Low-carbon electric heating optimization regulation and control method

Abstract

The invention discloses a low-carbon type electric heating optimization regulation and control method. The method comprises the following steps that firstly, multi-source data are collected and preprocessed; step 2, dynamic carbon quota distribution; step 3, optimizing the power of the carbon flow tracking drive; 4, building thermal inertia compensation control; 5, learning a user behavior mode; and 6, performing multi-mode cooperative control. According to the method, through multi-source data fusion and dynamic carbon quota distribution, carbon emission is accurately quantified and reasonably distributed, and the optimization level is improved; by means of carbon flow tracking and building thermal inertia compensation, accurate power regulation and control are achieved, control delay is eliminated, and energy loss is reduced; and meanwhile, a personalized heating curve is generated through user behavior mode learning, the comfort level and carbon emission are balanced, strategies are flexibly switched through multi-mode cooperative control, different scene requirements are met, the carbon emission is effectively reduced through the overall scheme, and the user experience and the system adaptability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building energy management, and in particular to a low-carbon electric heating optimization control method. BACKGROUND

[0002] The traditional electric heating system controls the start and stop of the heating equipment through a single temperature threshold, which has problems such as insufficient carbon emission optimization, control lag and energy waste caused by building thermal inertia, and increased hardware costs caused by complex algorithms.

[0003] The prior art only adjusts the power by presetting the carbon intensity threshold, without considering the real-time carbon flow dynamic characteristics of the power grid and the thermal characteristics of the building envelope, resulting in a disconnection between the control strategy and the actual carbon emission target, and lacking an adaptive learning mechanism for user behavior patterns. Therefore, a low-carbon electric heating optimization control method is proposed. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a low-carbon electric heating optimization control method to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a low-carbon electric heating optimization control method, comprising the following steps: Step 1, multi-source data acquisition and preprocessing: Real-time collection of environmental temperature, humidity, light intensity and user activity trajectory data by deploying Internet of Things sensor networks in the building, combined with minute-level carbon intensity data released by the regional power grid dispatching center, to build a multi-dimensional data set; Cleaning and normalizing the original data to generate a standardized input vector; Step 2, dynamic carbon quota allocation: Based on the building historical energy consumption data and the total amount of regional power grid carbon quota, the sliding window method is used to calculate the carbon budget per unit time; Establish a heat load prediction model combined with building thermal characteristic parameters to dynamically allocate the upper limit of the instantaneous carbon quota of each heating unit; Step 3, carbon flow tracking driven power optimization: According to the real-time carbon intensity data, a carbon flow tracking matrix is constructed to analyze the distribution of carbon emission flow on the power grid side; An improved greedy algorithm is used to control the on-off of the solid state relay array, and the heating circuit during the low carbon intensity period is preferentially started; Step 4, building thermal inertia compensation control: Establish a building envelope heat conduction model to predict the future indoor temperature trend; Design a feedforward compensator to correct the power output and eliminate the control delay caused by thermal inertia; Step five, user behavior pattern learning: Identify user's daily routine through unsupervised learning algorithm, generate personalized heating demand curve; Adjust the preset temperature threshold dynamically combined with carbon quota constraint, balance comfort and carbon emission; Step six, multi-mode collaborative control: Define three working scenarios: economic mode, comfort mode and emergency mode, automatically switch control strategy according to carbon market price signal and building thermal safety threshold; In emergency mode, enable silicon controlled voltage regulating circuit for fine power regulation; In step one, the Internet of Things sensor network includes infrared human body sensing module and door and window opening and closing state monitoring unit, through Kalman filtering algorithm to fuse multiple source data, generate corrected user activity heat map, combined with regional power grid carbon intensity data, build multi-dimensional data set; in step two, the sliding window length is dynamically adjusted according to the regional power grid day-ahead forecast accuracy, the carbon budget allocation in the window uses entropy weight method, and the building envelope thermal performance attenuation factor is introduced to weight and correct the historical energy consumption data; The dynamic carbon quota allocation mechanism ensures the rational use of carbon resources and improves energy efficiency. The power optimization strategy driven by carbon flow tracking effectively reduces carbon emissions while ensuring heating effect. Building thermal inertia compensation control reduces temperature fluctuations, improves residential comfort, user behavior pattern learning makes the heating system more personalized to meet the needs of different users, and multi-mode collaborative control ensures the flexibility and safety of the system in different scenarios, realizing the perfect combination of economic, comfortable and emergency functions.

