A heating energy regulation method based on building energy saving and emission reduction

Abstract

The present invention discloses a heating energy regulation method based on building energy conservation and emission reduction, comprising the following steps: S1, collecting internal and external environmental data of the building; S2, preprocessing the environmental data and cleaning the noise data; S3, using a deep reinforcement learning algorithm to train the scheduling strategy of the HVAC system; S4, dividing the building into multiple areas, and scheduling the HVAC equipment in each area based on a multi-agent system; S5, adjusting the operating parameters of the HVAC system in each area through fuzzy reasoning control; S6, combining particle swarm optimization with genetic algorithm for multi-objective scheduling optimization; S7, integrating renewable energy systems and intelligent energy storage equipment to optimize energy supply and energy storage; S8, using real-time feedback and prediction models to dynamically adjust the load scheduling strategy. The present invention adjusts the load scheduling strategy according to the energy consumption data of the building and the future energy consumption forecast through a real-time feedback mechanism, thereby achieving the dual goals of building energy conservation and emission reduction and indoor comfort.

Description

Technical Field

[0001] The invention relates to the technical field of intelligent buildings, and in particular to a heating energy regulation method based on building energy saving and emission reduction. Background Art

[0002] As the world pays more attention to energy conservation and environmental protection, improving building energy efficiency has become an important issue. In modern buildings, HVAC systems are one of the main sources of energy consumption. Traditional HVAC systems usually adopt a scheduling method based on fixed rules, and the system will operate according to a preset schedule and set temperature. This method lacks flexibility and is difficult to cope with real-time changes in the building's internal and external environment, resulting in energy waste and inefficiency.

[0003] Existing traditional HVAC systems mainly rely on manually set control modes and cannot be flexibly adjusted according to the temperature requirements of various areas within the building and the activities of personnel. Especially under extreme weather conditions, HVAC systems are difficult to achieve precise temperature regulation, which often results in energy waste. For example, during peak hours in winter or summer, the system does not have sufficient adaptive capabilities to adjust according to the actual needs of various areas within the building, resulting in uneconomical energy use.

[0004] In addition, although some modern systems have introduced environmental perception technology, using sensors such as temperature and humidity to collect data and make corresponding adjustments, these systems still rely mainly on rules and preset algorithms to operate, lacking intelligent dynamic scheduling mechanisms. Although some studies in recent years have tried to combine machine learning and optimization algorithms to improve scheduling efficiency, these methods usually only focus on optimizing a single goal, such as temperature control or energy saving, and ignore the comprehensive scheduling needs of multiple goals and multiple factors.

[0005] Although advanced algorithms such as reinforcement learning have been applied to the scheduling optimization of HVAC systems, the challenges faced by existing technologies are still great. First, the training process of the algorithm requires a large amount of historical data, and these data are usually delayed, resulting in the system's untimely response to emergencies. Secondly, although technologies such as reinforcement learning have self-learning capabilities, their training process is time-consuming and computationally intensive, and it is difficult to adapt to the needs of different building environments. In addition, the optimization of building energy management by existing technologies mostly remains at the level of a single goal, fails to consider the integration of green energy and intelligent energy storage systems, and cannot achieve more comprehensive energy optimization and management.

[0006] Therefore, how to provide a heating energy regulation method based on building energy conservation and emission reduction is an urgent problem that technicians in this field need to solve. Summary of the invention

[0007] One purpose of the present invention is to propose a heating energy regulation method based on building energy conservation and emission reduction. The present invention proposes an intelligent HVAC system scheduling method by combining multiple advanced technologies such as deep reinforcement learning, fuzzy reasoning control, particle swarm optimization and genetic algorithm. The method can significantly improve the energy efficiency of buildings and optimize system operation under complex environmental conditions. Compared with the traditional rule-based scheduling method, the present invention has significant advantages and can better adapt to changes in the internal and external environment of the building to achieve flexible and accurate temperature control management. By introducing deep reinforcement learning, multi-agent systems and advanced optimization algorithms, the system can dynamically adjust the scheduling strategy according to real-time environmental data and energy consumption information, thereby improving the intelligence level and operation efficiency of the system. , Provide an intelligent, efficient, energy-saving and adaptable to complex environmental changes HVAC system scheduling method, thereby improving system performance while contributing to green buildings and sustainable development.

