accuracy. For example, the ratio of energy consumption to temperature may be relatively small, while the energy consumption difference may be large. After standardization, they can be calculated within the same range to avoid imbalance between features. At the same time, composite features can capture more complex energy consumption patterns by combining different ratio features and difference features. This combination of features can reveal a deeper connection between seasonal changes and energy consumption fluctuations in the model, so that the model does not rely solely on a single feature, but can comprehensively analyze the driving factors of energy consumption changes from multiple dimensions. This will help improve the prediction accuracy of the model. The combination of ratio features and difference features can accurately capture the changing trends of energy consumption in different seasons. For example, load changes in winter may have different energy consumption drivers than those in spring and summer. By analyzing these composite features, the system can identify specific energy consumption patterns in winter or summer, thereby optimizing load forecasts. Composite features can provide richer information for the prediction model, making the prediction results more interpretable. Decision makers can clarify the specific influencing factors of energy consumption in different seasons based on composite features, so as to better understand the source of energy consumption fluctuations in actual operations and take corresponding measures.
[0059] In this embodiment, the second execution module further includes: A drawing unit, for drawing an energy consumption prediction value on the X-axis of the graph based on a preset graph of energy consumption of the building terminal, and drawing an energy consumption residual on the Y-axis of the graph, and combining them to obtain an energy consumption scatter plot; A second judgment unit is used to judge whether the residuals on the energy consumption scatter plot are randomly distributed; The second execution unit is configured to input the energy consumption scatter plot into the baseline model, perform cross-validation through the baseline model, and obtain fluctuation information of the baseline model on different data subsets.
[0060] In this embodiment, the system draws the energy consumption forecast value on the X-axis and the energy consumption residual on the Y-axis based on the pre-set chart of energy consumption of the building terminal, and combines them to obtain an energy consumption scatter plot. Then the system determines whether the residuals on the energy consumption scatter plot are randomly distributed to execute the corresponding steps. For example, when the system determines that the residuals on the energy consumption scatter plot are not randomly distributed, the system will believe that the baseline model may have systematic deviations, and the model does not accurately capture the true relationship between energy consumption and related features. The system will check whether the features used in the model cover all key factors that may affect energy consumption, especially whether the external factors such as weather, economic activities, and the operating status of building equipment are fully reflected. At the same time, the model's ability to recognize energy consumption patterns is enhanced by introducing more virtual variables, adjusting the construction method of features, or considering interactive features, and generating an energy consumption scatter plot again to check whether the residual distribution is random to ensure that the model can better reflect actual energy consumption fluctuations. For example, when the system determines that The residuals on the energy consumption scatter plot belong to random distribution. At this time, the system will think that the baseline model can accurately capture the true relationship between energy consumption and related features. The system will input the energy consumption scatter plot into the baseline model, and cross-validate the baseline model to obtain the fluctuation information of the baseline model on different data subsets. The system can evaluate the stability and consistency of the model by cross-validating on different data subsets to ensure that the model can maintain good prediction performance in different scenarios. At the same time, by cross-validating on different data subsets, multiple groups of model performance indicators (such as mean square error, determination coefficient, etc.) can be obtained to fully understand the prediction ability of the model. In addition, the hyperparameters of the model can be adjusted during the cross-validation process to find the best model configuration and improve the prediction accuracy and robustness of the model. In the cross-validation process, the influence of each feature on the model prediction results can be analyzed to help identify which features still play an important role under different conditions, thereby providing a basis for subsequent feature selection and model improvement.
[0061] In this embodiment, the judgment module further includes: A detection unit, configured to detect network delay information when the energy consumption data is transmitted based on a preset network bandwidth of the building terminal; A third judgment unit, used to judge whether the network delay information exceeds a preset delay period; The third execution unit is configured to, if yes, add a timestamp tag in the data collection link and the data display link of the energy consumption data, and calculate the overall delay of the energy consumption data according to the timestamp tag.