[0006] Preferably, the Internet of Things sensor network in step one includes infrared human body sensing module and door and window opening and closing state monitoring unit for correcting user activity trajectory data; the infrared human body sensing module is installed at 0.8-1.2 meters below the room ceiling, with a coverage angle not less than 120°, and at least one is arranged in each room; the door and window opening and closing state monitoring unit is installed on the inner edge of the door and window frame, with one for each door and window; the sensor uses RS485 bus or ZigBee wireless protocol wiring, wired wiring needs to be protected by metal bellows, and the distance between wireless nodes is not more than 30 meters; The correction process uses Kalman filtering algorithm to fuse multiple source sensor data to generate a corrected user activity heat map; In step one, the infrared human body sensing module and the door and window opening and closing state monitoring unit in the Internet of Things sensor network transmit data to the central processing unit through a wireless communication protocol. The central processing unit uses Kalman filtering algorithm, combines the time stamp and data accuracy of each sensor, and performs weighted fusion on multi-source data. Through iterative updating of state estimation value and covariance matrix, a high-precision user activity heat map is generated, which is used for subsequent optimization of control strategy. The combination of the infrared human body sensing module and the door and window state monitoring unit can more comprehensively capture user behavior and environmental changes. The application of Kalman filtering algorithm effectively reduces data noise and error, generating a high-precision user activity heat map. This process not only provides a solid data foundation for subsequent dynamic carbon quota allocation and power optimization, but also realizes personalized regulation of the heating system through accurate user behavior prediction, significantly improving user experience and energy utilization efficiency, and promoting the achievement of low-carbon and environmental protection goals.

[0007] Preferably, the sliding window length in step two is dynamically adjusted according to the regional power grid day-ahead prediction accuracy, and the carbon budget allocation in the window uses the entropy weight method to determine the weight coefficient of each heating unit. The entropy weight method introduces a building envelope thermal performance decay factor to weight and correct historical energy consumption data. The solid state relay array selection needs to meet the requirements of rated voltage ≥ AC220V and rated current ≥10A, and be installed in a separate compartment of the distribution box with a cooling distance ≥50mm. The power level of the silicon-controlled voltage regulating circuit needs to match 1.2-1.5 times the total power of the heating circuit, and a radiator and overcurrent protection device need to be configured during installation, with a grounding resistance ≤4Ω. For dynamic adjustment of the sliding window length, an evaluation model based on regional power grid day-ahead prediction accuracy can be established. When the accuracy is higher than the threshold, the window is shortened, otherwise it is lengthened. When determining the weight using the entropy weight method, the entropy value of each heating unit's historical energy consumption is calculated first, then the energy consumption data is weighted combined with the thermal performance decay factor, and finally the weight coefficient of each unit is calculated through the entropy weight formula to ensure more reasonable carbon budget allocation. By dynamically adjusting the sliding window length, the system can flexibly adapt to changes in power grid prediction accuracy, improving the timeliness and accuracy of carbon budget allocation. The entropy weight method combined with the building envelope thermal performance decay factor weights and corrects historical energy consumption data, fully considering the influence of building characteristics on energy consumption, making carbon budget allocation more scientific and reasonable. This not only helps to accurately control carbon emissions, but also optimizes energy utilization efficiency while ensuring heating demand, achieving a balance between low carbon and comfort, improving user experience, and having important significance for promoting the collaborative development of green buildings and smart grids.

[0008] Preferably, the improved greedy algorithm in step three introduces a carbon intensity change rate factor. When a sharp drop in carbon intensity is detected and the predicted duration exceeds T minutes, the heating circuit is started early. The T-minute threshold is dynamically set according to the building thermal inertia time constant; In step three, the improved greedy algorithm monitors the grid carbon intensity data in real time, analyzes the carbon intensity change rate using a sliding window, and automatically starts the heating loop in advance when a sharp drop in carbon intensity is detected and the predicted trend continues for more than T minutes. The T-minute threshold is dynamically determined by the building thermal inertia time constant, which is calculated by multiplying the building envelope thermal capacity parameter and the real-time temperature difference. The minimum common multiple of the thermal inertia time constant and the carbon intensity fluctuation period is taken as the T-minute threshold, ensuring the accuracy of the early start of the heating loop. The improved greedy algorithm captures and utilizes the low-carbon valley period of the grid by introducing the carbon intensity change rate factor. Starting the heating loop in advance can make full use of the low-carbon valley period for heating, reducing electricity consumption during the high-carbon period, and thus reducing carbon emissions. The dynamic setting of the T-minute threshold takes into account the building thermal inertia, avoiding temperature fluctuations caused by thermal inertia and ensuring the stability of indoor temperature. This intelligent control method not only improves the energy efficiency of the heating system, but also achieves a good balance between comfort and carbon emissions, in line with the development concept of low-carbon environmental protection.

[0009] Preferably, the heat conduction model in step four considers solar radiation heat gain and indoor equipment heat dissipation. Model parameters are identified online using the least squares method combined with the genetic algorithm, and the iteration step is adaptively adjusted according to the rate of change of environmental temperature and humidity. In step four, the heat conduction model is implemented as follows: First, establish a heat conduction equation that includes solar radiation heat gain and indoor equipment heat dissipation. Then, use the least squares method to preliminarily estimate the model parameters, and then use the genetic algorithm to globally optimize the search for parameters. During the identification process, the iteration step is dynamically adjusted according to the real-time rate of change of environmental temperature and humidity, ensuring that the model parameters can quickly and accurately converge to the optimal solution. The heat conduction model that considers solar radiation heat gain and indoor equipment heat dissipation, combined with the online identification method of the least squares method and the genetic algorithm, significantly improves the accuracy and adaptability of the model. The least squares method provides a preliminary estimate of the model parameters, while the genetic algorithm further optimizes the parameters through global search capability, ensuring that the model maintains high precision under various environmental conditions. In addition, the iteration step is adaptively adjusted according to the rate of change of environmental temperature and humidity, allowing the model to quickly respond to environmental changes and achieve more accurate thermal inertia compensation control, effectively improving the stability and energy efficiency of the system.

