Building energy consumption multi-scene prediction method and device based on data-mechanism fusion driving

Abstract

The invention discloses a building energy consumption multi-scene prediction method and device based on data-mechanism fusion driving, and relates to the technical field of building energy consumption multi-scene prediction, and the method comprises the steps: collecting building energy consumption data from a building management platform BMS, a meteorological station and historical energy consumption records through employing a data collection method, and obtaining the building energy consumption data; preprocessing the building energy consumption data by adopting a data preprocessing method to obtain a preprocessed building energy consumption data set; constructing a basic physical model for describing the heat change in the building based on the thermodynamic principle, and constructing a data driving model based on the preprocessed data and a machine learning algorithm; the basic physical model and the data driving model are combined to form a comprehensive data-mechanism fusion model, an energy consumption prediction fusion model is obtained, the preprocessed data are input into the energy consumption prediction fusion model, and an energy consumption prediction value is output; a scene analysis and simulation method is adopted, and a Monte Carlo simulation statistical method is utilized to generate a plurality of possible scenes.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building energy consumption multi-scene prediction, in particular to a building energy consumption multi-scene prediction method and device based on data-mechanism fusion driving. BACKGROUND

[0002] File encryption technology plays a crucial role in the field of information security today. With the rapid development of information technology, users' demand for file security and privacy protection is growing. Traditional encryption methods often rely on fixed keys and algorithms, which have problems such as complex key management, limited attack resistance, and cannot meet users' demand for personalized and intelligent encryption. Therefore, how to use advanced technology to improve the level of file encryption intelligence and security has become one of the problems to be solved.

[0003] In the field of building energy consumption multi-scene prediction, traditional encryption methods mainly rely on symmetric encryption algorithms and public key encryption algorithms such as AES, DES, RSA, etc. These algorithms ensure the confidentiality of data to some extent, but are difficult to cope with the dynamic changes and diversity of user behavior. At the same time, due to the fixedness of encryption algorithms and key management, there are security risks such as key leakage and insufficient encryption strength. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a building energy consumption multi-scene prediction method based on data-mechanism fusion driving to solve the above technical problems.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] The present application provides a building energy consumption multi-scene prediction method based on data-mechanism fusion driving, which includes:

[0008] Collecting building energy consumption data from building management platform BMS, weather station and historical energy consumption records using data collection method to obtain building energy consumption data;

[0009] Preprocessing the building energy consumption data using data preprocessing method to obtain preprocessed building energy consumption data set;

[0010] Based on the thermodynamic principle, a basic physical model describing the change of heat inside the building is constructed, and a data-driven model is constructed based on the preprocessed data and machine learning algorithm;

[0011] The basic physical model is combined with the data-driven model to form a comprehensive data-mechanism fusion model, an energy consumption prediction fusion model is obtained, and the preprocessed data is input into the energy consumption prediction fusion model to output an energy consumption prediction value.

[0012] A scenario analysis and simulation method is used, a plurality of possible scenarios are generated by using a Monte Carlo simulation statistical method, the possibility of building energy consumption under different conditions is evaluated, scenario simulation is performed based on the energy consumption prediction value, and a plurality of energy consumption prediction results are output.

[0013] Based on the plurality of energy consumption prediction results, an energy efficiency optimization suggestion method is used to provide specific energy saving suggestions for building managers, and an energy consumption suggestion strategy is obtained.

[0014] As a preferred scheme of the building energy consumption multi-scenario prediction method based on data-mechanism fusion driving, the data collection method is used to collect building energy consumption data from the building management platform BMS, the weather station and the historical energy consumption record, and the specific steps are as follows:

[0015] The API interface is used to obtain indoor temperature, humidity and equipment running state data from the building management platform BMS;

[0016] The open API interface provided by the weather station is used to collect weather parameters such as outdoor temperature, wind speed and solar radiation intensity;

[0017] The historical energy consumption record query method is used to extract energy consumption values in a past time period, and the data is integrated to form a unified data set, and building energy consumption data is obtained.