[0008] According to an embodiment of the present invention, a heating energy regulation method based on building energy conservation and emission reduction includes the following steps:

[0009] S1. Deploy a variety of environmental sensors through the Internet of Things technology to collect data on the internal and external environment of the building and generate an initial data set;

[0010] S2. Preprocessing the initial data set, including data cleaning, denoising and standardization, constructing an environmental data model, and generating a standardized environmental data model;

[0011] S3. Based on deep reinforcement learning algorithm, use environmental data model to train the scheduling strategy of HVAC system to optimize the energy efficiency and comfort control of HVAC equipment;

[0012] S4. Based on the environmental data model and scheduling strategy, use the multi-agent system to divide the building into zones, and dispatch the corresponding HVAC equipment according to the environmental data and needs of each zone;

[0013] S5. Based on the environmental data and scheduling strategy of each area, the fuzzy inference control mechanism is used to dynamically adjust the HVAC system operating parameters of each area;

[0014] S6. Based on the environmental data model, combined with the scheduling strategy and fuzzy reasoning control output, a combination of particle swarm optimization and genetic algorithm is used to perform multi-objective optimization scheduling;

[0015] S7. Based on the multi-objective optimization scheduling results, integrate renewable energy systems and intelligent energy storage equipment to optimize green energy supply and storage;

[0016] S8. Based on renewable energy supply and energy storage status, combined with real-time feedback mechanism, using building energy consumption data and future energy consumption forecasts, adjust the load scheduling strategy.

[0017] Optionally, the S2 specifically includes:

[0018] S2 1. Preprocess the initial data set and clean the collected environmental data to remove missing values ​​and outliers;

[0019] S22, using a denoising method based on statistical analysis to process the noise in the environmental data;

[0020] S23, standardize the environmental data and define the data standardization function as:

[0021]

[0022] Among them, x is the original data value, μ is the data mean, σ is the data standard deviation, x norm is the standardized data value;

[0023] S24. Construct an environmental data model based on the standardized environmental data, and define the model as follows:

[0024] y=f(x norm ,θ);

[0025] Among them, y is the output environment prediction result, x norm is the standardized input data, θ is the model parameter, and f(·) is the environmental data prediction function;

[0026] S25. Use machine learning methods to train environmental data models to minimize prediction errors;

[0027] S26. Generate a standardized environmental data model, verify the environmental data model, calculate the model error, and use a cross-validation method to evaluate the performance of the model.

[0028] Optionally, the S3 specifically includes:

[0029] S31. Build a deep reinforcement learning model, consider the HVAC system as an intelligent scheduling decision environment, define the state space as S, the action space as A, and the reward function as R;

[0030] S32. Utilize environmental data x norm As the input state S of the deep reinforcement learning model t , taking the system state at each time step t as the basis for the current decision;

[0031] S33, based on the current state S t , generate an action a through the neural network strategy t ∈A, action a t To dispatch control instructions for HVAC equipment;

[0032] S34, according to the action taken a t , execute the scheduling strategy and observe the reward r fed back by the system t , the reward function is defined as:

[0033] r t =f(S t , a t , S t+1 );

[0034] Among them, S t is the current state, a t For the action to be performed, S t+1 is the next state transferred to after executing the action, and f(·) is the reward function;

[0035] S35, use the reinforcement learning algorithm to update the intelligent scheduling strategy so that the cumulative reward R total maximize:

[0036]

[0037] Among them, γ is the discount factor, T is the training period, r t is the reward for each time step;

[0038] S36, updating the parameter θ in the neural network through the back propagation algorithm to optimize the scheduling strategy and converge to the optimal scheduling strategy;

[0039] S37. During the training process, the experience replay mechanism is used to store historical states, actions, and rewards, and batch updates are performed through random sampling.

[0040] Optionally, the S4 specifically includes:

[0041] S41. Divide the building into multiple functional areas according to the needs and environmental characteristics of different functional areas in the building, and each functional area is managed and scheduled by an independent intelligent agent;

[0042] S42, each intelligent agent calculates the operating parameters of the HVAC equipment in the area based on the environmental data of the area and the comfort requirements of the area;

[0043] S43, each intelligent agent generates a HVAC equipment scheduling strategy for the area based on the temperature, humidity and air quality data of the area and the real-time status of the equipment in the area;

[0044] S44, when generating a scheduling strategy, each intelligent agent optimizes the equipment operation strategy by combining the power and energy efficiency of the equipment in the area;

[0045] S45, the agent coordinates with other agents through the information sharing mechanism to adjust the scheduling strategy;

[0046] S46. Each intelligent agent adjusts its scheduling strategy based on external feedback and real-time data and schedules the corresponding HVAC equipment.