[0010] Preferably, the unsupervised learning algorithm in step five uses an improved K-means++ clustering algorithm. The number of cluster centers is adaptively adjusted according to seasonal changes and user behavior pattern migration metrics, which are calculated using the dynamic time warping algorithm. For the unsupervised learning algorithm, the improved K-means++ clustering is implemented. First, the benchmark value of the number of cluster centers is initialized according to seasonal changes. Then, the dynamic time warping algorithm is used to calculate the similarity of user behavior patterns in different time periods, and the migration measure value is obtained. According to the comparison of the measure value and the preset threshold, the number of cluster centers is dynamically adjusted to adapt to the seasonal changes and long-term migration of user behavior patterns. The improved K-means++ clustering algorithm is used, and the number of cluster centers is adaptively adjusted according to the seasonal changes and the user behavior pattern migration measure value. This method significantly improves the accuracy and flexibility of the heating demand curve identification. By calculating the migration measure value through the dynamic time warping algorithm, the subtle changes in user behavior patterns can be accurately captured, making the clustering results more consistent with the actual use. This adaptive adjustment mechanism ensures that the heating system can provide personalized comfort experience in different seasons and user behavior changes, while optimizing energy utilization and achieving an effective balance between comfort and carbon emissions.

[0011] Preferably, the economic mode in step six sets two-level carbon quota warning thresholds. When the carbon budget remaining amount is lower than the first threshold, basic heating is started, and when it is lower than the second threshold, unnecessary heating circuits are turned off. The two-level carbon quota warning thresholds are dynamically corrected according to user comfort feedback data. During the operation of the electric heating system, user feedback data on indoor temperature comfort, such as "too cold", "moderate" and "too hot", are collected through the user terminal. The system analyzes these feedbacks using natural language processing technology, combines the current carbon budget remaining amount and historical usage data, and dynamically adjusts the first and second carbon quota warning thresholds through machine learning algorithms to optimize the heating strategy and ensure energy saving while meeting user comfort requirements. The economic mode sets two-level carbon quota warning thresholds and dynamically corrects them. This provides a flexible and user-friendly energy-saving heating solution. When the carbon budget is close to depletion, the system first starts basic heating to ensure basic warmth while avoiding excessive energy consumption. As the carbon budget further decreases, the system automatically turns off unnecessary heating circuits to prioritize core area heating, effectively extending the heating time. More importantly, by dynamically adjusting the warning thresholds based on user comfort feedback, the system can continuously optimize the heating strategy to achieve a balance between energy saving and comfort, improve user experience, and promote the popularization of low-carbon life.

[0012] Preferably, the emergency mode in step six sets a minimum heating safety temperature, and when the predicted indoor temperature is lower than the set minimum heating safety temperature, all heating circuits are forcibly enabled and the building insulation layer electric heating wire is triggered to assist in temperature rise; the building insulation layer electric heating wire uses nickel-chromium alloy material, the power density is ≤15 W / m², is laid in the inside of the insulation layer, the interval is 10-15 cm, needs to be matched with a temperature fuse (action temperature 70±5 ℃) and a leakage protection device; In the emergency mode, the system will monitor the indoor temperature in real time and compare it with the preset minimum heating safety temperature. Once the predicted indoor temperature is lower than the threshold, the system will automatically trigger the emergency response mechanism, forcibly enable all available heating circuits, and activate the electric heating wire in the building insulation layer to quickly raise the indoor temperature, ensuring the safety and comfort of the indoor environment; The setting of the emergency mode significantly improves the safety and reliability of the system. In extreme weather conditions or sudden situations that cause a sharp drop in indoor temperature, this mode can quickly respond by forcibly enabling all heating circuits and insulation layer electric heating wires, effectively preventing the indoor temperature from being too low and protecting the safety of people and equipment in the building. This proactive defense mechanism not only reflects the thoroughness of the system design, but also enhances users' trust and satisfaction with the electric heating system.

[0013] Preferably, the thermal inertia time constant is calculated by multiplying the building envelope thermal capacity parameter and the real-time temperature difference, and the T-minute threshold is set as the least common multiple of the thermal inertia time constant and the carbon intensity fluctuation period; The building envelope thermal capacity parameter is obtained through laboratory determination, the real-time temperature difference is calculated from the environmental temperature data collected by the Internet of Things sensor network in real time, and the carbon intensity fluctuation period is determined by analyzing historical carbon intensity data using Fourier transform to determine its main periodic component; By combining the building envelope thermal capacity parameter and the real-time temperature difference to calculate the thermal inertia time constant, and setting the T-minute threshold as the least common multiple of the thermal inertia time constant and the carbon intensity fluctuation period, the starting time of the heating circuit is precisely controlled. This design fully considers the influence of building thermal inertia on indoor temperature changes and the influence of carbon intensity fluctuations on heating efficiency, thereby ensuring indoor temperature comfort while effectively reducing carbon emissions and improving the energy efficiency and environmental performance of the electric heating system.