[0018] As a preferred scheme of the building energy consumption multi-scenario prediction method based on data-mechanism fusion driving, the data preprocessing method is used to preprocess the building energy consumption data to obtain a preprocessed building energy consumption data set, and the specific steps are as follows:

[0019] The data cleaning method is used to remove outliers in the original data;

[0020] By calculating the difference between each data point and its adjacent point, data points significantly deviating from the normal range are identified and removed, and a preliminary cleaned data set is obtained;

[0021] The missing value filling method is used to fill the missing values in the preliminary cleaned data set, the linear interpolation method is used to fill the missing data points according to the time series characteristics of the data, and a data set without missing values is obtained;

[0022] The feature engineering method is used to extract and convert the data set without missing values;

[0023] The feature extraction and conversion includes standardization, normalization processing and creating new feature variables based on physical meaning and statistical principles.

[0024] As a preferred scheme of the building energy consumption multi-scenario prediction method based on data-mechanism fusion driving of the application, the basic physical model describing the heat change in the building is constructed based on the principle of thermodynamics, and a data-driven model is constructed based on the preprocessed data and machine learning algorithm, and the specific steps are:

[0025] The heat change in the building is modeled by using the energy balance equation method;

[0026] According to the building material properties, structural design factors and environmental parameters, the heat exchange per unit time is calculated to obtain the basic physical model E mech (t).

[0027]

[0028] Where, k is the thermal conductivity of building materials, A is the wall area, d is the wall thickness, η is the solar radiation absorption efficiency, T in (t) is the indoor temperature at time t, T out (t) is the outdoor temperature at time t, I solar (t) is the solar radiation intensity at time t.

[0029] The preprocessed data is analyzed by using the machine learning algorithm training method;

[0030] The long short-term memory network LSTM is selected to construct the data-driven model E data (t).

[0031] f LSTM (X(t),h t-1 ,c t-1 )=σ(W x ·X(t)+W h ·h t-1 +b);

[0032] E data (t)=f LSTM (X(t),h t-1 ,c t-1 );

[0033] Where, f LSTM is the core calculation function of the long short-term memory network, W x is the input data weight matrix, W his the hidden state weight matrix, b is the bias vector, σ is the Sigmoid function, X(t) is the input feature vector at time t, including indoor temperature, humidity, equipment running state, outdoor temperature, wind speed and solar radiation intensity, h t-1 is the hidden state vector at time t-1, c t-1 is the cell state vector at time t-1.

[0034] As a preferred scheme of the building energy consumption multi-scene prediction method based on data-mechanism fusion driving, the basic physical model is combined with the data-driven model to form a comprehensive data-mechanism fusion model, an energy consumption prediction fusion model is obtained, and the preprocessed data is input into the energy consumption prediction fusion model to output the energy consumption prediction value, and the specific steps are as follows:

[0035] The weight coefficient α is used to adjust the method to balance the relative importance between the basic physical model E mech (t) and the data-driven model E data (t), and a comprehensive data-mechanism fusion model is constructed, and the expression is:

[0036]

[0037] Wherein, E mech (t) is the energy consumption value predicted by the basic physical model at time t, E data (t) is the energy consumption value predicted by the data-driven model at time t, and α is the weight coefficient for adjusting the relative contribution of the two;

[0038] The method of optimizing the weight coefficient α based on historical data is adopted, and the optimal α * is determined by minimizing the prediction error, and the expression is:

[0039]

[0040] Wherein, E actual,i is the actual energy consumption value at the i th time point, is the energy consumption value predicted by the data-mechanism fusion model at the i th time point, and n is the total number of time points;

[0041] The preprocessed data set is input into the energy consumption prediction fusion model;

[0042] The input feature vector X(t) of each time point t is predicted to obtain the energy consumption prediction value

[0043] As a preferred scheme of the building energy consumption multi-scenario prediction method based on data-mechanism fusion driving, the scene analysis and simulation method is used to generate a plurality of possible scenarios by using a Monte Carlo simulation statistical method, to evaluate the possibility of building energy consumption under different conditions, and to perform scene simulation based on the energy consumption prediction value to output a plurality of energy consumption prediction results, and the specific steps are as follows:

[0044] The Monte Carlo simulation method is used to randomly sample the key parameters affecting the building energy consumption to obtain a random sample set.

