Self-adaptive air conditioner heating control method and system based on multi-source perception

Abstract

The embodiment of the invention discloses a self-adaptive air conditioner heating control method and system based on multi-source perception.The method comprises the steps that multi-source environment data such as indoor environment data, outdoor environment data and air conditioner operation data are collected in real time, and active adjustment behaviors and environment state data of a user are combined; a personalized comfort model is established and continuously updated through a machine learning algorithm, and a dynamic target comfort temperature interval is output to reflect comfort temperature preferences of a user in different environments; the multi-source environment data and the target comfortable temperature interval serve as input, maximization of user comfort, minimization of energy consumption and reduction of compressor starting and stopping times serve as targets, rolling optimization calculation is conducted through an optimization algorithm, and an air conditioner control instruction set containing a dynamic target temperature set value, a compressor operation instruction and a fan rotating speed instruction is dynamically generated; and executing an instruction set, monitoring an effect and user feedback, and feeding data back to the model establishment and optimization calculation step to realize self-adaptive control.

Description

TECHNICAL FIELD

[0001] The embodiment of the present application relates to the air conditioner control technical field, and particularly relates to a multi-source sensing based adaptive air conditioner heating control method and system. BACKGROUND

[0002] In the air conditioner heating control technical field, the traditional air conditioner heating mode has been developed for a long time. The early air conditioner heating is mainly based on simple temperature set point control, and the user sets a target room temperature, and the air conditioner performs heating operation according to the set value. With the development of technology, some air conditioners increase the preset operation program, such as setting different temperatures according to different time periods, but the overall still relies on a single control logic. In the running process of the traditional air conditioner, the indoor temperature information is mainly obtained by a single room temperature sensor, and the overall sensing of the human actual cold and hot feeling, indoor humidity, outdoor humidity, outdoor temperature and other environmental factors is lacked. When facing the indoor and outdoor environment changes, such as cold wind invasion caused by opening doors and windows, heat changes caused by sunlight, personnel increase and decrease and the like, the response ability of the traditional air conditioner is poor, and the heating strategy often cannot be adjusted in time. In order to ensure that the indoor temperature can reach the set value, the traditional air conditioner often uses the "overshoot" heating method, that is, excessive heating, which undoubtedly causes the waste of energy. At the same time, in order to maintain the set temperature, the compressor is frequently started and stopped, which not only increases the energy consumption, but also accelerates the wear and tear of the equipment. Moreover, the traditional air conditioner cannot learn and adapt to the temperature preferences and work and rest habits of different users, and cannot provide personalized heating experience for the user.

[0003] It can be seen that the traditional air conditioner heating mode has problems of insufficient comfort, response lag, low energy efficiency and lack of individualization, and is difficult to meet the high-quality demand of modern users for air conditioner heating. SUMMARY

[0004] The embodiment of the present application provides a multi-source sensing based adaptive air conditioner heating control method and system.

[0005] In a first aspect, the embodiment of the present application provides a multi-source sensing based adaptive air conditioner heating control method, applied to a multi-source sensing based adaptive air conditioner heating control system, and the method comprises the following steps.

[0006] Real-time collection of multi-source environmental data, wherein the multi-source environmental data comprises indoor environmental data, outdoor environmental data and air conditioner self-operation data;

[0007] Based on the user active adjustment behavior and environmental state data in the multi-source environmental data, an individualized comfort model is established and continuously updated by a machine learning algorithm, and the individualized comfort model outputs a dynamic target comfort temperature interval, which is used to reflect the comfort temperature preference of the user in different environmental states;

[0008] The multi-source environment data and the target comfortable temperature interval output by the personalized comfort model are taken as inputs to maximize user comfort, minimize energy consumption and reduce the number of compressor start-stop times, and a set of air conditioner control instructions is dynamically generated through an optimization algorithm for rolling optimization calculation, the set of air conditioner control instructions at least including a dynamic target temperature setting value, a compressor operation instruction and a fan speed instruction, and the dynamic target temperature setting value is dynamically adjusted within the target comfortable temperature interval.

[0009] The set of air conditioner control instructions is executed, and the execution effect and user feedback are continuously monitored, and the monitored execution effect data and user feedback data are fed back as new input data to the establishing and updating step and the optimization calculation step of the personalized comfort model.

[0010] In a second aspect, an embodiment of the present application provides a multi-source sensing based adaptive air conditioner heating control system, comprising:

[0011] a processor;

[0012] a storage device having a computer program stored thereon,

[0013] When the computer program is executed by the processor, the processor implements the multi-source sensing based adaptive air conditioner heating control method.

[0014] An embodiment of the present application provides a readable storage medium having a program or instructions stored thereon, and the program or instructions are executed by a processor to implement the steps of the multi-source sensing based adaptive air conditioner heating control method.

[0015] Therefore, the embodiment of the present application has the following beneficial effects: first, the indoor environment data, outdoor environment data and air conditioner running data are collected in real time, various factors affecting the heating effect of the air conditioner can be comprehensively obtained, and based on the user active adjustment behavior and environment state data in the multi-source environment data, a personalized comfort model is established and continuously updated through a machine learning algorithm, the dynamic target comfortable temperature interval output by the model can accurately reflect the comfortable temperature preference of the user in different environment states, the problem of lack of personalization of the traditional air conditioner is solved, and the comfort experience of the user is improved.

[0016] Secondly, the multi-source environmental data and the target comfortable temperature interval output by the personalized comfort model are taken as inputs to maximize user comfort, minimize energy consumption and reduce the number of compressor start-stop times, and a rolling optimization calculation is performed through an optimization algorithm to dynamically generate an air conditioner control instruction set containing a dynamic target temperature setting value, a compressor operation instruction and a fan speed instruction, which can dynamically adjust the air conditioner operation parameters according to real-time environment and user demand, avoid the problems of excessive heating and frequent start-stop of the traditional air conditioner, improve the energy utilization efficiency and reduce the equipment wear.