[0014] Preferably, the feedforward compensator uses a fuzzy PID control algorithm, and the proportional coefficient of the fuzzy PID control algorithm is corrected in real time according to the product of the indoor and outdoor temperature difference change rate and the building envelope thermal conductivity; The fuzzy PID control algorithm in the feedforward compensator first acquires the indoor-outdoor temperature difference change rate and the building envelope thermal conductivity through a sensor, multiplies the two to obtain an input variable of the fuzzy PID, performs fuzzy reasoning according to a preset fuzzy rule base to determine a proportional coefficient correction amount, and finally applies the corrected proportional coefficient to the PID control to realize accurate correction of the power output. The fuzzy PID control algorithm in the feedforward compensator has the advantage that the proportional coefficient can be corrected in real time according to the product of the indoor-outdoor temperature difference change rate and the building envelope thermal conductivity. This dynamic adjustment mechanism enables the system to more accurately adapt to the heat conduction characteristics under different environmental conditions, effectively eliminates the control delay caused by thermal inertia, and through real-time adjustment of the control parameters, the system can respond more quickly and accurately to environmental changes, ensure that the indoor temperature is stable within the set range, improve user comfort, and at the same time realize more efficient energy utilization and reduce unnecessary energy consumption.

[0015] In summary, compared with the prior art, the present application provides a low-carbon electric heating optimization control method, which has the following advantages: The present application realizes accurate quantification and reasonable allocation of carbon emissions through multi-source data acquisition and preprocessing and dynamic carbon quota allocation, which has the advantage of improving the optimization level of carbon emissions. The method uses the Internet of Things sensor network to collect various environmental data in the building and regional power grid minute-level carbon intensity data in real time, constructs a multi-dimensional data set and generates a standardized input vector, providing a comprehensive and accurate data basis for subsequent analysis. At the same time, based on the historical energy consumption data of the building and the total amount of regional power grid carbon quota, the sliding window method is used to calculate the carbon budget per unit time, and a heat load prediction model is established based on the building thermal characteristic parameters to dynamically allocate the upper limit of the instantaneous carbon quota of each heating unit, making the carbon emission allocation more scientific and reasonable, effectively avoiding the problem of insufficient carbon emission optimization. Through carbon flow tracking driven power optimization and building thermal inertia compensation control, accurate power regulation and elimination of control delay are realized, which has the advantage of reducing energy waste. The method constructs a carbon flow tracking matrix according to real-time carbon intensity data, analyzes the carbon emission flow distribution on the power grid side, and uses an improved greedy algorithm to control the on-off of the solid-state relay array, preferentially enabling the heating circuit during the low-carbon intensity period to reduce carbon emissions. In addition, a building envelope heat conduction model is established to predict the future indoor temperature change trend, and a feedforward compensator is designed to correct the power output, eliminating the control delay caused by thermal inertia, improving the response speed and regulation accuracy of the system, and reducing energy waste. Through user behavior pattern learning and multi-mode collaborative control, the balance of comfort and carbon emission and the flexible switching of control strategies are realized, which has the benefits of improving user experience and adaptability, the method identifies user work and rest rules through an unsupervised learning algorithm, generates a personalized heating demand curve, and dynamically adjusts the preset temperature threshold in combination with the carbon quota constraint, while meeting the user comfort demand, reduces carbon emission, meanwhile, three working scenes of economic mode, comfort mode and emergency mode are defined, the control strategy is automatically switched according to the carbon market price signal and the building thermal safety threshold, and in the emergency mode, a silicon controlled regulator circuit is also started to carry out fine power regulation, further improving the adaptability and reliability of the system. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is the step diagram of the low-carbon type electric heating optimization control method. DETAILED DESCRIPTION