[0047] Optionally, the S5 specifically includes:

[0048] S51. Based on the environmental data and scheduling strategy of each area, define fuzzy sets and build a fuzzy control rule base;

[0049] S52, inputting the collected real-time environmental data into the fuzzy inference system, performing fuzzification processing on the input variables, and generating a membership value;

[0050] S53, according to the fuzzified input variables and the predefined fuzzy control rule base, fuzzy reasoning is performed to calculate the fuzzy result of each output variable. The fuzzy reasoning process is expressed as:

[0051] R j : IF x1 IS A1 AND x2 IS A2 THEN y j IS B j ;

[0052] Among them, x1, x2 are input variables, A1, A2 are corresponding fuzzy sets, y j is the output variable;

[0053] S54, by using the defuzzification method, the fuzzy reasoning output is converted into actual control instructions, and the control parameters are obtained using the center average method:

[0054]

[0055] Among them, μ i is the membership value, B i is the possible value of the fuzzy output, y j is the actual control parameter;

[0056] S55. Dynamically adjust the HVAC system operating parameters of each area through real-time feedback.

[0057] Optionally, the S6 specifically includes:

[0058] S61. Based on the environmental data model, obtain environmental data and energy efficiency requirements of each area, and define multiple optimization objectives;

[0059] S62. According to the defined objectives, a multi-objective scheduling model is constructed, and the objective function in the model is:

[0060] f(x)=α1·f1(x)+α2·f2(x)+…+α n ·f n (x);

[0061] Among them, f1(x), f2(x), ..., f n (x) are the objective functions, α1, α2, ..., α n is the target weight coefficient;

[0062] S63. Use the particle swarm optimization algorithm to optimize each target and update the particle position and velocity. The particle position update formula is:

[0063]

[0064] in, is the position of the particle in the kth iteration, is the speed of the particle; the speed update formula is:

[0065]

[0066] Among them, w is the inertia weight, c1, c2 are learning factors, r1, r2 are random numbers, p i is the historical optimal position of the particle, g i is the global optimal position;

[0067] S64, the particle swarm population is evolved through genetic algorithm, selection, crossover and mutation operations are used to generate new populations, and the quality of individuals is evaluated according to the fitness function, and the best individuals are selected for the next generation of reproduction;

[0068] S65. Achieving the global optimal solution for multi-objective scheduling by combining particle swarm optimization and genetic algorithm;

[0069] S66. Adjust the HVAC system scheduling strategy according to the optimal solution and perform multi-objective optimization scheduling.

[0070] Optionally, the S8 specifically includes:

[0071] S81. Based on the renewable energy supply and storage status, the building's energy consumption data, including the use of electricity, heat and cooling energy, can be obtained in real time through intelligent sensors and data acquisition systems;

[0072] S82. Based on historical energy consumption data and real-time data, the regression analysis method is used to predict future energy consumption. The prediction formula is as follows:

[0073] P(t)=a·P(t-1)+b·P(t-2)+c;

[0074] Where P(t) is the predicted energy consumption at time t, a and b are regression coefficients, and c is a constant term;

[0075] S83. Using the difference between the predicted energy consumption data and the real-time energy consumption data, a load adjustment signal is generated, and a fuzzy logic controller is used for adjustment. The output signal of the fuzzy logic controller is:

[0076] ΔP=f(ΔE);

[0077] Among them, ΔP is the load adjustment signal, and ΔE is the energy consumption difference;

[0078] S84, adjusting the energy supply of the building and the load of the energy storage device according to the regulation signal;

[0079] S85, monitor the energy consumption changes of buildings in real time and adjust the scheduling strategy based on real-time feedback;

[0080] S86. When sudden load fluctuations occur, adjust the load dispatching strategy.

[0081] The beneficial effects of the present invention are:

[0082] This paper proposes an intelligent HVAC system scheduling method by combining advanced technologies such as deep reinforcement learning, fuzzy reasoning control, particle swarm optimization and genetic algorithm. Compared with traditional scheduling methods, this paper has significant advantages and can better adapt to environmental changes and improve the energy efficiency and comfort of the system.

[0083] First, the Internet of Things technology is used to deploy environmental sensors to collect real-time environmental data inside and outside the building, such as temperature, humidity, air quality, etc., to provide a basis for real-time decision-making for the system. Compared with traditional methods, this invention can adjust the system operation in real time and avoid energy waste caused by data delays.