[0017] Then, the air conditioner control instruction set is executed and the execution effect and user feedback are continuously monitored, and the monitoring data is fed back to the establishment and update step and the optimization calculation step of the personalized comfort model as new inputs to form an adaptive control system, which can timely respond to changes in indoor and outdoor environments and user preferences, ensure that the air conditioner always operates in the best state, and realize dynamic, accurate, comfortable and energy-saving heating control, thereby improving the overall performance of the air conditioner heating control. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 A flowchart of a multi-source sensing adaptive air conditioner heating control method provided by an embodiment of the present application.

[0019] Figure 2 A schematic diagram of the basic structure of a multi-source sensing adaptive air conditioner heating control system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned objects, features and advantages of the present application more apparent and easy to understand, the embodiments of the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0021] Referring to Figure 1 The figure is a flowchart of a multi-source sensing adaptive air conditioner heating control method provided by an embodiment of the present application, and the method can be applied to a multi-source sensing adaptive air conditioner heating control system. As shown in Figure 1 The method can include steps 110-140.

[0022] Step 110: Real-time acquisition of multi-source environmental data, the multi-source environmental data including indoor environmental data, outdoor environmental data and air conditioner self-operation data.

[0023] This embodiment takes the air conditioning heating control scene in the living room as an example. In order to achieve accurate air conditioning heating control, it is necessary to comprehensively and accurately collect multi-source environmental data. First, temperature sensors and humidity sensors are arranged at different positions in the room to collect indoor environmental data. In the human activity area, such as around the sofa, near the dining table and other positions, sensors are set up. These positions can directly reflect the actual temperature and humidity of the environment where the user is located, and the user spends most of the time in this area. The temperature and humidity of this area have a greater impact on user comfort. At the same time, sensors are also installed near the air conditioner, which can obtain local environmental information around the air conditioner, so as to analyze the influence of the change of the surrounding environment on the performance of the air conditioner when the air conditioner is running.

[0024] Outdoor environmental data is obtained through networking, including outdoor temperature, humidity, wind speed, weather conditions and other information. Outdoor environment has a significant impact on indoor temperature. For example, in cold weather, low outdoor temperature will accelerate the loss of indoor heat, and the air conditioner needs to provide more heating to maintain indoor temperature. The air conditioner's own running data is collected by using the built-in sensors of the air conditioner, including running mode, set temperature, compressor state, fan speed, outlet temperature, actual power consumption, etc. These data can accurately represent the current running state and performance of the air conditioner.

[0025] After collecting multi-source environmental data, time stamp alignment processing is performed. The data collection time of different sensors may differ, and time stamp alignment can ensure the consistency of the data in time, which is convenient for subsequent analysis and processing. Then, abnormal value detection and elimination is performed. Due to sensor failure, external interference and other factors, the collected data may have abnormal values. Eliminating these abnormal values can improve data quality. Finally, the processed target environmental data is classified and stored according to indoor environmental data, outdoor environmental data and air conditioner's own running data, forming a structured multi-source environmental data set.

[0026] Step 120: Based on the user's active adjustment behavior and environmental state data in the multi-source environmental data, an individualized comfort model is established and continuously updated through a machine learning algorithm. The individualized comfort model outputs a dynamic target comfort temperature interval, which is used to reflect the user's comfort temperature preference under different environmental conditions.

[0027] After collecting multi-source environmental data, the user's active adjustment behavior and environmental state data are used to construct and update the individualized comfort model. In the initial stage of air conditioner startup, a baseline comfort temperature range is provided as the initial parameter of the model. This range is determined by considering human physiological characteristics and energy saving requirements. When the user first turns on the air conditioner, the system defaults to this range as the comfort temperature reference.

[0028] The user's active adjustment behavior to temperature in different environment states, such as adjusting the set temperature up or down, turning on or off the air conditioner, etc. In the living room scene, the user will make different adjustments at different times according to his own feelings. He may adjust the temperature up during the day and adjust the temperature down at night. The environment state data corresponding to these active adjustment behaviors is input into the machine learning algorithm as training samples.

[0029] Step 121: In the initial stage after the air conditioner is turned on, the reference comfort temperature range is provided as the initial parameter of the personalized comfort model.

[0030] At the beginning of the air conditioner being turned on, the initial parameter of the personalized comfort model is set according to the general comfort temperature standard. The reference comfort temperature range is obtained through a large number of experiments and researches, which can meet the basic comfort needs of most users to a certain extent, providing a starting point for the subsequent learning and adjustment of the model, so that the model can be optimized based on the actual behavior of the user and the change of the environment.

[0031] Step 122: Record the user's active adjustment behavior to temperature in different environment states, including adjusting the set temperature up or down and turning on or off the air conditioner.

[0032] In the daily use of the living room, the user will actively adjust the temperature of the air conditioner according to his own cold and hot feelings, changes in indoor and outdoor environment and other factors. For example, when the number of people in the room increases and the user feels hot, he may adjust the set temperature down. When the outdoor temperature drops suddenly and the indoor feels cold, the user may adjust the set temperature up. Turning on or off the air conditioner is also an active adjustment behavior. When the user leaves the room for a period of time, he may choose to turn off the air conditioner to save energy. When he returns to the room, he turns on the air conditioner again. These behaviors reflect the user's preference for comfortable temperature in different environment states.

[0033] Step 123: The environment state data corresponding to the user's active adjustment behavior is input into the machine learning algorithm as training samples.

[0034] The recorded user's active adjustment behavior is associated with the environment state data at that time to form training samples. The environment state data includes indoor and outdoor temperature, humidity, weather conditions, time and other factors. These training samples are the basis for the machine learning algorithm to learn the user's comfort temperature preference. Through analysis and learning of a large number of training samples, the algorithm can find the user's adjustment rules in different environment conditions.

[0035] Step 124: Through the machine learning algorithm, the user's individual comfort temperature preference in different environment states is analyzed to establish a personalized comfort model.

[0036] Step 1241: Feature extraction is performed on the environmental state data in the training samples to obtain a plurality of environmental feature variables, including an indoor-outdoor temperature difference, an indoor humidity value, an outdoor weather type code, and a time period identifier.