[0062] In this embodiment, the system detects the network delay information when the energy consumption data is transmitted based on the network bandwidth preset by the building terminal, and then the system determines whether the network delay information exceeds the preset delay period to execute the corresponding steps; for example, when the system determines that the network delay information when the energy consumption data is transmitted does not exceed the preset delay period, the system will consider that the network connection is normal, the data transmission is smooth, and the real-time energy consumption situation in the building can be reflected in time. The system will maintain the current data transmission status to ensure that the energy consumption data is continuously uploaded to the building terminal, record the current delay information, and update the system monitoring panel to ensure real-time monitoring of the data transmission status, and perform data integrity checks to ensure that no data is lost or damaged during the transmission process, generate real-time feedback reports, and show the current energy consumption data status and network transmission status to the management personnel to ensure that all relevant personnel understand the system operation status; for example, when the system determines that the network delay information when the energy consumption data is transmitted If the pre-set delay period is exceeded, the system will consider that the data transmission is abnormal and cannot reflect the real-time energy consumption in the building in a timely manner. The system will add timestamp tags in the data collection and data display links of the energy consumption data, and adaptively calculate the overall delay of the energy consumption data according to different timestamp tags. The system can accurately record the collection and transmission time of each data point through timestamp tags, helping the system to more accurately reflect the energy consumption situation and reduce misjudgments due to delays. At the same time, through the analysis of different timestamp tags, the system can identify the patterns and trends of delays, so as to understand the reasons for the increase in delays under specific conditions, which is convenient for subsequent optimization of network configuration. In addition, when network delays occur, the system can quickly locate the specific delay link (such as data collection or data transmission), which is convenient for technical personnel to perform targeted troubleshooting and repairs. The adaptive calculation of the overall delay can help the system adjust the data processing strategy. For example, when the network conditions are poor, reduce the frequency of data transmission or choose a more efficient data compression method.
[0063] In this embodiment, the second determination module further includes: A generating unit, configured to calculate a proportion of feature missing values based on pre-statistical features of the effective feature set, obtain a distribution of the feature missing values according to the proportion, and generate a corresponding missing value heat map; A fourth judgment unit, used to judge whether the missing value heat map matches a preset random missing; The fourth execution unit is used to identify the association information between the feature missing value and the pre-statistical feature if no, and collect the corresponding missing source according to the association information.
[0064] In this embodiment, the system calculates the proportion of missing values of features based on the pre-statistical features of the effective feature set, obtains the distribution of missing values of features according to different proportions, generates a corresponding missing value heat map, and then the system determines whether the missing value heat map matches the pre-set random missingness to execute the corresponding steps; for example, when the system determines that the missing value heat map can match the pre-set random missingness, the system will believe that the missing value distribution of these features is not affected by systematic factors, and the missing values are randomly generated rather than caused by some specific reasons. The system will record the results of the missing value heat map and its matching status, and generate an analysis report so that the team can understand the status of the missing data and its impact on the analysis. At the same time, because the missing values are random, the system can continue to evaluate and verify the stability and accuracy of the model without making major adjustments to the data, and although the current missing values are random, the system still needs to monitor the missingness of subsequent data to ensure that the data quality continues to meet expectations. Standard, and by analyzing the pattern of missing values, the system can improve the data collection strategy to reduce the occurrence of missing values in the future, thereby improving the overall quality of the data; for example, when the system determines that the missing value heat map cannot match the pre-set random missing, the system will consider that the distribution of missing values of these features is affected, and the system will identify the association information between the missing values of the features and the pre-statistical features, and collect the corresponding missing sources based on different associated information; by analyzing the association between missing values and other features, the system can identify potential causes of missing values, such as sensor failure, data collection errors or external environmental factors, which helps to solve the problem in a targeted manner. At the same time, identifying non-random missing values can help model designers understand the limitations of the data set, consider these factors when building the model, and enhance the robustness of the model to missing data. Identifying the source of feature missingness can provide important information for building managers, helping them consider the potential impact of these data missingness on energy consumption analysis and prediction when making decisions.
[0065] In this embodiment, the unified module also includes: an identification unit, configured to detect original units in the energy consumption data, and identify different units to be converted based on the original units; A fifth judgment unit, used to judge whether the different units can be uniformly converted into the data unit type; The fifth execution unit is used to mark the inconsistent items of the different units if not, perform unit conversion on the corresponding data of the different units row by row according to the inconsistent items, and compare the data distribution and statistical characteristics before and after the conversion.