[0017] The application provides a technical solution, a low-carbon type electric heating optimization control method, please refer to Figure 1 , comprising the following steps: Step one, multi-source data acquisition and preprocessing: Real-time collection of environmental temperature, humidity, light intensity and user activity trajectory data through the Internet of Things sensor network deployed in the building, combined with the minute-level carbon intensity data released by the regional power grid dispatching center, a multi-dimensional data set is constructed; The original data is cleaned and normalized to generate a standardized input vector; Step two, dynamic carbon quota allocation: Based on the building historical energy consumption data and the total amount of regional power grid carbon quota, the sliding window method is used to calculate the carbon budget per unit time; Combined with the building thermal characteristic parameters, a heat load prediction model is established to dynamically allocate the upper limit of the instantaneous carbon quota of each heating unit; Step three, carbon flow tracking driven power optimization: According to the real-time carbon intensity data, a carbon flow tracking matrix is constructed to analyze the distribution of carbon emission flow on the power grid side; An improved greedy algorithm is used to control the on-off of the solid state relay array, and the heating circuit in the low carbon intensity period is preferentially started; Step four, building thermal inertia compensation control: A building envelope heat conduction model is established to predict the future indoor temperature change trend; A feedforward compensator is designed to correct the power output and eliminate the control delay caused by thermal inertia; Step five, user behavior pattern learning: An unsupervised learning algorithm is used to identify user work and rest rules, and generate a personalized heating demand curve; The preset temperature threshold is dynamically adjusted in combination with the carbon quota constraint to balance comfort and carbon emissions. Step six, multi-mode cooperative control: Three working scenarios, namely, economic mode, comfort mode and emergency mode, are defined, and the control strategy is automatically switched according to the carbon market price signal and the building thermal safety threshold. In the emergency mode, the silicon-controlled voltage regulating circuit is enabled for fine power regulation. In step one, the Internet of Things sensor network includes an infrared human body sensing module and a door and window opening and closing state monitoring unit. The multi-source data is fused and processed by the Kalman filtering algorithm to generate a corrected user activity heat map. In combination with the regional power grid carbon intensity data, a multi-dimensional data set is constructed. In step two, the sliding window length is dynamically adjusted according to the regional power grid day-ahead prediction accuracy. The carbon budget allocation in the window uses the entropy weight method, and the building envelope thermal performance attenuation factor is introduced to weight and correct the historical energy consumption data. The dynamic carbon quota allocation mechanism ensures the rational use of carbon resources and improves energy efficiency. The power optimization strategy driven by carbon flow tracking effectively reduces carbon emissions while ensuring heating effect. The building thermal inertia compensation control reduces temperature fluctuations and improves residential comfort. User behavior pattern learning makes the heating system more personalized to meet the needs of different users. Multi-mode cooperative control ensures the flexibility and safety of the system in different scenarios, achieving a perfect combination of economic, comfortable and emergency functions. Hardware integration: The central processing unit and the sensor network are connected through an industrial-grade switch (supporting Modbus protocol). The wiring needs to distinguish between strong current (AC 220V) and weak current (DC 24V) with a spacing of ≥30 cm. The heating loop cable cross-sectional area is selected according to the rated current of 1.5 times, and is laid through galvanized steel pipes.

[0018] Software debugging: The thermal conduction model parameters are calibrated offline through a special debugging tool. The step response method is used to test the system delay time to ensure that the response time of the feedforward compensator is ≤5 minutes. The multi-mode switching logic needs to be tested at least 3 times through simulation of carbon market price fluctuations and temperature sudden change scenarios.

[0019] Please refer to Figure 1 , the Internet of Things sensor network in step one includes an infrared human body sensing module and a door and window opening and closing state monitoring unit for correcting user activity trajectory data. The infrared human body sensing module is installed at 0.8-1.2 meters below the room ceiling with a coverage angle of not less than 120°, and at least one is arranged in each room. The door and window opening and closing state monitoring unit is installed on the inside edge of the door and window frame, with one for each door and window. The sensors are wired with RS485 bus or ZigBee wireless protocol. The wired wiring needs to be protected by metal bellows, and the distance between wireless nodes should not exceed 30 meters. The correction process adopts Kalman filtering algorithm to fuse the multi-source sensor data and generate the corrected user activity heat map. In step one, the infrared human body sensing module and the door and window opening and closing state monitoring unit in the Internet of Things sensor network transmit data to the central processing unit through the wireless communication protocol. The central processing unit adopts Kalman filtering algorithm, combines the time stamp and data accuracy of each sensor, and performs weighted fusion on the multi-source data. Through iterative updating of state estimation value and covariance matrix, a high-precision user activity heat map is generated, which is used for subsequent optimization of control strategy. The combination of the infrared human body sensing module and the door and window state monitoring unit can more comprehensively capture user behavior and environmental changes. The application of Kalman filtering algorithm effectively reduces data noise and error, generating a high-precision user activity heat map. This process not only provides a solid data foundation for subsequent dynamic carbon quota allocation and power optimization, but also realizes personalized regulation of the heating system through accurate user behavior prediction, significantly improving user experience and energy utilization efficiency, and promoting the achievement of low-carbon and environmental protection goals.

[0020] Please refer to Figure 1 , the sliding window length in step two is dynamically adjusted according to the regional power grid day-ahead prediction accuracy. The carbon budget allocation in the window uses the entropy weight method to determine the weight coefficient of each heating unit. The entropy weight method introduces the building envelope thermal performance decay factor to weight and correct the historical energy consumption data. The solid state relay array selection needs to meet the requirements of rated voltage ≥ AC220V and rated current ≥10A. It is installed in the independent compartment of the distribution box with a cooling distance of ≥50mm. The power level of the silicon controlled rectifier voltage regulating circuit needs to match 1.2-1.5 times of the total power of the heating circuit. When installing, a radiator and an overcurrent protection device need to be configured, and the grounding resistance needs to be ≤4Ω. For dynamic adjustment of the sliding window length, an evaluation model based on the regional power grid day-ahead prediction accuracy can be established. When the accuracy is higher than the threshold, the window is shortened, otherwise it is lengthened. When determining the weight using the entropy weight method, the entropy value of the historical energy consumption of each heating unit is calculated first. Then, the energy consumption data is weighted by combining the thermal performance decay factor. Finally, the weight coefficient of each unit is calculated through the entropy weight formula to ensure that the carbon budget allocation is more reasonable. By dynamically adjusting the sliding window length, the system can flexibly adapt to changes in power grid prediction accuracy, improving the timeliness and accuracy of carbon budget allocation. The entropy weight method combines the building envelope thermal performance decay factor to weight and correct the historical energy consumption data, fully considering the influence of building characteristics on energy consumption, making the carbon budget allocation more scientific and reasonable. This not only helps to accurately control carbon emissions, but also optimizes energy utilization efficiency while ensuring heating demand, achieving a balance between low carbon and comfort, improving user experience, and having important significance for promoting the collaborative development of green buildings and smart grids.