[0037] In-depth analysis is performed on the environmental state data in the training samples to extract key environmental feature variables. The indoor-outdoor temperature difference reflects the degree of heat exchange between the indoor and outdoor environments, which has an important influence on the heating demand of the air conditioner. The indoor humidity value affects the thermal sensation of the human body, and the perception of comfortable temperature by the human body is different under different humidity. The outdoor weather type code classifies different weather conditions, such as sunny, cloudy, rainy, etc., and the outdoor environment has different influences on the indoor temperature under different weather types. The time period identifier takes into account the activity patterns and temperature demand differences of users at different times of the day, for example, the comfortable temperature preferences of users may be different during the day and at night.

[0038] Step 1242: The set temperature after the user's active adjustment behavior is taken as the target variable, and a mapping relationship between the environmental feature variables and the target variable is constructed.

[0039] The set temperature after the user's active adjustment behavior is taken as the target variable, and it is linked with the extracted environmental feature variables. By analyzing a large number of training samples, the potential law between the environmental feature variables and the target variable is found, and a mapping relationship is constructed, which can describe the user's expected comfortable temperature setting value under different combinations of environmental features.

[0040] Step 1243: Reinforcement learning algorithm or collaborative filtering algorithm is used to train the mapping relationship, and the user's temperature adjustment behavior is taken as the feedback signal to adjust the model parameters.

[0041] Optionally, a suitable machine learning algorithm is selected to train the constructed mapping relationship. Reinforcement learning algorithm gives rewards or punishments according to the user's temperature adjustment behavior through continuous trial and error and feedback, so as to adjust the parameters of the model, so that the model can better predict the user's comfortable temperature preference. Collaborative filtering algorithm optimizes the model parameters of the current user by analyzing the similar behaviors and preferences of other users. The user's temperature adjustment behavior is taken as the feedback signal to update the model in real time, so that the model can adapt to the dynamic changes of user preferences.

[0042] Step 1244: When a new user's active adjustment behavior occurs, the similarity between the new environmental state data corresponding to the new user's active adjustment behavior and the historical training samples is judged. If the similarity is higher than a preset threshold, the target comfortable temperature interval is predicted based on the mapping relationship in the historical data, and if the similarity is lower than the preset threshold, the new environmental state data is taken as a supplementary training sample, and the model parameters are optimized through iterative updating.

[0043] When new user-initiated adjustment behavior occurs, the corresponding new environmental state data is compared with the historical training samples. The similarity between the two is calculated. If the similarity is high, it means that the current environment is similar to the historical situation, and the target comfort temperature interval can be directly predicted based on the mapping relationship in the historical data. If the similarity is low, it means that a new environmental situation has occurred or the user's preference has changed significantly. At this time, the new environmental state data is used as a supplementary training sample to retrain the model. Through iterative updating, the model parameters are optimized so that the model can adapt to new situations.

[0044] Step 125: During the operation of the air conditioner, new user-initiated adjustment behavior and environmental state data are continuously received, and the parameters of the personalized comfort model are updated through cyclic iteration, so that the output target comfort temperature interval is used to continuously adapt to changes in user preferences.

[0045] During the continuous operation of the air conditioner, new user-initiated adjustment behavior and environmental state data are continuously collected. These new data are added to the training samples, and the personalized comfort model is trained and optimized again. Through cyclic iteration, the parameters of the model are continuously updated, so that the target comfort temperature interval output by the model can timely reflect the dynamic adjustment of user preferences over time and environmental changes.

[0046] Step 130: The multi-source environmental data and the target comfort temperature interval output by the personalized comfort model are used as inputs to maximize user comfort, minimize energy consumption, and reduce the number of compressor start-stop times as optimization objectives. Through optimization algorithms, rolling optimization calculations are performed to dynamically generate air conditioner control instruction sets, including dynamic target temperature set values, compressor operation instructions, and fan speed instructions. The dynamic target temperature set value is dynamically adjusted within the target comfort temperature interval.

[0047] The collected multi-source environmental data and the target comfort temperature interval output by the personalized comfort model are used as inputs for multi-objective optimization decision-making. First, the real-time collected multi-source environmental data is preprocessed to extract key information, such as the difference between the current indoor temperature and the target comfort temperature interval, the temperature rise and fall rate, the influence coefficient of outdoor environmental parameters on indoor temperature, and the current running parameters of the air conditioner.

[0048] The optimization objective function is defined, and the three objectives of user comfort, energy consumption, and compressor start-stop times are quantified. The user comfort objective is measured by the proportion of time that the temperature is maintained within the target interval and the temperature fluctuation amplitude. The longer the temperature is maintained within the target interval and the smaller the fluctuation amplitude, the higher the user's comfort. The energy consumption objective is quantified by the actual power consumption per unit time. Reducing power consumption can achieve the purpose of energy saving. The compressor start-stop objective is quantified by the number of start-stops per unit time. Reducing the number of start-stops can prolong the service life of the compressor and reduce energy consumption.

[0049] Dynamic weight coefficients are set for the optimization objectives, which are adjusted according to real-time environmental conditions and user usage scenarios. For example, when the user is active during the day, more attention is paid to user comfort, and the weight coefficient of the user comfort objective can be appropriately increased. At night when the user is resting, more attention may be paid to energy consumption and compressor start-stop times, and the weight coefficients are adjusted accordingly.

[0050] Step 131: Preprocess the real-time collected multi-source environmental data, extract the difference between the current indoor temperature and the target comfortable temperature interval, the temperature rise and fall rate, the influence coefficient of outdoor environmental parameters on indoor temperature, and the current running parameters of the air conditioner.

[0051] The collected multi-source environmental data is cleaned and organized to remove noise and outliers. Then, key information is extracted, the difference between the current indoor temperature and the target comfortable temperature interval reflects the temperature amplitude that needs to be adjusted by the air conditioner. The temperature rise and fall rate can help determine whether the heating or cooling capacity of the air conditioner meets the demand. The influence coefficient of outdoor environmental parameters on indoor temperature considers the effect of outdoor environment on indoor temperature, such as outdoor low temperature which can accelerate the loss of indoor heat and require the air conditioner to provide more heating capacity. The current running parameters of the air conditioner include compressor frequency, fan speed, etc.