[0084] Secondly, the introduction of deep reinforcement learning algorithms enables the system to self-learn and optimize scheduling strategies based on real-time data. The adaptive ability of reinforcement learning allows the system to continuously adjust the control strategy according to the actual environment, avoiding the rigidity of traditional rule-driven systems.

[0085] Through the multi-agent system, the present invention divides the building into multiple zones, and each zone independently dispatches HVAC equipment to accurately control temperature and energy efficiency. This not only improves comfort, but also avoids the temperature difference problem that may be caused by unified dispatch of the entire building.

[0086] In addition, the present invention integrates renewable energy and intelligent energy storage equipment to optimize energy use and reduce dependence on traditional energy. The utilization of green energy and the management of the energy storage system further improve the energy efficiency of the system.

[0087] Finally, based on the real-time feedback mechanism, the system can dynamically adjust the scheduling strategy according to energy consumption data and future forecasts. This feedback adjustment improves the system's adaptive ability and ensures optimal operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0089] Figure 1 A flowchart of a heating energy regulation method based on building energy conservation and emission reduction proposed by the present invention;

[0090] Figure 2 This is an architectural diagram of a heating energy regulation method based on building energy conservation and emission reduction proposed by the present invention. DETAILED DESCRIPTION

[0091] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0092] refer to Figure 1 and Figure 2 A heating energy regulation method based on building energy conservation and emission reduction comprises the following steps:

[0093] S1. Deploy a variety of environmental sensors through the Internet of Things technology to collect data on the internal and external environment of the building and generate an initial data set;

[0094] S2. Preprocessing the initial data set, including data cleaning, denoising and standardization, constructing an environmental data model, and generating a standardized environmental data model;

[0095] S3. Based on deep reinforcement learning algorithm, use environmental data model to train the scheduling strategy of HVAC system to optimize the energy efficiency and comfort control of HVAC equipment;

[0096] S4. Based on the environmental data model and scheduling strategy, use the multi-agent system to divide the building into zones, and dispatch the corresponding HVAC equipment according to the environmental data and needs of each zone;

[0097] S5. Based on the environmental data and scheduling strategy of each area, the fuzzy inference control mechanism is used to dynamically adjust the HVAC system operating parameters of each area;

[0098] S6. Based on the environmental data model, combined with the scheduling strategy and fuzzy reasoning control output, a combination of particle swarm optimization and genetic algorithm is used to perform multi-objective optimization scheduling;

[0099] S7. Based on the multi-objective optimization scheduling results, integrate renewable energy systems and intelligent energy storage equipment to optimize green energy supply and storage;

[0100] S8. Based on renewable energy supply and energy storage status, combined with real-time feedback mechanism, using building energy consumption data and future energy consumption forecasts, adjust the load scheduling strategy.

[0101] In this implementation, S2 specifically includes:

[0102] S21. Preprocess the initial data set and clean the collected environmental data to remove missing values ​​and outliers.

[0103] S22, using a denoising method based on statistical analysis to process the noise in the environmental data;

[0104] S23, standardize the environmental data and define the data standardization function as:

[0105]

[0106] Among them, x is the original data value, μ is the data mean, σ is the data standard deviation, x norm is the standardized data value;

[0107] S24. Construct an environmental data model based on the standardized environmental data, and define the model as follows:

[0108] y=f(x norm ,θ);

[0109] Among them, y is the output environment prediction result, x norm is the standardized input data, θ is the model parameter, and f(·) is the environmental data prediction function;

[0110] S25. Use machine learning methods to train environmental data models to minimize prediction errors;

[0111] S26. Generate a standardized environmental data model, verify the environmental data model, calculate the model error, and use a cross-validation method to evaluate the performance of the model.

[0112] In this implementation, S3 specifically includes:

[0113] S31. Build a deep reinforcement learning model, consider the HVAC system as an intelligent scheduling decision environment, define the state space as S, the action space as A, and the reward function as R;

[0114] S32. Utilize environmental data x norm As the input state S of the deep reinforcement learning model t, taking the system state at each time step t as the basis for the current decision;

[0115] S33, based on the current state S t , generate an action a through the neural network strategy t ∈A, action a t To dispatch control instructions for HVAC equipment;

[0116] S34, according to the action taken a t , execute the scheduling strategy and observe the reward r fed back by the system t , the reward function is defined as:

[0117] r t =f(S t , a t , S t+1 );

[0118] Among them, S t is the current state, a t For the action to be performed, S t+1 is the next state transferred to after executing the action, and f(·) is the reward function;

[0119] S35, use the reinforcement learning algorithm to update the intelligent scheduling strategy so that the cumulative reward R total maximize:

[0120]

[0121] Among them, γ is the discount factor, T is the training period, r t is the reward for each time step;

[0122] S36, updating the parameter θ in the neural network through the back propagation algorithm to optimize the scheduling strategy and converge to the optimal scheduling strategy;

[0123] S37. During the training process, the experience replay mechanism is used to store historical states, actions, and rewards, and batch updates are performed through random sampling.