[0045] The random sample is input into the energy consumption prediction fusion model, the energy consumption of each random sample is predicted, the corresponding energy consumption prediction value under the sample condition is calculated, and the expression is as follows:

[0046]

[0047] Wherein, E mech (t) (i) is the predicted energy consumption value of the basic physical model under the i th sample condition, E data (t) (i) is the predicted energy consumption value of the data-driven model under the i th sample condition, and a is a weight coefficient.

[0048] The energy consumption prediction results of all random samples are summarized and analyzed by using the statistical analysis method, the average energy consumption prediction value and its standard deviation under different conditions are calculated to evaluate the possibility of building energy consumption under different external conditions, and the expression is as follows:

[0049]

[0050] Wherein, N is the number of random samples, is the average energy consumption prediction value, and σ E (t) is the average energy consumption standard deviation.

[0051] The plurality of energy consumption prediction results are output, and the average energy consumption prediction value and the standard deviation calculated under different conditions are taken as the final scene analysis result.

[0052] As a preferred scheme of the building energy consumption multi-scenario prediction method based on data-mechanism fusion driving, the scene analysis and simulation method is used to generate a plurality of possible scenarios by using a Monte Carlo simulation statistical method, to evaluate the possibility of building energy consumption under different conditions, and to perform scene simulation based on the energy consumption prediction value to output a plurality of energy consumption prediction results, and the specific steps are as follows:

[0053] The plurality of energy consumption prediction results are output, and the average energy consumption prediction value and the standard deviation calculated under different conditions are taken as the final scene analysis result.

[0054] An energy efficiency optimization model is established, based on the identified high energy consumption period, combined with the actual operation of the building, an optimization model is constructed to minimize the total energy consumption, the expression is:

[0055]

[0056] Wherein, n is the total number of time points, λ is the risk coefficient, for balancing the relationship between energy consumption expectation value and volatility;

[0057] The linear programming algorithm is used to find the optimal control parameter combination;

[0058] The control parameter combination includes indoor temperature setting value, equipment start-stop time and lighting system brightness adjustment, so that the total energy consumption is minimized under the premise of meeting the comfort requirement;

[0059] According to the optimal control parameter combination obtained by solving, specific energy-saving measure suggestions are made;

[0060] The energy-saving measure suggestions include adjusting the running temperature range of the air conditioner, optimizing the on-off time table of the lighting system and reasonably arranging the equipment use time.

[0061] The present application provides a kind of based on data-mechanism fusion driven building energy consumption multi-scene prediction device, comprising:

[0062] Data collection module, data preprocessing module, model construction module, energy consumption prediction module, scenario simulation module and energy efficiency optimization module;

[0063] The data collection module is used to collect building energy consumption related data from building management platform BMS, weather station and historical energy consumption records;

[0064] The data preprocessing module is used to clean and pretreat the collected building energy consumption data to obtain a high-quality data set;

[0065] The model construction module is used to construct a basic physical model and a data-driven model, and combine the two to form a comprehensive data-mechanism fusion model;

[0066] The energy consumption prediction module is used to input the pretreated data into the energy consumption prediction fusion model and output the energy consumption prediction value at each time point;

[0067] The scenario simulation module is used to generate multiple possible scenarios using the Monte Carlo simulation method to evaluate the possibility of building energy consumption under different conditions;

[0068] The energy efficiency optimization module is used to provide specific energy-saving suggestions for building managers based on multiple energy consumption prediction results.