[0052] Step 132: Define the optimization objective function, the user comfort objective is quantified by the proportion of time that the temperature is maintained within the target interval and the temperature fluctuation amplitude, the energy consumption objective is quantified by the actual power consumption per unit time, and the compressor start-stop objective is quantified by the number of start-stops per unit time.

[0053] An integrated optimization objective function is constructed, which includes user comfort, energy consumption, and compressor start-stop frequency. For the user comfort objective, the proportion of time that the temperature is maintained within the target interval is calculated, and the temperature fluctuation amplitude is calculated. The higher the proportion of time that the temperature is maintained within the target interval, the better the air conditioner can meet the user's comfort needs; the smaller the temperature fluctuation amplitude, the more stable the user's thermal sensation. The energy consumption objective is quantified by measuring the actual power consumption of the air conditioner per unit time, and reducing power consumption can reduce energy consumption and use cost. The compressor start-stop objective is quantified by counting the number of compressor starts and stops per unit time, and frequent starts and stops increase the wear and tear and energy consumption of the compressor, and reducing the number of starts and stops can improve the reliability and energy-saving effect of the compressor.

[0054] Step 133: Set dynamic weight coefficients for optimization objectives, which are adjusted according to real-time environmental conditions and user usage scenarios.

[0055] According to the real-time environmental conditions and user usage scenarios, different weight coefficients are assigned to each optimization objective. In different situations, users may have different priorities for each objective. For example, in some scenarios, users may pay more attention to comfort, so the weight coefficient of the user comfort objective can be increased; in energy- constrained or energy-saving scenarios, the weight coefficient of the energy consumption objective can be appropriately increased. By dynamically adjusting the weight coefficients, the optimization objective function can better adapt to different actual situations.

[0056] Step 134: Use model predictive control algorithm to solve multi-objective optimization problem in finite time domain, according to the difference between current indoor temperature and target temperature and temperature rise and fall rate, find the matching compressor frequency and fan speed through step-by-step approximation algorithm.

[0057] Step 1341: Establish a building thermal characteristic model, which is used to predict the indoor temperature variation trend in the subsequent period under the specified air conditioning output condition.

[0058] A building thermal characteristic model is constructed, which takes into account the structure, material, thermal insulation performance of the building, and the influence of outdoor environment on indoor temperature. Through this model, the indoor temperature variation trend in the subsequent period under different air conditioning output conditions can be predicted, providing a basis for finding the optimal compressor frequency and fan speed, and understanding the indoor temperature variation under different control parameters in advance.

[0059] Step 1342: Based on the building thermal characteristic model, predict the time and energy consumption for indoor temperature to reach the target comfort temperature interval under different compressor frequency and fan speed combinations, and substitute the prediction results into the optimization objective function to calculate the comprehensive objective value of different combinations.

[0060] The building thermal characteristic model is used to simulate and predict different compressor frequency and fan speed combinations. The time required for the indoor temperature to reach the target comfort temperature interval and the corresponding energy consumption are predicted for each combination. These prediction results are substituted into the optimization objective function to calculate the comprehensive target value for different combinations. The comprehensive target value considers multiple factors such as user comfort, energy consumption, and compressor start-stop frequency, allowing a comprehensive evaluation of the pros and cons of each combination.

[0061] Step 1343: Iteratively adjust the compressor frequency and fan speed using the step-by-step approximation algorithm. After each adjustment, re-predict and calculate the comprehensive target value until the combination with the optimal comprehensive target value is found.

[0062] The step-by-step approximation algorithm is used to iteratively optimize the compressor frequency and fan speed. Starting from an initial combination, the compressor frequency and fan speed are adjusted step by step based on the current comprehensive target value. After each adjustment, the building thermal characteristic model is used for prediction, and a new comprehensive target value is calculated. This process is repeated until the combination with the optimal comprehensive target value is found. This process is a continuous trial and optimization process that gradually approaches the optimal solution, improving the control effect of the air conditioner.

[0063] Step 1344: During the iteration process, determine whether the current predicted indoor temperature fluctuation amplitude is within the preset range. If it exceeds the preset range, adjust the fan speed to improve air distribution and adjust the compressor frequency to control the refrigeration capacity.

[0064] During the iterative adjustment process, the current predicted indoor temperature fluctuation amplitude is monitored in real time. If the fluctuation amplitude exceeds the preset range, it indicates that the indoor temperature stability is poor, which may affect user comfort. At this time, the fan speed is adjusted to improve indoor air distribution, making the indoor temperature distribution more uniform. At the same time, the compressor frequency is adjusted to control the refrigeration capacity, reducing temperature fluctuations and ensuring that the indoor temperature can be stabilized within the target comfort temperature interval.

[0065] Step 135: Generate a control instruction set containing dynamic target temperature set value, compressor operation mode, and fan speed, with the dynamic target temperature set value dynamically adjusted within the target comfort temperature interval according to the optimization results.

[0066] According to the optimal combination obtained by optimization calculation, the air conditioner control instruction set is generated, which includes dynamic target temperature set value, compressor operation mode, and fan speed. The dynamic target temperature set value will be dynamically adjusted within the target comfort temperature interval output by the personalized comfort model according to the optimization results to ensure that the indoor temperature is always within the user's comfortable range. The compressor operation mode and fan speed will also be set according to the optimization results to make the air conditioner run in the optimal state, achieving the goal of maximizing user comfort, minimizing energy consumption, and reducing compressor start-stop frequency.

[0067] Step 140: Execute the air conditioner control instruction set and continuously monitor the execution effect and user feedback, feed the monitored execution effect data and user feedback data back to the personalized comfort model establishment and update step and optimization calculation step as new input data.

[0068] The generated air conditioner control instruction set is sent to the air conditioner execution mechanism, such as the compressor, fan motor, etc., so that the air conditioner runs according to the instructions. At the same time, the indoor temperature change, actual energy consumption data and compressor start-stop times are monitored in real time as execution effect data. In the living room scene, the temperature sensor installed in the room is used to obtain the change of indoor temperature in real time, and the power meter is used to record the actual power consumption of the air conditioner, and the running state of the compressor is monitored to count the start-stop times.