[0124] In this implementation manner, the S4 specifically includes:

[0125] S41. Divide the building into multiple functional areas according to the needs and environmental characteristics of different functional areas in the building, and each functional area is managed and scheduled by an independent intelligent agent;

[0126] S42, each intelligent agent calculates the operating parameters of the HVAC equipment in the area based on the environmental data of the area and the comfort requirements of the area;

[0127] S43, each intelligent agent generates a HVAC equipment scheduling strategy for the area based on the temperature, humidity and air quality data of the area and the real-time status of the equipment in the area;

[0128] S44, when generating a scheduling strategy, each intelligent agent optimizes the equipment operation strategy by combining the power and energy efficiency of the equipment in the area;

[0129] S45, the agent coordinates with other agents through the information sharing mechanism to adjust the scheduling strategy;

[0130] S46. Each intelligent agent adjusts its scheduling strategy based on external feedback and real-time data and schedules the corresponding HVAC equipment.

[0131] In this implementation manner, S5 specifically includes:

[0132] S51. Based on the environmental data and scheduling strategy of each area, define fuzzy sets and build a fuzzy control rule base;

[0133] S52, inputting the collected real-time environmental data into the fuzzy inference system, performing fuzzification processing on the input variables, and generating a membership value;

[0134] S53, according to the fuzzified input variables and the predefined fuzzy control rule base, fuzzy reasoning is performed to calculate the fuzzy result of each output variable. The fuzzy reasoning process is expressed as:

[0135] R j :IFx1 IS A1 AND x2 IS A2 THEN y j IS B j ;

[0136] Among them, x1, x2 are input variables, A1, A2 are corresponding fuzzy sets, y j is the output variable;

[0137] S54, by using the defuzzification method, the fuzzy reasoning output is converted into actual control instructions, and the control parameters are obtained using the center average method:

[0138]

[0139] Among them, μ i is the membership value, B i is the possible value of the fuzzy output, y j is the actual control parameter;

[0140] S55. Dynamically adjust the HVAC system operating parameters of each area through real-time feedback.

[0141] In this implementation manner, S6 specifically includes:

[0142] S61. Based on the environmental data model, obtain environmental data and energy efficiency requirements of each area, and define multiple optimization objectives;

[0143] S62. According to the defined objectives, a multi-objective scheduling model is constructed, and the objective function in the model is:

[0144] f(x)=α1·f1(x)+α2·f2(x)+…+α n ·f n (x);

[0145] Among them, f1(x), f2(x), ..., f n (x) are the objective functions, α1, α2, ..., α n is the target weight coefficient;

[0146] S63. Use the particle swarm optimization algorithm to optimize each target and update the particle position and velocity. The particle position update formula is:

[0147]

[0148] in, is the position of the particle in the kth iteration, is the speed of the particle; the speed update formula is:

[0149]

[0150] Among them, w is the inertia weight, c1, c2 are learning factors, r1, r2 are random numbers, p i is the historical optimal position of the particle, g i is the global optimal position;

[0151] S64, the particle swarm population is evolved through genetic algorithm, selection, crossover and mutation operations are used to generate new populations, and the quality of individuals is evaluated according to the fitness function, and the best individuals are selected for the next generation of reproduction;

[0152] S65. Achieving the global optimal solution for multi-objective scheduling by combining particle swarm optimization and genetic algorithm;

[0153] S66. Adjust the HVAC system scheduling strategy according to the optimal solution and perform multi-objective optimization scheduling.