[0066] In this embodiment, the system detects the original units in the energy consumption data, identifies the different units to be converted based on the original units, and then determines whether these different units can be uniformly converted into data unit types to execute the corresponding steps; for example, when the system determines that the different units to be converted can be uniformly converted into data unit types, the system will consider that these units have the same dimensions and can be converted through standard conversion formulas or ratios. The system will clarify the conversion relationship between each original unit and the data unit type, and at the same time, according to the defined conversion rules, convert the data to be converted item by item, and update the converted data to the data set, and record the original units and the converted units for traceability and auditing. Once the data is unified into a standard unit type, further data analysis, modeling and visualization can be performed; for example, when the system determines that the different units to be converted cannot be uniformly converted into data unit types, the system will consider that these units have the same dimensions and can be converted through standard conversion formulas or ratios. If the units do not have the same dimensions and cannot be converted uniformly, the system will mark the inconsistent items of different units, convert the corresponding data of different units line by line based on these inconsistent items, and compare the data distribution and statistical characteristics before and after the conversion; the system can avoid errors or deviations that may be introduced during global processing by converting units line by line, especially in complex situations, such as large differences in dimensions between units or unclear conversion rules. The system processes each piece of data separately according to its characteristics to ensure the accuracy of data conversion. At the same time, when encountering units that cannot be converted, the system will not simply discard the data, but will compare the data distribution and statistical characteristics before and after the conversion to ensure that meaningful information is retained. This method can maintain a high level of data integrity in a complex data environment and mark the inconsistent items of different units to ensure that these inconsistencies can be traced back in the subsequent analysis process to prevent confusion or misuse of data in the wrong units, so that problems can be quickly located and corrected when they occur.
[0067] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. An energy-saving optimization control method based on load curve, characterized in that: The following steps are involved: Based on the data unit type preset by the building terminal, the data format of the energy consumption data is unified into the data unit type, wherein the data unit type specifically includes temperature, power and timestamp; Determining whether the energy consumption data can be presented to the building terminal in real time; If not, then collect corresponding various multi-source data from different data sources through the preset sensors of the building terminal, identify the time series of the multi-source data, apply the preset data fusion according to the time series, integrate the multi-source data to obtain fused data, extract corresponding feature parameters from the fused data, and construct a valid feature set based on the feature parameters, wherein the time series specifically includes trend changes and periodic changes, and the feature parameters specifically include weather impact index, economic activity level and historical load data; Determining whether the valid feature set requires preset data cleaning; If not necessary, a preset correlation coefficient is used to collect the linear relationship between each feature in the effective feature set and the load variable, a corresponding baseline model is established on the effective feature set, the mean square error and determination coefficient of the baseline model are calculated, and the load prediction result of the building terminal is constructed through the baseline model. The load prediction result is compared with the actual load value to generate a load line chart.
2. The energy-saving optimization control method based on load curve according to claim 1 is characterized in that: Before the step of collecting corresponding various multi-source data from different data sources through the sensors preset by the building terminal, the method further includes: Based on the preset working range of the sensor, the collection area covered by the sensor in the building is divided by the building terminal; Determining whether the acquisition area can detect a preset overlapping monitoring edge; If possible, an overlapping time window is constructed based on the overlapping monitoring edges, and the collected measurement values of multiple sensors in the overlapping time window are identified. Data features to be fused are generated based on the collected measurement values, and a preset Kalman filter is applied to assign weights to the collected measurement values, and the weighted average of the data features is calculated, wherein the data features specifically include temperature, humidity and energy consumption.
3. The energy-saving optimization control method based on load curve according to claim 1 is characterized in that: The step of constructing a valid feature set according to the feature parameters further includes: Based on the preset virtual variables, the seasonal characteristics of the energy consumption of the building terminals are constructed, and the ratio characteristics and the difference characteristics between the seasonal characteristics are calculated, wherein the virtual variables specifically include spring, summer, autumn and winter, the ratio characteristics are specifically the ratio of energy consumption to temperature, and the difference characteristics are specifically the difference between the current load and the past load; Determine whether there are missing values in the ratio feature and the difference feature; If not, the ratio feature and the difference feature are standardized to ensure that they are always on the same scale, and the ratio feature and the difference feature are combined with other features to form a composite feature, and the energy consumption impact corresponding to seasonal changes is captured according to the composite feature.
4. The energy-saving optimization control method based on load curve according to claim 1 is characterized in that: The step of establishing a corresponding baseline model on the effective feature set further includes: Based on the preset chart of energy consumption of the building terminal, the X-axis is drawn from the chart as the energy consumption prediction value, and the Y-axis is drawn from the chart as the energy consumption residual, and the energy consumption scatter plot is obtained by combining them; Determining whether the residuals on the energy consumption scatter plot are randomly distributed; If yes, the energy consumption scatter plot is input into the baseline model, and cross-validation is performed through the baseline model to obtain fluctuation information of the baseline model on different data subsets.