[0069] The user's active adjustment behavior to the temperature is received as user feedback data. Users will adjust the air conditioner temperature according to their actual feelings, and these adjustment behaviors reflect the user's satisfaction and comfort demand for the current temperature. The execution effect data is analyzed to determine whether the current indoor temperature is maintained within the target comfortable temperature range, whether the energy consumption is within the expected range, and whether the compressor start-stop times meet the optimization target.

[0070] Step 141: Send the air conditioner control instruction set to the air conditioner execution mechanism, and monitor the indoor temperature change, actual energy consumption data and compressor start-stop times in real time as execution effect data.

[0071] The generated air conditioner control instruction set is sent to the air conditioner execution mechanism through the communication interface, so that the compressor, fan and other components run according to the instructions. At the same time, the real-time monitoring system is started to continuously monitor the indoor temperature change, actual energy consumption data and compressor start-stop times. The change of indoor temperature reflects the heating effect of the air conditioner, the actual energy consumption data reflects the energy consumption of the air conditioner, and the compressor start-stop times are related to the service life and energy saving effect of the equipment.

[0072] Step 142: Receive the user's active adjustment behavior to the temperature as user feedback data.

[0073] During the operation of the air conditioner, users will actively adjust the air conditioner temperature according to their own cold and hot feelings. User adjustment instructions can be received through air conditioner remote controllers, mobile phone APPs, etc., and these instructions are used as user feedback data, which can directly reflect the user's satisfaction and comfort demand for the current temperature, and are important inputs for the update and optimization of the personalized comfort model.

[0074] Step 143: Analyze the execution effect data to determine whether the current indoor temperature is maintained within the target comfortable temperature range, whether the energy consumption is within the expected range, and whether the compressor start-stop times meet the optimization target.

[0075] Step 1431: Set the allowed threshold range of indoor temperature fluctuations, calculate the ratio of actual temperature fluctuation amplitude to the allowed threshold range, and if the ratio is greater than a preset value, determine that the temperature control effect is not up to standard.

[0076] An allowed threshold range is set for indoor temperature fluctuations, which is determined according to the user's comfort requirements and the performance characteristics of the air conditioner. Calculate the ratio of actual temperature fluctuation amplitude to the allowed threshold range, and if the ratio is greater than a preset value, it means that the actual temperature fluctuation exceeds the acceptable range, and the temperature control effect is not ideal. At this time, the control parameters of the air conditioner need to be adjusted to improve the stability of temperature control.

[0077] Step 1432: Calculate the difference between actual energy consumption and theoretical minimum energy consumption per unit time, and if the difference is greater than a preset energy consumption deviation threshold, determine that the energy consumption control effect is not up to standard.

[0078] By measuring the actual energy consumption of the air conditioner per unit time, and comparing it with the theoretical minimum energy consumption. Calculate the difference between the two, and if the difference is greater than the preset energy consumption deviation threshold, it means that the energy consumption exceeds the expected range, and the energy consumption control effect is not good. It may be necessary to check the running state of the air conditioner and optimize the control strategy to reduce energy consumption.

[0079] Step 1433: Count the number of compressor starts and stops per unit time, and compare it with the preset start-stop number threshold, and if it exceeds the threshold, determine that the compressor protection effect is not up to standard.

[0080] Count the number of compressor starts and stops per unit time, and compare it with the preset start-stop number threshold. If the start-stop number exceeds the threshold, it means that the compressor is frequently started and stopped, which may affect its service life and increase energy consumption. At this time, the control strategy of the air conditioner needs to be adjusted to reduce the number of compressor starts and stops, and improve the protection effect of the compressor.

[0081] Step 1434: Comprehensive evaluation of the compliance of temperature, energy consumption and compressor start-stop number, generate evaluation results, including the compliance degree of each index and the main deviation direction.

[0082] Comprehensively consider the compliance of temperature control, energy consumption control and compressor protection, and conduct a comprehensive evaluation of the execution effect. The generated evaluation results include the compliance degree of each index, such as temperature control compliance rate, energy consumption control compliance rate, compressor protection compliance rate, etc., and the main deviation direction.

[0083] Step 144: Compare the execution effect data and user feedback data with historical data to identify the type and cause of deviation. If the deviation is caused by changes in user preferences, feedback the data back to the personalized comfort model establishment and update step. If the deviation is caused by environmental changes or equipment characteristic changes, feedback the data back to the optimization calculation step.

[0084] Compare the current execution effect data and user feedback data with historical data to identify the type and cause of deviation. If the deviation is due to changes in user preferences, such as recent changes in user sensitivity to temperature, feedback these data to the personalized comfort model establishment and update step to adjust and optimize the model, so that the model can better reflect the user's current comfort needs. If the deviation is caused by environmental changes, such as sudden drop in outdoor temperature, or equipment characteristic changes, such as compressor performance decline, feedback the data to the optimization calculation step to adjust the parameters of the optimization algorithm and generate more suitable air conditioning control instruction set.

[0085] Step 145: Continuously update the parameters of the personalized comfort model and the input parameters of the optimization algorithm using feedback data through iterative loops.

[0086] Through continuous iterative loops, feedback data is used to update the parameters of the personalized comfort model and the input parameters of the optimization algorithm. Each iteration adjusts and optimizes the model and algorithm based on new data, improving the control effect of the air conditioner. Over time, the model and algorithm can better adapt to changes in user preferences and environmental dynamics, achieving continuous optimization of air conditioning heating control.

[0087] In a non-limiting embodiment, the method further comprises:

[0088] Step 150: Perform correlation analysis on the execution effect data and user feedback data to extract a comfort deviation feature set related to user comfort perception, including amplitude and duration characteristics of actual temperature fluctuations deviating from the target comfort temperature interval.

[0089] To further optimize the personalized comfort model and air conditioning control strategy, perform correlation analysis on the execution effect data and user feedback data. In the living room scenario, collect indoor temperature change data, user temperature adjustment behavior, and air conditioner running parameters, etc. execution effect data and user feedback data within a period of time, and conduct in-depth mining.