[0021] Referring to Figure 1 , the improved greedy algorithm in step three introduces a carbon intensity change rate factor. When a steep drop in carbon intensity is detected and the predicted duration exceeds T minutes, the heating circuit is activated in advance. The T-minute threshold is dynamically set according to the building thermal inertia time constant. In step three, the improved greedy algorithm monitors real-time grid carbon intensity data, analyzes the carbon intensity change rate using a sliding window, and automatically activates the heating circuit in advance when a steep drop in carbon intensity is detected and the predicted trend lasts more than T minutes. The T-minute threshold is dynamically determined by the building thermal inertia time constant, which is calculated by multiplying the building envelope thermal capacity parameter and the real-time temperature difference. The minimum common multiple of the thermal inertia time constant and the carbon intensity fluctuation period is taken as the T-minute threshold, ensuring the accuracy of the early activation of the heating circuit. The improved greedy algorithm captures and utilizes the low-carbon valley period of the grid by introducing a carbon intensity change rate factor. Early activation of the heating circuit can make full use of the low-carbon valley period for heating, reducing electricity consumption during high-carbon periods, thereby reducing carbon emissions. The dynamic setting of the T-minute threshold takes into account the building thermal inertia, avoiding temperature fluctuations caused by thermal inertia and ensuring the stability of indoor temperature. This intelligent control method not only improves the energy efficiency of the heating system but also achieves a good balance between comfort and carbon emissions, in line with the development concept of low-carbon environmental protection.

[0022] Referring to Figure 1 , the heat conduction model in step four considers solar radiation heat gain and indoor equipment heat dissipation. Model parameters are identified online using the least squares method combined with the genetic algorithm, and the iteration step is adaptively adjusted according to the environmental temperature and humidity change rate. In step four, the heat conduction model is implemented as follows: First, establish a heat conduction equation that includes solar radiation heat gain and indoor equipment heat dissipation. Then, use the least squares method to preliminarily estimate the model parameters, and then use the genetic algorithm to perform global optimization search on the parameters. During the identification process, the iteration step is dynamically adjusted according to the real-time change rate of environmental temperature and humidity, ensuring that the model parameters can quickly and accurately converge to the optimal solution. The heat conduction model that considers solar radiation heat gain and indoor equipment heat dissipation, combined with the online identification method of the least squares method and the genetic algorithm, significantly improves the accuracy and adaptability of the model. The least squares method provides a preliminary estimate of the model parameters, while the genetic algorithm further optimizes the parameters through global search capability, ensuring that the model maintains high precision under various environmental conditions. In addition, the iteration step is adaptively adjusted according to the environmental temperature and humidity change rate, allowing the model to quickly respond to environmental changes and achieve more accurate thermal inertia compensation control, effectively improving the stability and energy efficiency of the system.

[0023] Referring to Figure 1, the unsupervised learning algorithm in step five employs an improved K-means++ clustering, with the number of cluster centers self-adaptively adjusted according to seasonal changes and user behavior pattern migration metrics calculated by dynamic time warping algorithm; For the specific implementation of the unsupervised learning algorithm using improved K-means++ clustering, first, initialize the benchmark value of the number of cluster centers according to seasonal changes, then calculate the similarity of user behavior patterns in different time periods using dynamic time warping algorithm to obtain migration metrics, and dynamically adjust the number of cluster centers according to the comparison of the metrics with the preset threshold to adapt to the seasonal changes and long-term migration of user behavior patterns; By using the improved K-means++ clustering algorithm and self-adaptively adjusting the number of cluster centers according to seasonal changes and user behavior pattern migration metrics, this method significantly improves the accuracy and flexibility of heating demand curve identification. By calculating the migration metrics using dynamic time warping algorithm, subtle changes in user behavior patterns can be accurately captured, making the clustering results more in line with actual usage. This adaptive adjustment mechanism ensures that the heating system can provide personalized comfort experience in different seasons and user behavior changes, while optimizing energy utilization and achieving an effective balance between comfort and carbon emissions.