5. The energy-saving optimization control method based on load curve according to claim 1 is characterized in that: The step of determining whether the energy consumption data can be presented to the building terminal in real time also includes: Based on the preset network bandwidth of the building terminal, detecting the network delay information when the energy consumption data is transmitted; Determining whether the network delay information exceeds a preset delay period; If so, a timestamp tag is added in the data collection link and the data display link of the energy consumption data, and the overall delay of the energy consumption data is calculated according to the timestamp tag.
6. The energy-saving optimization control method based on load curve according to claim 1 is characterized in that: The step of determining whether the valid feature set requires preset data cleaning also includes: Based on the pre-statistical features of the effective feature set, the proportion of feature missing values is calculated, the distribution of the feature missing values is obtained according to the proportion, and a corresponding missing value heat map is generated; Determine whether the missing value heat map matches a preset random missing; If not, then identify the association information between the feature missing value and the pre-statistical feature, and collect the corresponding missing source based on the association information.
7. The energy-saving optimization control method based on load curve according to claim 1 is characterized in that: The step of unifying the data format of the energy consumption data into the data unit type based on the data unit type preset by the building terminal further includes: detecting original units in the energy consumption data, and identifying different units to be converted based on the original units; Determining whether the different units can be uniformly converted into the data unit type; If not, the inconsistent items of the different units are marked, and the corresponding data of the different units are converted row by row according to the inconsistent items, and the data distribution and statistical characteristics before and after the conversion are compared.
8. An energy-saving optimization control system based on load curve, characterized in that: include: A unification module, used for unifying the data format of the energy consumption data into the data unit type preset by the building terminal based on the data unit type, wherein the data unit type specifically includes temperature, power and timestamp; A determination module, used to determine whether the energy consumption data can be presented to the building terminal in real time; an execution module, for, if not, collecting corresponding various multi-source data from different data sources through the preset sensors of the building terminal, identifying the time series of the multi-source data, applying preset data fusion according to the time series, integrating the multi-source data to obtain fused data, extracting corresponding feature parameters from the fused data, and constructing a valid feature set according to the feature parameters, wherein the time series specifically includes trend changes and periodic changes, and the feature parameters specifically include weather impact index, economic activity level and historical load data; A second judgment module is used to judge whether the valid feature set requires preset data cleaning; The second execution module is used to use a preset correlation coefficient if it is not needed, collect the linear relationship between each feature in the effective feature set and the load variable, establish a corresponding baseline model on the effective feature set, calculate the mean square error and determination coefficient of the baseline model, construct the load prediction result of the building terminal through the baseline model, compare the load prediction result with the actual load value, and generate a load line graph.
9. The energy-saving optimization control system based on load curve according to claim 8, characterized in that: Also includes: A division module, used to divide the collection area covered by the sensor in the building through the building terminal based on the preset working range of the sensor; A third judgment module is used to judge whether the acquisition area can detect a preset overlapping monitoring edge; The third execution module is used to, if possible, construct an overlapping time window according to the overlapping monitoring edges, identify the collected measurement values of multiple sensors in the overlapping time window, generate data features to be fused according to the collected measurement values, apply a preset Kalman filter to weight the collected measurement values, and calculate the weighted average of the data features, wherein the data features specifically include temperature, humidity and energy consumption.
10. The energy-saving optimization control system based on load curve according to claim 8, characterized in that: The execution module also includes: A calculation unit, configured to construct seasonal characteristics of energy consumption of the building terminal based on preset virtual variables, and calculate ratio characteristics and difference characteristics between the seasonal characteristics, wherein the virtual variables specifically include spring, summer, autumn and winter, the ratio characteristics specifically are the ratio of energy consumption to temperature, and the difference characteristics specifically are the difference between current load and past load; A judging unit, used for judging whether there are missing values in the ratio feature and the difference feature; an execution unit, for, if not, standardizing the ratio feature and the difference feature to ensure that the ratio feature and the difference feature are always on the same scale, combining the ratio feature and the difference feature with other features to form a composite feature, and capturing the energy consumption impact corresponding to seasonal changes according to the composite feature.
Citation Information
Patent Citations
Prediction method of forest canopy biomass
CN110109118A
Energy-saving control management method, system and equipment for smart identifier
CN117689118A
Intelligent power utilization monitoring system
CN118868404A
Cited By
Energy consumption monitoring intelligent management platform
CN120146635A
Distributed sewage treatment intelligent control system
CN120315353A