[0090] Extract comfort deviation feature set related to user comfort perception, focusing on the amplitude feature and duration feature of the actual temperature fluctuation deviating from the target comfort temperature interval. If the actual temperature is consistently higher or lower than the target comfort temperature interval for a period of time, and the deviation amplitude is large and the duration is long, the user will feel uncomfortable. Through the extraction and analysis of these features, the user's comfort perception in different environments can be more accurately understood.

[0091] Step 151: Based on the comfort deviation feature set, extract the user preference drift feature vector using the feature decomposition method, which is used to represent the dynamic adjustment trend of user comfort preference with environmental state changes.

[0092] Based on the extracted comfort deviation feature set, use the feature decomposition method to extract the user preference drift feature vector. With the passage of time, seasonal changes and changes in indoor and outdoor environments, the user's preference for comfortable temperature may drift. The feature decomposition method can quantitatively represent this drift trend, and by analyzing the changes in the feature vector, the dynamic adjustment trend of user comfort preference can be understood.

[0093] Step 152: Compare the user preference drift feature vector with the current environmental feature variable weight distribution of the personalized comfort model to generate a preference drift correction coefficient matrix.

[0094] Compare the user preference drift feature vector with the current environmental feature variable weight distribution of the personalized comfort model. The environmental feature variable weight distribution reflects the model's judgment weight for the comfortable temperature under different environmental characteristics. By comparing, find out the differences between the two, generate a preference drift correction coefficient matrix, which can reflect which environmental feature variable weights need to be adjusted according to the user's preference drift.

[0095] Step 153: Use the preference drift correction coefficient matrix to dynamically adjust the environmental feature variable weight distribution of the personalized comfort model, so that the output of the target comfort temperature interval reflects the adjustment trend represented by the user preference drift feature vector.

[0096] According to the generated preference drift correction coefficient matrix, dynamically adjust the environmental feature variable weight distribution of the personalized comfort model. By adjusting the weight distribution, the model output of the target comfort temperature interval can better reflect the adjustment trend represented by the user preference drift feature vector, so that the model can adapt to the changes in user comfort preference in time and provide target comfort temperature intervals that better meet user needs.

[0097] Step 154: Take the adjusted target comfort temperature interval as the current output of the personalized comfort model, which is used as the input for the next round of optimization calculation step.

[0098] The adjusted target comfort temperature interval is taken as the current output of the personalized comfort model and is input into the next round of optimization calculation steps. During the optimization calculation process, a new set of air conditioner control instructions is generated in combination with multi-source environmental data and the adjusted target comfort temperature interval. In this way, the control of the air conditioner can better meet the changing comfort needs of the user, and further optimization of the heating control of the air conditioner is achieved.

[0099] As another non-limiting embodiment, the method further comprises:

[0100] Step 160: Feature fusion is performed on the execution effect data and the user feedback data to extract a system dynamic response feature parameter set reflecting the dynamic response characteristics of the air conditioning system, the system dynamic response feature parameter set including a temperature regulation rate deviation feature, an energy consumption fluctuation feature, and a compressor start-stop frequency feature.

[0101] In order to comprehensively evaluate the performance and control effect of the air conditioning system, feature fusion is performed on the execution effect data and the user feedback data. In the living room scenario, the execution effect data such as indoor temperature change data, actual energy consumption data, and compressor start-stop frequency are comprehensively processed with the feedback data such as user temperature adjustment.

[0102] The system dynamic response feature parameter set reflecting the dynamic response characteristics of the air conditioning system is extracted, including a temperature regulation rate deviation feature, an energy consumption fluctuation feature, and a compressor start-stop frequency feature. The temperature regulation rate deviation feature reflects the difference between the actual temperature regulation rate of the air conditioner and the expected rate, the energy consumption fluctuation feature reflects the stability of the energy consumption of the air conditioner, and the compressor start-stop frequency feature is related to the operating state and service life of the compressor.

[0103] Step 161: Comparing the system dynamic response feature parameter set with a preset ideal operating state parameter set to generate a control efficiency evaluation index vector, the control efficiency evaluation index vector including a temperature control accuracy index, an energy consumption control index, and a device protection index.

[0104] The extracted system dynamic response feature parameter set is compared and analyzed with a preset ideal operating state parameter set. The ideal operating state parameter set is set according to the design performance of the air conditioner and the expectations of the user. Through the comparison, a control efficiency evaluation index vector is generated, which includes a temperature control accuracy index, an energy consumption control index, and a device protection index. The temperature control accuracy index reflects the control accuracy of the air conditioner on the indoor temperature, the energy consumption control index measures the energy saving effect of the air conditioner, and the device protection index evaluates the protection degree of the compressor and other devices.

[0105] Step 162: Based on the control performance evaluation index vector, dynamically correct the multi-objective weight coefficient configuration matrix of the optimization algorithm, which is used to adjust the weight proportion of maximizing user comfort, minimizing energy consumption and reducing compressor start-stop times in rolling optimization calculation.

[0106] According to the generated control performance evaluation index vector, dynamically correct the multi-objective weight coefficient configuration matrix of the optimization algorithm, which determines the weight proportion of maximizing user comfort, minimizing energy consumption and reducing compressor start-stop times in rolling optimization calculation. If the temperature control precision index is low, it means that the importance of temperature control is not enough, and the weight coefficient of the user comfort target can be appropriately increased; if the energy consumption control index is not ideal, the weight coefficient of the energy consumption target is increased. By dynamically adjusting the weight coefficient, the optimization algorithm can make more reasonable decisions according to the actual operation of the air conditioning system.

[0107] Step 163: Input the corrected multi-objective weight coefficient configuration matrix into the optimization calculation step to adjust the generation logic of dynamic target temperature setting value, compressor operation instruction and fan speed instruction, so that the output of air conditioning control instruction set matches the current system dynamic response characteristic parameter set.

[0108] The corrected multi-objective weight coefficient configuration matrix is input into the optimization calculation step to adjust the generation logic of dynamic target temperature setting value, compressor operation instruction and fan speed instruction. According to the new weight coefficient, the optimization algorithm will recalculate the optimal control parameters and generate a new air conditioning control instruction set that better matches the current system dynamic response characteristic parameter set. In this way, the control of the air conditioner can better adapt to the actual operating state of the system and improve the control effect.