[0069] The present application has the beneficial effects that: by integrating multi-source data, a more accurate and comprehensive data basis is provided, laying a solid foundation for subsequent data preprocessing, model construction and energy consumption prediction, not only improving the prediction accuracy, but also enhancing the adaptability and robustness of the model under different conditions, the high-quality data set helps to improve the training efficiency and prediction accuracy of the model, reduces the error caused by data quality problems, the standardization and normalization processing in the feature engineering process enables data of different scales to be compared and analyzed under the same standard, further optimizing the model performance, the combination mode of data-mechanism fusion model not only retains the stability of the physical model, but also has the flexibility of data-driven model, significantly improving the accuracy and reliability of energy consumption prediction, by using the Monte Carlo simulation method to generate multiple possible scenarios, the possibility of building energy consumption under different conditions is evaluated, realizing multi-dimensional prediction of future energy consumption trend, the method not only considers the energy consumption under the current known conditions, but also predicts various scenarios that may occur in the future, helping managers better understand and cope with uncertainty. BRIEF DESCRIPTION OF DRAWINGS

[0070] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0071] Figure 1 The flowchart of the building energy consumption multi-scenario prediction method based on data-mechanism fusion driving of the present application.

[0072] Figure 2 The schematic diagram of the building energy consumption multi-scenario prediction device based on data-mechanism fusion driving of the present application. DETAILED DESCRIPTION

[0073] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail in conjunction with the drawings of the specification.

[0074] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited by the specific embodiments disclosed below.

[0075] Second, the "one embodiment" or "an embodiment" referred to herein means a particular feature, structure, or characteristic including an implementation that can be included in at least one implementation of the application. The appearances of "in one embodiment" or "in an embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily all referring to a single, alternative embodiment, or a single, alternative implementation.

[0076] Referring to Figure 1 and Figure 2 , a building energy consumption multi-scenario prediction method based on data-mechanism fusion driving is provided, comprising the following steps:

[0077] S1, using a data collection method to collect building energy consumption data from a building management platform BMS, a meteorological station and historical energy consumption records, to obtain building energy consumption data;

[0078] Further, API interface is used to obtain indoor temperature, humidity and equipment running state data from the building management platform BMS;

[0079] The weather parameters of outdoor temperature, wind speed and solar radiation intensity are collected by using the open API interface provided by the meteorological station;

[0080] The energy consumption values in the past time period are extracted by using the historical energy consumption record query method, and are integrated with the above data to form a unified data set, to obtain building energy consumption data;

[0081] It should be noted that through the integration of multiple data sources, not only the changes of the internal environment of the building and the external weather conditions can be comprehensively reflected, but also more accurate prediction basis can be provided by combining historical energy consumption records. The method ensures the real-time and accuracy of the data, and provides solid data support for the subsequent steps.

[0082] S2, using a data preprocessing method to preprocess the building energy consumption data, to obtain a preprocessed building energy consumption data set;

[0083] Further, a data cleaning method is used to remove outliers in the original data;

[0084] By calculating the difference between each data point and its adjacent point, the data points significantly deviating from the normal range are identified and removed, to obtain a preliminary cleaned data set;

[0085] A missing value filling method is used to fill the missing values in the preliminary cleaned data set. According to the time series characteristics of the data, the linear interpolation method is used to fill the missing data points, to obtain a data set without missing values;

[0086] A feature engineering method is used to extract and convert the data set without missing values;

[0087] Feature extraction and transformation include standardization, normalization processing and creation of new feature variables based on physical meaning and statistical principles;

[0088] It should be noted that the data preprocessing step is crucial to improve the prediction accuracy of the model. By cleaning and filling missing values, data noise can be reduced and data quality can be improved. Feature engineering further enhances the expressiveness of data, enabling the model to better capture the potential rules in the data.