[0154] In this implementation manner, S8 specifically includes:

[0155] S81. Based on the renewable energy supply and storage status, the building's energy consumption data, including the use of electricity, heat and cooling energy, can be obtained in real time through intelligent sensors and data acquisition systems;

[0156] S82. Based on historical energy consumption data and real-time data, the regression analysis method is used to predict future energy consumption. The prediction formula is as follows:

[0157] P(t)=a·P(t-1)+b·P(t-2)+c;

[0158] Where P(t) is the predicted energy consumption at time t, a and b are regression coefficients, and c is a constant term;

[0159] S83. Using the difference between the predicted energy consumption data and the real-time energy consumption data, a load adjustment signal is generated, and a fuzzy logic controller is used for adjustment. The output signal of the fuzzy logic controller is:

[0160] ΔP=f(ΔE);

[0161] Among them, AP is the load adjustment signal, ΔE is the energy consumption difference;

[0162] S84, adjusting the energy supply of the building and the load of the energy storage device according to the regulation signal;

[0163] S85, monitor the energy consumption changes of buildings in real time and adjust the scheduling strategy based on real-time feedback;

[0164] S86. When sudden load fluctuations occur, adjust the load dispatching strategy.

[0165] Embodiment 1:

[0166] In order to verify the feasibility of the present invention in implementation, the present invention is applied to the energy management system of a commercial building, which is located in Pudong New District, Shanghai, with a construction area of ​​about 50,000 square meters, 10 floors of office area and 2 floors of underground parking. Due to the frequent personnel activities and diverse office equipment in the office area of ​​the building, and the fact that the periphery of the building is greatly affected by seasonal weather, the operating efficiency and energy management issues of the HVAC system have always been the focus of attention of managers. In order to prove the effectiveness of the method of the present invention, this embodiment implemented a three-month trial operation in the building.

[0167] In this embodiment, first, environmental sensors such as temperature and humidity sensors, air quality sensors, and light sensors are deployed through the Internet of Things technology and distributed in various areas of the building. These sensors collect environmental data of each floor and different areas in real time, including indoor and outdoor temperature, humidity, carbon dioxide concentration, air quality index and other information. The system collects data every 5 minutes, and the collected data is transmitted to the central data platform.

[0168] Subsequently, the data processing module is used to pre-process the collected environmental data, including denoising, missing value filling and standardization. Through these steps, the system obtains accurate environmental data models for each floor and different areas, providing a high-quality data foundation for the subsequent scheduling strategy formulation.

[0169] Next, the system applies a deep reinforcement learning algorithm to train the HVAC system's scheduling strategy. Specifically, the system automatically generates an operation scheduling strategy for the air conditioning system based on the preprocessed data, combined with the load demand in different areas of the building and external weather forecast data. During the training process of the reinforcement learning model, the system is encouraged to choose an operation strategy that maximizes energy efficiency while ensuring indoor comfort by setting a reward function.

[0170] At the same time, the system uses a multi-agent algorithm to divide the building into zones and schedules based on the real-time needs of each zone. For example, the system will flexibly adjust the operating intensity of the air-conditioning equipment according to the density of people on the floor, the operation of office equipment, and external weather changes. Based on the scheduling data of each zone, the system adjusts the operating parameters of the air-conditioning equipment in real time to ensure that the temperature and humidity of each zone remain within the optimal range.

[0171] To further optimize energy use, the system also integrates renewable energy systems and smart energy storage devices. Solar panels are installed on the roof of the building to collect solar energy and store excess energy through smart energy storage devices. This energy can be used to supply the HVAC system during peak load periods, thereby reducing dependence on traditional power grids, reducing energy costs and carbon emissions.

[0172] Finally, the system makes dynamic adjustments through a feedback mechanism. During system operation, the energy consumption and environmental changes inside and outside the building are monitored in real time. Based on the future energy consumption prediction model, the system can predict the upcoming peak period and adjust the scheduling strategy in advance to avoid overload and maximize the efficiency of the air conditioning system.

[0173] Table 1 Comparison of energy consumption and cost of air conditioning system

[0174]

[0175] Table 2 Comparison of comfort improvement and energy saving data of air conditioning system

[0176]

[0177] Table 3 Comparison of comfort improvement and energy saving data of air conditioning system

[0178]

[0179]

[0180] Table 1 shows the significant savings in energy consumption and costs of the HVAC system after applying the method of the present invention. In three months, a total of 31,500 kWh of energy consumption was saved, and 21.00% of the total cost was saved, amounting to 12,200 yuan. Table 2 shows the improvement in comfort after the application of the method of the present invention. The comfort score increased from an average of 71 points to 86 points, an increase of 21.13%. While ensuring the improvement in comfort, energy consumption and costs were also saved. Table 3 shows that the green energy utilization rate of the building has been significantly improved after integrating renewable energy systems and intelligent energy storage devices. After application, the green energy utilization rate increased from 30% to 45%, achieving a 15% increase.