[0024] Please refer to Figure 1 , the economic mode sets two-level carbon quota warning thresholds, and when the carbon budget remaining amount is below the first threshold, basic heating is started, and when it is below the second threshold, unnecessary heating circuits are turned off; The two-level carbon quota warning thresholds are dynamically corrected according to user comfort feedback data; During the operation of the electric heating system, user feedback data on indoor temperature comfort, such as "too cold", "moderate" and "too hot", are collected through the user terminal. The system analyzes these feedbacks using natural language processing technology, combines the current carbon budget remaining amount and historical usage data, and dynamically adjusts the first and second carbon quota warning thresholds through machine learning algorithms to optimize the heating strategy and ensure energy saving while meeting user comfort requirements; The benefits of setting two-level carbon quota warning thresholds and dynamically correcting them in the economic mode are that it provides a flexible and user-friendly energy-saving heating solution. When the carbon budget is close to depletion, the system first starts basic heating to ensure basic warmth while avoiding excessive energy consumption. As the carbon budget further decreases, the system automatically turns off unnecessary heating circuits to prioritize heating in core areas, effectively extending the heating time. More importantly, by dynamically adjusting the warning thresholds based on user comfort feedback, the system can continuously optimize the heating strategy to achieve a balance between energy saving and comfort, improve user experience, and promote the popularization of low-carbon life.

[0025] Please refer to Figure 1The emergency mode in step six sets a minimum heating safety temperature. When the predicted indoor temperature is lower than the set minimum heating safety temperature, all heating circuits are forcibly enabled and the building insulation layer electric heating wire is triggered to assist in temperature rise. The building insulation layer electric heating wire uses nickel-chromium alloy material with a power density of ≤15 W / m², is laid on the inside of the insulation layer with a spacing of 10-15 cm, and needs to be matched with a temperature fuse (action temperature 70±5℃) and a leakage protection device. In the emergency mode, the system will monitor the indoor temperature in real time and compare it with the preset minimum heating safety temperature. Once the predicted indoor temperature is lower than the threshold, the system will automatically trigger the emergency response mechanism, forcibly enable all available heating circuits, and simultaneously activate the electric heating wire in the building insulation layer to quickly raise the indoor temperature, ensuring the safety and comfort of the indoor environment. The setting of the emergency mode significantly improves the safety and reliability of the system. In extreme weather conditions or sudden situations that cause a sharp drop in indoor temperature, this mode can quickly respond by forcibly enabling all heating circuits and insulation layer electric heating wires, effectively preventing the indoor temperature from being too low and protecting the safety of people and equipment in the building. This proactive defense mechanism not only reflects the thoroughness of the system design, but also enhances users' trust and satisfaction with the electric heating system.

[0026] Please refer to Figure 1 The thermal inertia time constant is calculated by multiplying the building envelope thermal capacity parameter and the real-time temperature difference. The T-minute threshold is set as the least common multiple of the thermal inertia time constant and the carbon intensity fluctuation period. The building envelope thermal capacity parameter is obtained through laboratory measurement, the real-time temperature difference is calculated from the environmental temperature data collected by the Internet of Things sensor network in real time, and the carbon intensity fluctuation period is determined by analyzing historical carbon intensity data using Fourier transform. By combining the building envelope thermal capacity parameter and the real-time temperature difference to calculate the thermal inertia time constant, and setting the T-minute threshold as the least common multiple of the thermal inertia time constant and the carbon intensity fluctuation period, the starting time of the heating circuit is precisely controlled. This design fully considers the influence of building thermal inertia on indoor temperature changes and the influence of carbon intensity fluctuations on heating efficiency, thereby ensuring indoor temperature comfort while effectively reducing carbon emissions and improving the energy efficiency and environmental performance of the electric heating system.

[0027] Please refer to Figure 1 The feedforward compensator uses a fuzzy PID control algorithm, and the proportional coefficient of the fuzzy PID control algorithm is modified in real time according to the product of the indoor-outdoor temperature difference rate and the building envelope thermal conductivity coefficient. The fuzzy PID control algorithm in the feedforward compensator first obtains the indoor-outdoor temperature difference change rate and the building envelope thermal conductivity through the sensor, multiplies the two to obtain the input variable of the fuzzy PID, performs fuzzy reasoning according to the preset fuzzy rule base to determine the proportional coefficient correction amount, and finally applies the corrected proportional coefficient to the PID control to realize accurate correction of the power output. The fuzzy PID control algorithm in the feedforward compensator has the advantage of being able to correct the proportional coefficient in real time according to the product of the indoor-outdoor temperature difference change rate and the building envelope thermal conductivity. This dynamic adjustment mechanism enables the system to more accurately adapt to the heat conduction characteristics under different environmental conditions, effectively eliminates the control delay caused by thermal inertia, and through real-time adjustment of the control parameters, the system can respond more quickly and accurately to environmental changes, ensure that the indoor temperature is stable within the set range, improve user comfort, and at the same time realize more efficient energy utilization and reduce unnecessary energy consumption.