[0089] S3, based on the principle of thermodynamics, a basic physical model describing the heat change in the building interior is constructed, and a data-driven model is constructed based on the preprocessed data and machine learning algorithm;

[0090] Further, the energy balance equation method is used to model the heat change in the building interior;

[0091] According to the building material properties, structural design factors and environmental parameters, the heat exchange per unit time is calculated to obtain the basic physical model E mech (t), the expression is:

[0092]

[0093] Where k is the thermal conductivity of the building material, A is the wall area, d is the wall thickness, η is the solar radiation absorption efficiency, T in (t) is the indoor temperature at time t, T out (t) is the outdoor temperature at time t, I solar (t) is the solar radiation intensity at time t;

[0094] The preprocessed data is analyzed by using machine learning algorithm training method;

[0095] The long short-term memory network LSTM is selected to construct the data-driven model E data (t), the expression is:

[0096] f LSTM (X(t),h t-1 ,c t-1 )=σ(W x ·X(t)+W h ·h t-1 +b);

[0097] E data (t)=f LSTM (X(t),h t-1 ,c t-1 );

[0098] Where f LSTM is the core calculation function of the long short-term memory network, Wx is the input data weight matrix, W h is the hidden state weight matrix, b is the bias vector, σ is the Sigmoid function, X(t) is the input feature vector at time t, including indoor temperature, humidity, equipment running state, outdoor temperature, wind speed and solar radiation intensity, h t-1 is the hidden state vector at time t-1, c t-1 is the cell state vector at time t-1.

[0099] It should be noted that the combination of the basic physical model and the data-driven model fully utilizes the advantages of physical knowledge and data-driven, which can accurately describe the physical process of heat change in the building and flexibly adapt to complex and variable actual environment, improving the robustness and prediction accuracy of the model.

[0100] S4, combine the basic physical model with the data-driven model to form a comprehensive data-mechanism fusion model, obtain an energy consumption prediction fusion model, and input the preprocessed data into the energy consumption prediction fusion model to output the energy consumption prediction value;

[0101] Further, a weight coefficient α adjustment method is used to balance the relative importance between the basic physical model E mech (t) and the data-driven model E data (t), and a comprehensive data-mechanism fusion model is constructed, and the expression is:

[0102]

[0103] Wherein, E mech (t) is the energy consumption value predicted by the basic physical model at time t, E data (t) is the energy consumption value predicted by the data-driven model at time t, and α is the weight coefficient for adjusting the relative contribution of the two;

[0104] An optimization method based on historical data is used to optimize the weight coefficient α, and the optimal α * is determined by minimizing the prediction error, and the expression is:

[0105]

[0106] Wherein, E actual,i is the actual energy consumption value at the i th time point, is the energy consumption value predicted by the data-mechanism fusion model at the i th time point, and n is the total number of time points;

[0107] Input the preprocessed data set into the energy consumption prediction fusion model;

[0108] The input feature vector X(t) of each time point t is predicted to obtain the energy consumption prediction value of each time point

[0109] It should be noted that by introducing the weight coefficient α, the contribution ratio of the two models can be dynamically adjusted according to the actual situation, thereby maximizing the prediction accuracy on the basis of ensuring the stability of the model, and the method is especially suitable for complex situations that are difficult to fully explain by physical models.

[0110] S5, using scenario analysis and simulation method, using Monte Carlo simulation statistical method to generate multiple possible scenarios, evaluate the possibility of building energy consumption under different conditions, and based on the energy consumption prediction value, perform scenario simulation, output multiple energy consumption prediction results;

[0111] Further, the Monte Carlo simulation method is used to randomly sample the key parameters affecting building energy consumption to obtain a random sample set;

[0112] Input the random sample into the energy consumption prediction fusion model, perform energy consumption prediction for each random sample, calculate the corresponding energy consumption prediction value under the sample condition, and the expression is:

[0113]

[0114] Wherein, E mech (t) (i) is the predicted energy consumption value of the basic physical model under the i th sample condition, E data (t) (i) is the predicted energy consumption value of the data-driven model under the i th sample condition, and α is the weight coefficient;

[0115] The energy consumption prediction results of all random samples are summarized and analyzed by statistical analysis method, the average energy consumption prediction value and its standard deviation under different conditions are calculated to evaluate the possibility of building energy consumption under different external conditions, and the expression is:

[0116]

[0117] Wherein, N is the number of random samples, is the average energy consumption prediction value, and σ E (t) is the average energy consumption standard deviation;

[0118] Output multiple energy consumption prediction results, and the average energy consumption prediction value and standard deviation calculated under different conditions are taken as the final scenario analysis results;

[0119] It should be noted that the Monte Carlo simulation method simulates multiple possible scenarios by randomly sampling key parameters, provides a comprehensive prediction of future energy consumption trends, which not only helps to identify high-risk periods, but also provides a scientific basis for developing response strategies.