[0181] From the analysis of the above data tables, it can be concluded that the method of the present invention has significant advantages over traditional methods in terms of energy saving, comfort improvement, and energy consumption optimization. Especially in the application of different building types, the scheduling strategy of the present invention can adjust the operation mode of the HVAC system in real time by combining environmental data, equipment operating status and external factors, thereby achieving the effect of maximizing energy saving and optimizing environmental comfort.

[0182] From the specific data, it can be seen that the method of the present invention can adapt to different building environments and intelligently optimize energy consumption, thereby effectively improving the energy efficiency of the building. In addition, the dynamic scheduling strategy implemented by the present invention through deep learning algorithms and multi-agent collaboration mechanisms is more intelligent than traditional static methods, and can adjust the load scheduling strategy based on real-time feedback and future energy consumption predictions, thereby improving the response speed and accuracy of the system.

[0183] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A heating energy regulation method based on building energy conservation and emission reduction, characterized in that: The steps include: S1. Deploy a variety of environmental sensors through the Internet of Things technology to collect data on the internal and external environment of the building and generate an initial data set; S2. Preprocessing the initial data set, including data cleaning, denoising and standardization, constructing an environmental data model, and generating a standardized environmental data model; S3. Based on deep reinforcement learning algorithm, use environmental data model to train the scheduling strategy of HVAC system to optimize the energy efficiency and comfort control of HVAC equipment; S4. Based on the environmental data model and scheduling strategy, use the multi-agent system to divide the building into zones, and dispatch the corresponding HVAC equipment according to the environmental data and needs of each zone; S5. Based on the environmental data and scheduling strategy of each area, the fuzzy inference control mechanism is used to dynamically adjust the HVAC system operating parameters of each area; S6. Based on the environmental data model, combined with the scheduling strategy and fuzzy reasoning control output, a combination of particle swarm optimization and genetic algorithm is used to perform multi-objective optimization scheduling; S7. Based on the multi-objective optimization scheduling results, integrate renewable energy systems and intelligent energy storage equipment to optimize green energy supply and storage; S8. Based on renewable energy supply and energy storage status, combined with real-time feedback mechanism, using building energy consumption data and future energy consumption forecasts, adjust the load scheduling strategy.

2. A heating energy regulation method based on building energy conservation and emission reduction according to claim 1, characterized in that: The S2 specifically includes: S21. Preprocess the initial data set and clean the collected environmental data to remove missing values ​​and outliers. S22, using a denoising method based on statistical analysis to process the noise in the environmental data; S23, standardize the environmental data and define the data standardization function as: Among them, x is the original data value, μ is the data mean, σ is the data standard deviation, x norm is the standardized data value; S24. Construct an environmental data model based on the standardized environmental data, and define the model as follows: y=f(xnorm,θ); Among them, y is the output environment prediction result, x norm is the standardized input data, θ is the model parameter, and f(·) is the environmental data prediction function; S25. Use machine learning methods to train environmental data models to minimize prediction errors; S26. Generate a standardized environmental data model, verify the environmental data model, calculate the model error, and use a cross-validation method to evaluate the performance of the model.

3. A heating energy regulation method based on building energy conservation and emission reduction according to claim 1, characterized in that: The S3 specifically includes: S31. Build a deep reinforcement learning model, consider the HVAC system as an intelligent scheduling decision environment, define the state space as S, the action space as A, and the reward function as R; S32. Utilize environmental data x norm As the input state S of the deep reinforcement learning model t , taking the system state at each time step t as the basis for the current decision; S33, based on the current state S t , generate an action a through the neural network strategy t ∈A, action a t To dispatch control instructions for HVAC equipment; S34, according to the action taken a t , execute the scheduling strategy and observe the reward r fed back by the system t , the reward function is defined as: rt=f(St,at,St +1 ); Among them, S t is the current state, a t For the action to be performed, S t+1 is the next state transferred to after executing the action, and f(·) is the reward function; S35, use the reinforcement learning algorithm to update the intelligent scheduling strategy so that the cumulative reward R total maximize: Among them, γ is the discount factor, T is the training period, r t is the reward for each time step; S36, updating the parameter θ in the neural network through the back propagation algorithm to optimize the scheduling strategy and converge to the optimal scheduling strategy; S37. During the training process, the experience replay mechanism is used to store historical states, actions, and rewards, and batch updates are performed through random sampling.