[0109] Step 164: Use the adjusted optimization algorithm parameters as the initial configuration parameters for the next round of rolling optimization calculation to form a cycle control optimization.

[0110] The adjusted optimization algorithm parameters are used as the initial configuration parameters for the next round of rolling optimization calculation, and through continuous cycle iteration, the control strategy of the air conditioner is continuously optimized. Each iteration adjusts the algorithm parameters according to the new system dynamic response characteristics and user feedback data, so that the air conditioning system can always maintain a good operating state and achieve continuous optimization of the air conditioning heating control.

[0111] It can be seen that the application firstly collects indoor environment data, outdoor environment data and air conditioner self-operation data in real time, can comprehensively acquire various factors influencing the heating effect of the air conditioner, based on the user active adjustment behavior and environment state data in the multi-source environment data, establishes and continuously updates the personalized comfort degree model through the machine learning algorithm, and the dynamic target comfort temperature interval output by the model can accurately reflect the comfort temperature preference of the user under different environment states, solves the problem of lack of personalization of the traditional air conditioner, and improves the comfort experience of the user.

[0112] Secondly, the multi-source environment data and the target comfort temperature interval output by the personalized comfort degree model are taken as inputs, the optimization algorithm is used for rolling optimization calculation with the optimization objectives of maximizing the user comfort degree, minimizing the energy consumption and reducing the number of compressor start-stop, and the air conditioner control instruction set containing the dynamic target temperature setting value, the compressor operation instruction and the fan speed instruction is dynamically generated, the air conditioner operation parameters can be dynamically adjusted according to the real-time environment and the user demand, the problems of excessive heating and frequent start-stop of the traditional air conditioner are avoided, the energy utilization efficiency is improved, and the equipment wear is reduced.

[0113] Then, the air conditioner control instruction set is executed and the execution effect and the user feedback are continuously monitored, the monitoring data is fed back to the establishment and update steps of the personalized comfort degree model and the optimization calculation step as new inputs, and the adaptive control system is formed, the system can timely respond to the changes of the indoor and outdoor environments and the change of the user preference, ensures that the air conditioner always operates in the best state, and realizes the dynamic, accurate, comfortable and energy-saving heating control, and the performance of the air conditioner heating control is improved as a whole.

[0114] Referring to Figure 2 The basic structure of the multi-source perception adaptive air conditioner heating control system 200 provided by the embodiment of the application is shown in the figure, and the multi-source perception adaptive air conditioner heating control system 200 comprises:

[0115] a processor 201;

[0116] a storage device 202, wherein the storage device 202 stores a computer program 2020.

[0117] When the computer program 2020 is executed by the processor 201, the processor 201 realizes the multi-source perception adaptive air conditioner heating control method.

[0118] On the basis, a readable storage medium is provided, the readable storage medium stores a program or instructions, and the program or instructions are executed by the processor to realize the steps of the above method.

[0119] It should be noted that the various embodiments described in the specification are intended to be exemplary only and that the scope of the application is not intended to be limited to the embodiments described in the specification.

Claims

1. A multi-source perception based adaptive air conditioning and heating control method, characterized in that, The application relates to a method for dynamically generating air conditioner control instructions based on user comfort temperature preferences. The method comprises the following steps: Real-time acquisition of multi-source environment data, including indoor environment data, outdoor environment data and air conditioner self-operation data; Based on user active adjustment behavior and environment state data in the multi-source environment data, a personalized comfort temperature model is established and continuously updated through a machine learning algorithm, and the personalized comfort temperature model outputs a dynamic target comfort temperature interval, which is used to reflect user comfort temperature preferences under different environment states; The multi-source environment data and the target comfort temperature interval output by the personalized comfort temperature model are taken as inputs, and an optimization algorithm is used for rolling optimization calculation to dynamically generate an air conditioner control instruction set, which at least includes a dynamic target temperature setting value, a compressor operation instruction and a fan rotating speed instruction, and the dynamic target temperature setting value is dynamically adjusted within the target comfort temperature interval; 2. The method of claim 1, wherein, The air conditioner control instruction set is executed, and the execution effect and user feedback are continuously monitored, and the monitored execution effect data and user feedback data are fed back to the establishment and updating steps and the optimization calculation steps of the personalized comfort temperature model as new input data. The method for dynamically generating air conditioner control instructions based on user comfort temperature preferences comprises the following steps: In an initial stage after the air conditioner is turned on, a benchmark comfort temperature range is provided as an initial parameter of the personalized comfort temperature model; User active adjustment behavior to temperature under different environment states is recorded, and the active adjustment behavior includes increasing the setting temperature, decreasing the setting temperature and turning on / off the air conditioner; Environment state data corresponding to the user active adjustment behavior is input into a machine learning algorithm as a training sample; The machine learning algorithm is used to analyze the comfort temperature preferences of a user individual under different environment states in the training sample, and the personalized comfort temperature model is established; 3. The method of claim 2, wherein, During air conditioner operation, new user active adjustment behavior and environment state data are continuously received, and the parameters of the personalized comfort temperature model are updated through cyclic iteration, so that the output target comfort temperature interval is used to continuously adapt to changes in user preferences. The method for dynamically generating air conditioner control instructions based on user comfort temperature preferences comprises the following steps: Feature extraction is performed on the environment state data in the training sample to obtain a plurality of environment feature variables, and the environment feature variables include an indoor-outdoor temperature difference, an indoor humidity value, an outdoor weather type code and a time period identifier; A setting temperature after user active adjustment behavior is taken as a target variable, and a mapping relationship between the environment feature variables and the target variable is constructed; A reinforcement learning algorithm or a collaborative filtering algorithm is used to train the mapping relationship, and user temperature adjustment behavior is taken as a feedback signal to adjust model parameters. When a new user initiates a new active adjustment behavior, the similarity between the new environment state data corresponding to the new user active adjustment behavior and the historical training samples is determined, and if the similarity is higher than a preset threshold, the target comfortable temperature interval is predicted based on the mapping relationship in the historical data, and if the similarity is lower than the preset threshold, the new environment state data is taken as a supplementary training sample, and the model parameters are optimized through an iterative updating manner.