[0120] S6. Based on multiple energy consumption prediction results, use energy efficiency optimization recommendation methods to provide building managers with specific energy-saving suggestions and obtain energy consumption recommendation strategies;

[0121] Furthermore, we analyze various energy consumption forecast results, conduct a comprehensive evaluation of the average energy consumption forecast values ​​and their standard deviations under different scenarios, and identify high and low energy consumption periods;

[0122] An energy efficiency optimization model is established. Based on the identified high energy consumption periods and the actual operation of the building, an optimization model is constructed to minimize the total energy consumption. The expression is:

[0123]

[0124] Where n is the total number of time points, and λ is the risk coefficient, which is used to balance the relationship between the expected value of energy consumption and volatility;

[0125] Use linear programming algorithm to find the optimal control parameter combination;

[0126] The control parameter combination includes indoor temperature setpoint, equipment start and stop time, and lighting system brightness adjustment, so as to minimize total energy consumption while meeting comfort requirements;

[0127] Formulate specific energy-saving measures based on the optimal control parameter combination obtained;

[0128] Recommended energy-saving measures include adjusting the operating temperature range of air conditioners, optimizing the on / off schedule of lighting systems, and rationally arranging equipment usage hours;

[0129] It should be noted that the energy efficiency optimization model not only takes into account the average level of energy consumption, but also takes into account the volatility of energy consumption, ensuring that the proposed energy-saving measures are both effective and robust. By combining the actual operating conditions, the specific energy-saving suggestions formulated are more operational and help achieve the goals of energy conservation and emission reduction.

[0130] A multi-scenario building energy consumption prediction device driven by data-mechanism fusion is also provided, including:

[0131] Data collection module, data preprocessing module, model building module, energy consumption prediction module, scenario simulation module and energy efficiency optimization module;

[0132] Data collection module, used to collect building energy consumption related data from the building management platform BMS, weather station and historical energy consumption records;

[0133] The data preprocessing module is used to clean and preprocess the collected building energy consumption data to obtain a high-quality data set;

[0134] a model construction module configured to construct a basic physical model and a data-driven model, and combine the two to form a comprehensive data-mechanism fusion model;

[0135] an energy consumption prediction module configured to input the preprocessed data into the energy consumption prediction fusion model, and output energy consumption prediction values at each time point;

[0136] a scenario simulation module configured to generate a plurality of possible scenarios by using a Monte Carlo simulation method, and evaluate the possibility of building energy consumption under different conditions;

[0137] an energy efficiency optimization module configured to provide specific energy-saving suggestions for building managers based on a plurality of energy consumption prediction results.

[0138] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A multi-scenario prediction method for building energy consumption based on data-mechanism fusion, characterized by: include: The building energy consumption data is collected from the building management platform BMS, weather station and historical energy consumption records using data collection methods to obtain building energy consumption data; The building energy consumption data is preprocessed using a data preprocessing method to obtain a preprocessed building energy consumption data set; Build a basic physical model based on thermodynamic principles to describe the heat changes inside the building, and build a data-driven model based on pre-processed data and machine learning algorithms; Combine the basic physical model with the data-driven model to form a comprehensive data-mechanism fusion model to obtain the energy consumption prediction fusion model. Input the pre-processed data into the energy consumption prediction fusion model to output the energy consumption prediction value. Adopting scenario analysis and simulation methods, using Monte Carlo simulation statistical methods to generate multiple possible scenarios, evaluate the possibility of building energy consumption under different conditions, and conduct scenario simulation based on energy consumption forecast values ​​to output multiple energy consumption forecast results; Based on a variety of energy consumption prediction results, the energy efficiency optimization recommendation method is used to provide building managers with specific energy-saving suggestions and obtain energy consumption recommendation strategies.