4. A heating energy regulation method based on building energy conservation and emission reduction according to claim 1, characterized in that: The S4 specifically includes: S41. Divide the building into multiple functional areas according to the needs and environmental characteristics of different functional areas in the building, and each functional area is managed and scheduled by an independent intelligent agent; S42, each intelligent agent calculates the operating parameters of the HVAC equipment in the area based on the environmental data of the area and the comfort requirements of the area; S43, each intelligent agent generates a HVAC equipment scheduling strategy for the area based on the temperature, humidity and air quality data of the area and the real-time status of the equipment in the area; S44, when generating a scheduling strategy, each intelligent agent optimizes the equipment operation strategy by combining the power and energy efficiency of the equipment in the area; S45, the agent coordinates with other agents through information sharing mechanism to adjust the scheduling strategy; S46. Each intelligent agent adjusts its scheduling strategy based on external feedback and real-time data and schedules the corresponding HVAC equipment.

5. A heating energy regulation method based on building energy conservation and emission reduction according to claim 1, characterized in that: The S5 specifically includes: S51. Based on the environmental data and scheduling strategy of each area, define fuzzy sets and build a fuzzy control rule base; S52, inputting the collected real-time environmental data into the fuzzy inference system, performing fuzzification processing on the input variables, and generating a membership value; S53, according to the fuzzified input variables and the predefined fuzzy control rule base, fuzzy reasoning is performed to calculate the fuzzy result of each output variable. The fuzzy reasoning process is expressed as: Rj:IFx1 ISA1 ANDx2 ISA2 THENyjIS Bj; Among them, x1, x2 are input variables, A1, A2 are corresponding fuzzy sets, y j is the output variable; S54, by using the defuzzification method, the fuzzy reasoning output is converted into actual control instructions, and the control parameters are obtained using the center average method: Among them, μ i is the membership value, B i is the possible value of the fuzzy output, y j is the actual control parameter; S55. Dynamically adjust the HVAC system operating parameters of each area through real-time feedback.

6. A heating energy regulation method based on building energy conservation and emission reduction according to claim 1, characterized in that: The S6 specifically includes: S61. Based on the environmental data model, obtain environmental data and energy efficiency requirements of each area, and define multiple optimization objectives; S62. According to the defined objectives, a multi-objective scheduling model is constructed, and the objective function in the model is: f(x)=α1·f1(x)+α2·f2(x)+…+αn·fn(x); Among them, f1(x),f2(x),…,f n (x) are the objective functions, α1, α2, …, α n is the target weight coefficient; S63. Use the particle swarm optimization algorithm to optimize each target and update the particle position and velocity. The particle position update formula is: in, is the position of the particle in the kth iteration, is the speed of the particle; the speed update formula is: Among them, w is the inertia weight, c1, c2 are learning factors, r1, r2 are random numbers, p i is the historical optimal position of the particle, g i is the global optimal position; S64, the particle swarm population is evolved through genetic algorithm, selection, crossover and mutation operations are used to generate new populations, and the quality of individuals is evaluated according to the fitness function, and the best individuals are selected for the next generation of reproduction; S65. Achieving the global optimal solution for multi-objective scheduling by combining particle swarm optimization and genetic algorithm; S66. Adjust the HVAC system scheduling strategy according to the optimal solution and perform multi-objective optimization scheduling.

7. A heating energy regulation method based on building energy conservation and emission reduction according to claim 1, characterized in that: The S8 specifically includes: S81. Based on the renewable energy supply and storage status, the building's energy consumption data, including the use of electricity, heat and cooling energy, can be obtained in real time through intelligent sensors and data acquisition systems; S82. Based on historical energy consumption data and real-time data, the regression analysis method is used to predict future energy consumption. The prediction formula is as follows: P(t)=a·P(t-1)+b·P(t-2)+c; Where P(t) is the predicted energy consumption at time t, a and b are regression coefficients, and c is a constant term; S83. Using the difference between the predicted energy consumption data and the real-time energy consumption data, a load adjustment signal is generated, and a fuzzy logic controller is used for adjustment. The output signal of the fuzzy logic controller is: ΔP=f(ΔE); Among them, ΔP is the load adjustment signal, and ΔE is the energy consumption difference; S84, adjusting the energy supply of the building and the load of the energy storage device according to the regulation signal; S85, monitor the energy consumption changes of buildings in real time and adjust the scheduling strategy based on real-time feedback; S86. When sudden load fluctuations occur, adjust the load dispatching strategy.

Images