[0028] It should be noted that in this document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0029] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, alternatives, and variations can be made in the embodiments without departing from the principles and spirits of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A low-carbon electric heating optimization control method, characterized in that, The method comprises the following steps: Step one, multi-source data collection and preprocessing: Real-time collection of environmental temperature, humidity, light intensity and user activity trajectory data through the Internet of Things sensor network deployed in the building, combined with the minute-level carbon intensity data released by the regional power grid dispatching center, to construct a multi-dimensional data set; Cleaning and normalization processing of the original data to generate standardized input vectors; Step two, dynamic carbon quota allocation: Based on the building historical energy consumption data and the total carbon quota of the regional power grid, the sliding window method is used to calculate the carbon budget per unit time; Combined with the building thermal characteristic parameters, a heat load prediction model is established to dynamically allocate the upper limit of the instantaneous carbon quota of each heating unit; Step three, power optimization driven by carbon flow tracking: According to the real-time carbon intensity data, a carbon flow tracking matrix is constructed to analyze the distribution of carbon emission flow on the power grid side; An improved greedy algorithm is used to control the on-off of the solid-state relay array, and the heating circuit in the low carbon intensity period is preferentially started; Step four, building thermal inertia compensation control: Establish a building envelope heat conduction model to predict the future indoor temperature trend; Design a feedforward compensator to correct the power output and eliminate the control delay caused by thermal inertia; Step five, user behavior pattern learning: Through unsupervised learning algorithm, the user's work and rest rules are identified, and the personalized heating demand curve is generated; Combined with the carbon quota constraint, the preset temperature threshold is dynamically adjusted to balance the comfort and carbon emission; Step six, multi-mode collaborative control: Define three working scenarios: economic mode, comfort mode and emergency mode, and automatically switch the control strategy according to the carbon market price signal and building thermal safety threshold; In the emergency mode, the silicon controlled rectifier circuit is used for fine power regulation.

2. The low-carbon electric heating optimization method according to claim 1, characterized in that: The Internet of Things sensor network in step one includes an infrared human body sensing module and a door and window opening and closing state monitoring unit for correcting user activity trajectory data; The infrared human body sensing module is installed at 0.8-1.2 meters below the room ceiling, with a coverage angle not less than 120°, and at least one is arranged in each room; the door and window opening and closing state monitoring unit is installed on the inner edge of the door and window frame, with one for each door and window; the sensors are wired with RS485 bus or ZigBee wireless protocol, the wired wiring needs to be protected by metal bellows, and the distance between wireless nodes is not more than 30 meters; The correction process uses Kalman filtering algorithm to fuse and process multi-source sensor data to generate a corrected user activity heat map.

3. The low-carbon electric heating optimization method according to claim 1, characterized in that: The sliding window length in step two is dynamically adjusted according to the accuracy of the regional power grid day-ahead forecast, and the carbon budget allocation in the window uses the entropy weight method to determine the weight coefficient of each heating unit; The entropy weight method introduces a building envelope thermal performance attenuation factor to weight and correct the historical energy consumption data; the solid-state relay array selection needs to meet the requirements of rated voltage ≥ AC220V and rated current ≥10A, and is installed in the independent compartment of the distribution box with a heat dissipation distance ≥50mm; the power level of the silicon controlled rectifier circuit needs to match 1.2-1.5 times of the total power of the heating circuit, and a radiator and overcurrent protection device need to be configured during installation, with a grounding resistance ≤4Ω.

4. The low-carbon electric heating optimization method according to claim 1, characterized in that: The improved greedy algorithm in step three introduces a carbon intensity change rate factor. When a sharp carbon intensity drop is detected and the predicted duration exceeds T minutes, the heating circuit is activated in advance. The T-minute threshold is dynamically set according to the building thermal inertia time constant.

5. The low-carbon electric heating optimization method according to claim 1, characterized in that: The heat conduction model in step four considers solar radiation heat gain and indoor equipment heat dissipation. Model parameters are identified online through the least squares method combined with the genetic algorithm, and the iteration step is adaptively adjusted according to the ambient temperature and humidity change rate.

6. The low-carbon electric heating optimization method according to claim 1, wherein: The unsupervised learning algorithm in step five uses an improved K-means++ clustering algorithm. The number of cluster centers is adaptively adjusted according to seasonal changes and user behavior pattern migration metrics, which are calculated using the dynamic time warping algorithm.

7. The low-carbon electric heating optimization method according to claim 1, wherein: The economic mode in step six sets two-level carbon quota warning thresholds. When the carbon budget remaining amount is below the first threshold, basic heating is started, and when it is below the second threshold, unnecessary heating circuits are turned off. The two-level carbon quota warning thresholds are dynamically corrected based on user comfort feedback data.

8. The low-carbon electric heating optimization method according to claim 1, characterized in that: The emergency mode in step six sets a minimum heating safety temperature. When the predicted indoor temperature is below the set minimum heating safety temperature, all heating circuits are forced to be activated, and the building insulation layer electric heating wire is triggered to assist in temperature rise. The building insulation layer electric heating wire uses nickel-chromium alloy material with a power density of ≤15 W / m². It is laid on the inside of the insulation layer with a spacing of 10-15 cm and needs to be equipped with a temperature fuse and a leakage protection device.

9. The low-carbon electric heating optimization method according to claim 4, characterized in that: The thermal inertia time constant is calculated by multiplying the building envelope heat capacity parameter by the real-time temperature difference. The T-minute threshold is set as the least common multiple of the thermal inertia time constant and the carbon intensity fluctuation period.

10. The low-carbon electric heating optimization method according to claim 1, characterized in that: The feedforward compensator uses a fuzzy PID control algorithm. The proportional coefficient of the fuzzy PID control algorithm is real-time corrected according to the product of the indoor-outdoor temperature difference change rate and the building envelope thermal conductivity.

Images