2. The multi-scenario prediction method for building energy consumption based on data-mechanism fusion drive according to claim 1 is characterized by: The data collection method is used to collect building energy consumption data from the building management platform BMS, weather station and historical energy consumption records. The specific steps are as follows: Use API interface to obtain indoor temperature, humidity and equipment operation status data from the building management platform BMS; The open API interface provided by the weather station is used to collect weather parameters such as outdoor temperature, wind speed, and solar radiation intensity; The energy consumption values ​​in the past time period are extracted by querying historical energy consumption records, and are integrated with the above data to form a unified data set to obtain building energy consumption data.

3. The multi-scenario prediction method for building energy consumption based on data-mechanism fusion drive according to claim 2 is characterized by: The data preprocessing method is used to preprocess the building energy consumption data to obtain a preprocessed building energy consumption data set. The specific steps are: Use data cleaning methods to remove outliers from the original data; By calculating the difference between each data point and its adjacent points, data points that significantly deviate from the normal range are identified and removed to obtain a preliminary cleaned data set; The missing value filling method is used to fill the missing values ​​in the data set after preliminary cleaning. According to the time series characteristics of the data, the linear interpolation method is used to fill the missing data points to obtain a data set without missing values; Feature engineering methods are used to extract and transform features from datasets without missing values; The feature extraction and conversion include standardization, normalization processing, and creation of new feature variables based on physical meaning and statistical principles.

4. The multi-scenario prediction method for building energy consumption based on data-mechanism fusion drive according to claim 3 is characterized by: The method constructs a basic physical model based on thermodynamic principles to describe the heat changes inside the building, and constructs a data-driven model based on preprocessed data and machine learning algorithms. The specific steps are as follows: The energy balance equation method is used to model the heat changes inside the building; According to the building material properties, structural design factors and environmental parameters, the heat exchange per unit time is calculated to obtain the basic physical model E mech (t), the expression is: Where k is the thermal conductivity of the building material, A is the wall area, d is the wall thickness, η is the solar radiation absorption efficiency, T in (t) is the indoor temperature at time t, T out (t) is the outdoor temperature at time t, I solar (t) is the solar radiation intensity at time t; Use machine learning algorithm training methods to analyze the preprocessed data; Select the long short-term memory network LSTM to build the data-driven model E data (t), the expression is: f LSTM (X(t),h t-1 ,c t-1 )=σ(W x ·X(t)+W h ·h t-1 +b); E data (t)=f LSTM (X(t),h t-1 ,c t-1 ); Among them, f LSTM is the core computing function of the long short-term memory network, W x is the input data weight matrix, W h is the hidden state weight matrix, b is the bias vector, σ is the Sigmoid function, X(t) is the input feature vector at time t, including indoor temperature, humidity, equipment operating status, outdoor temperature, wind speed and solar radiation intensity, h t-1 is the hidden state vector at time t-1, c t-1 is the cell state vector at time t-1.