4. The method of claim 1, wherein, The multi-source environment data and the target comfortable temperature interval output by the personalized comfort model are taken as inputs, and the optimization target is to maximize user comfort, minimize energy consumption and reduce the number of compressor start-stop times, and a rolling optimization calculation is performed through an optimization algorithm to dynamically generate an air conditioner control instruction set, including: The real-time collected multi-source environment data is preprocessed to extract the difference between the current indoor temperature and the target comfortable temperature interval, the temperature rise and fall rate, the influence coefficient of the outdoor environment parameter on the indoor temperature, and the current running parameter of the air conditioner; An optimization objective function is defined, the user comfort objective is quantified by the time proportion and temperature fluctuation amplitude of the temperature maintained in the target interval, the energy consumption objective is quantified by the actual power consumption per unit time, and the compressor start-stop objective is quantified by the number of start-stop times per unit time; A dynamic weight coefficient is set for the optimization objective, and the dynamic weight coefficient is adjusted according to the real-time environment state and user use scenario; A model predictive control algorithm is used to solve the multi-objective optimization problem in a limited time domain, and the matching compressor frequency and fan speed are found through a step-by-step approximation algorithm according to the difference between the current indoor temperature and the target temperature and the temperature rise and fall rate; A control instruction set containing a dynamic target temperature setting value, a compressor operating mode and a fan speed is generated, and the dynamic target temperature setting value is dynamically adjusted in the target comfortable temperature interval according to the optimization result.

5. The method of claim 4, wherein, The model predictive control algorithm is used to solve the multi-objective optimization problem in a limited time domain, and the matching compressor frequency and fan speed are found through a step-by-step approximation algorithm according to the difference between the current indoor temperature and the target temperature and the temperature rise and fall rate, including: A building thermal characteristic model is established, which is used to predict the indoor temperature variation trend in a preset subsequent period under a specified air conditioner output condition; Based on the building thermal characteristic model, the time and energy consumption for the indoor temperature to reach the target comfortable temperature interval under different compressor frequency and fan speed combinations are predicted, and the prediction results are substituted into the optimization objective function to calculate the comprehensive objective values of different combinations; The compressor frequency and fan speed are iteratively adjusted through a step-by-step approximation algorithm, and the comprehensive objective value is recalculated after each adjustment until the combination with the optimal comprehensive objective value is found; During the iteration process, it is determined whether the current predicted indoor temperature fluctuation amplitude is within a preset range, and if it exceeds the preset range, the fan speed is adjusted to improve air flow organization, and the compressor frequency is adjusted to control the refrigerating capacity.

6. The method of claim 1, wherein, The air conditioner control instruction set is executed, and the execution effect and user feedback are continuously monitored, and the monitored execution effect data and user feedback data are fed back to the establishment and updating steps and the optimization calculation steps of the personalized comfort model as new input data, including: The air conditioner control instruction set is sent to an air conditioner actuator, real-time monitoring of indoor temperature changes, actual energy consumption data and compressor start-stop times is performed as execution effect data; Receiving user-initiated temperature adjustment behavior as user feedback data; Analyzing the execution effect data to determine whether the current indoor temperature is maintained within the target comfortable temperature interval, whether the energy consumption is within the expected range, and whether the compressor start-stop times meet the optimization target; Comparing the execution effect data and the user feedback data with historical data to identify deviation types and causes, and if the deviation is caused by changes in user preferences, the data is fed back to the personalized comfort model establishment and updating step, and if the deviation is caused by environmental changes or equipment characteristic changes, the data is fed back to the optimization calculation step; Through a cyclic iteration method, the feedback data is used to continuously update the parameters of the personalized comfort model and the input parameters of the optimization algorithm.

7. The method of claim 6, wherein, The analysis of the execution effect data to determine whether the current indoor temperature is maintained within the target comfortable temperature interval, whether the energy consumption is within the expected range, and whether the compressor start-stop times meet the optimization target includes: Setting an allowed threshold range of indoor temperature fluctuations, calculating the ratio of the actual temperature fluctuation amplitude to the allowed threshold range, and if the ratio is greater than a preset value, determining that the temperature control effect is not up to standard; Calculating the difference between the actual energy consumption per unit time and the theoretical minimum energy consumption, and if the difference is greater than a preset energy consumption deviation threshold, determining that the energy consumption control effect is not up to standard; Counting the number of compressor start-stops per unit time, comparing it with a preset start-stop number threshold, and if it exceeds the threshold, determining that the compressor protection effect is not up to standard; Comprehensively evaluating the compliance of temperature, energy consumption and compressor start-stop times to generate an evaluation result, which includes the compliance degree of each index and the main deviation direction.

8. The method of claim 1, wherein, The real-time collection of multi-source environmental data includes: Collecting indoor environmental data through temperature sensors and humidity sensors arranged at different positions in the room, including human activity areas and near the air conditioner body; Obtaining outdoor environmental data through networking and collecting air conditioner self-operation data through built-in sensors; Timestamp alignment processing and outlier detection and elimination are performed on the collected multi-source environmental data to obtain target environmental data, which is classified and stored according to indoor environmental data, outdoor environmental data and air conditioner self-operation data to form a structured multi-source environmental data set.

9. The method of claim 1, wherein, The dynamic generation of the air conditioner control instruction set includes: Determining a dynamic target temperature set value based on the results of multi-objective optimization decision, which dynamically changes within the target comfortable temperature interval output by the personalized comfort model; Generating a compressor operation instruction based on the difference between the dynamic target temperature set value and the current indoor temperature, which includes the start-stop state and variable frequency of the compressor; Generating a fan speed instruction based on the indoor temperature distribution and target temperature uniformity requirements, which is used to adjust the speed level of the fan; Combining the dynamic target temperature set value, the compressor operation instruction and the fan speed instruction into the air conditioner control instruction set.

10. A multi-source perception based adaptive air conditioning and heating control system, comprising: It includes: A processor; A storage device having stored thereon a computer program which, when executed by the processor, causes the processor to implement the multi-source perception based adaptive air conditioning heating control method according to any one of claims 1-9.

Images