5. The multi-scenario prediction method for building energy consumption based on data-mechanism fusion drive according to claim 4 is characterized by: The basic physical model is combined with the data-driven model to form a comprehensive data-mechanism fusion model, and the energy consumption prediction fusion model is obtained. The pre-processed data is input into the energy consumption prediction fusion model to output the energy consumption prediction value. The specific steps are as follows: The weight coefficient α is used to adjust the balance of the basic physical model E mech (t) and data-driven model E data (t) to construct a comprehensive data-mechanism fusion model, which is expressed as: Among them, E mech (t) is the energy consumption value predicted by the basic physical model at time t, E data (t) is the energy consumption value predicted by the data-driven model at time t, and α is the weight coefficient used to adjust the relative contribution of the two; The method of optimizing the weight coefficient α based on historical data is used to determine the optimal α by minimizing the prediction error. * , the expression is: Among them, E actual,i is the actual energy consumption value at the i-th time point, is the energy consumption value predicted by the data-mechanism fusion model at the i-th time point, and n is the total number of time points; Input the preprocessed data set into the energy consumption prediction fusion model; Predict the input feature vector X(t) at each time point t to obtain the energy consumption prediction value at each time point 6. The multi-scenario prediction method for building energy consumption based on data-mechanism fusion drive according to claim 5 is characterized by: The scenario analysis and simulation method uses the Monte Carlo simulation statistical method to generate multiple possible scenarios, evaluate the possibility of building energy consumption under different conditions, and perform scenario simulation based on the energy consumption forecast value to output multiple energy consumption forecast results. The specific steps are as follows: The Monte Carlo simulation method is used to randomly sample the key parameters that affect building energy consumption and obtain a random sample set; Input random samples into the energy consumption prediction fusion model, perform energy consumption prediction for each random sample, and calculate the corresponding energy consumption prediction value under the sample conditions. The expression is: Among them, E mech (t) (i) is the predicted energy consumption value of the basic physical model under the i-th sample condition, E data (t) (i) is the predicted energy consumption value of the data-driven model under the i-th sample condition, and α is the weight coefficient; The energy consumption prediction results of all random samples are summarized and analyzed using statistical analysis methods. The average energy consumption prediction value and its standard deviation under different conditions are calculated to evaluate the possibility of building energy consumption under different external conditions. The expression is: Where N is the number of random samples, is the average energy consumption prediction value, σ E (t) is the standard deviation of average energy consumption; Output a variety of energy consumption forecast results, and use the average energy consumption forecast value and standard deviation calculated under different conditions as the final scenario analysis result.

7. The multi-scenario prediction method for building energy consumption based on data-mechanism fusion drive according to claim 6 is characterized by: Based on multiple energy consumption prediction results, the energy efficiency optimization suggestion method is used to provide specific energy-saving suggestions to building managers and obtain energy consumption suggestion strategies. The specific steps are as follows: Analyze various energy consumption forecast results, conduct a comprehensive assessment of the average energy consumption forecast values ​​and their standard deviations under different scenarios, and identify periods of high and low energy consumption; An energy efficiency optimization model is established. Based on the identified high energy consumption periods and the actual operation of the building, an optimization model is constructed to minimize the total energy consumption. The expression is: Where n is the total number of time points, and λ is the risk coefficient, which is used to balance the relationship between the expected value of energy consumption and volatility; Use linear programming algorithm to find the optimal control parameter combination; The control parameter combination includes indoor temperature setting value, equipment start and stop time, and lighting system brightness adjustment, so as to minimize total energy consumption while meeting comfort requirements; Formulate specific energy-saving measures based on the optimal control parameter combination obtained; The energy-saving measures suggested include adjusting the operating temperature range of air conditioners, optimizing the on / off schedule of lighting systems, and rationally arranging equipment usage time.

8. A device for predicting building energy consumption in multiple scenarios based on data-mechanism fusion drive, based on the method for predicting building energy consumption in multiple scenarios based on data-mechanism fusion drive according to any one of claims 1 to 7, characterized in that: include: Data collection module, data preprocessing module, model building module, energy consumption prediction module, scenario simulation module and energy efficiency optimization module; The data collection module is used to collect building energy consumption related data from the building management platform BMS, weather station and historical energy consumption records; The data preprocessing module is used to clean and preprocess the collected building energy consumption data to obtain a high-quality data set; The model building module is used to build a basic physical model and a data-driven model, and combine the two to form a comprehensive data-mechanism fusion model; The energy consumption prediction module is used to input the pre-processed data into the energy consumption prediction fusion model and output the energy consumption prediction value at each time point; The scenario simulation module is used to generate multiple possible scenarios using the Monte Carlo simulation method to evaluate the likelihood of building energy consumption under different conditions; The energy efficiency optimization module is used to provide building managers with specific energy-saving suggestions based on multiple energy consumption prediction results.

Images