Energy-saving control method and system for multi-split air conditioner

Abstract

The invention provides an energy-saving control method and system for a multi-split air conditioner. The energy-saving control method comprises the steps that a predicted heat load demand value is obtained according to a heat load prediction model of each indoor unit; obtaining a real-time energy efficiency coefficient and an energy consumption characteristic curve according to the operation state data of the outdoor unit; calculating a target operation frequency of an outdoor unit compressor and a target opening value of an electronic expansion valve of each indoor unit in combination with the predicted thermal load demand value, the real-time energy efficiency coefficient and the energy consumption characteristic curve; generating a real-time load correction coefficient in combination with indoor personnel data; correcting the predicted thermal load demand value according to the real-time load correction coefficient, and updating the target operation frequency and the target opening value by using the corrected thermal load demand value; and calculating a deviation value between the predicted thermal load demand value and the actual thermal load and an energy efficiency coefficient deviation value between the actual energy efficiency coefficient and the predicted energy efficiency coefficient, and correcting the weight parameter and the energy consumption characteristic curve of the neural network according to the deviation value. The overall energy consumption can be reduced, and the response precision and stability are improved.

Description

Technical Field

[0001] This invention relates to the field of air conditioning technology, and specifically to an energy-saving control method and system for multi-split air conditioners. Background Technology

[0002] Multi-split air conditioning systems (VRF systems) have become the preferred solution for large-scale building air conditioning due to their core advantages such as space efficiency, independent zone control, high energy efficiency, strong installation adaptability, flexible operation, and convenient maintenance. Their core technology lies in achieving precise and coordinated control of refrigerant flow and temperature through a communication network between indoor and outdoor units. However, current mainstream VRF systems still face challenges in their energy-saving control strategies, generally exhibiting key issues such as low system energy efficiency ratios and low control algorithm accuracy, which urgently require optimization and improvement. Summary of the Invention

[0003] Based on the above-mentioned problems, this invention proposes an energy-saving control method and system for multi-split air conditioners. Through this invention, the overall energy consumption of multi-split air conditioner systems can be significantly reduced, and the system's response accuracy to dynamic load changes and long-term energy-saving stability can be improved.

[0004] In view of this, one aspect of the present invention proposes an energy-saving control method for multi-split air conditioners, comprising: By collecting indoor temperature, humidity, number of people and intensity of people's activities as environmental feature data through smart sensors deployed in each indoor unit, the weight parameters of the neural network are optimized using the quantum annealing algorithm, and a heat load prediction model for each indoor unit in the future preset period is established, and the predicted heat load demand value of each indoor unit is output. Based on IoT technology, the compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow of the outdoor unit are acquired in real time as operating status data. A digital twin model of the outdoor unit is established through 3D modeling technology. The digital twin model calculates the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data. The predicted heat load demand value is fused with the real-time energy efficiency coefficient and the energy consumption characteristic curve. With the minimization of total system energy consumption as the optimization objective, a multi-objective optimization algorithm is used to calculate the target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit. The target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve. The image recognition module installed in each indoor unit identifies the distribution density and activity status of people in the room in real time, generates a real-time load correction coefficient for each indoor unit, dynamically corrects the predicted heat load demand value of each indoor unit based on the real-time load correction coefficient, and uses the corrected heat load demand value for optimization calculation to update the target operating frequency and the target opening value. The system continuously acquires the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to reach the temperature and humidity standard as feedback data. Through data analysis technology, it calculates the heat load deviation between the predicted heat load demand value and the actual heat load, as well as the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency. Based on the heat load deviation value and the energy efficiency coefficient deviation value, it corrects the weight parameters of the neural network and the energy consumption characteristic curve, respectively.

[0005] Optionally, the step of collecting indoor temperature, humidity, number of people, and activity intensity as environmental characteristic data through intelligent sensors deployed in each indoor unit, optimizing the weight parameters of the neural network using the quantum annealing algorithm, establishing a heat load prediction model for each indoor unit within a preset future time period, and outputting the predicted heat load demand value for each indoor unit includes: The collected environmental feature data are organized according to time series to construct a training dataset containing input feature vectors and corresponding actual heat load values; A first neural network is established, comprising an input layer, a hidden layer, and an output layer. The number of nodes in the input layer corresponds to the dimension of the environmental feature data, and the nodes in the output layer correspond to the predicted heat load demand value of each indoor unit. Initial random values ​​are assigned to the weight parameters of each layer of the first neural network. The optimization problem of the weight parameters of the first neural network is transformed into an energy minimization problem. The quantum annealing algorithm is used to perform a global search in the solution space of the weight parameters. By avoiding getting trapped in local optima through the quantum tunneling effect, the optimal combination of weight parameters that minimizes the prediction error on the training dataset is obtained through iterative optimization. The optimal weight parameter combination is loaded into the first neural network to form a heat load prediction model for each indoor unit. The heat load prediction model takes current and historical environmental characteristic data as input and outputs the predicted heat load demand value of each indoor unit in the future preset period. During system operation, the latest environmental characteristic data collected by the intelligent sensors are acquired in real time, input into the heat load prediction model, and the predicted heat load demand value of each indoor unit in the future preset time period is calculated and output.

[0006] Optionally, the step of acquiring real-time compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow rate of the outdoor unit based on IoT technology as operating status data, establishing a digital twin model of the outdoor unit using 3D modeling technology, and calculating the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data, includes: Sensor nodes are deployed at key locations of the compressor, heat exchanger, and refrigerant piping of the outdoor unit. The compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow rate are collected in real time via IoT communication protocols as operating status data, and the operating status data is transmitted to the server. Based on the structural parameters and spatial layout of the outdoor unit's physical components, such as the compressor, heat exchanger, fan, and refrigerant piping, a three-dimensional geometric model of the outdoor unit is established using three-dimensional modeling technology. This three-dimensional geometric model accurately maps the spatial positional relationships and geometric dimensions of each component. Based on the aforementioned three-dimensional geometric model, energy efficiency simulation sub-models are established for components such as compressors, heat exchangers, and fans. Each energy efficiency simulation sub-model calculates the current power consumption and heat transfer performance parameters of the corresponding component based on the component's operating status data. The energy efficiency simulation sub-models of each component are integrated according to the energy flow and mass flow relationship of the actual system to form a digital twin model of the outdoor unit. The digital twin model calculates the total cooling capacity and total power consumption of the outdoor unit synchronously based on the operating status data, thereby obtaining the current real-time energy efficiency coefficient of the outdoor unit. In the digital twin model, the compressor frequency is set as a variable, and the simulation is performed at preset frequency intervals within the allowed frequency range. The operating status data other than the compressor frequency is kept unchanged. The energy efficiency coefficient and power consumption value corresponding to each frequency level are calculated, and the simulation results are plotted as an energy consumption characteristic curve with the compressor frequency as the horizontal axis and the energy efficiency coefficient and power consumption as the vertical axis.

[0007] Optionally, the step of fusing the predicted heat load demand value with the real-time energy efficiency coefficient and the energy consumption characteristic curve, and using a multi-objective optimization algorithm to calculate the target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit's electronic expansion valve with the goal of minimizing the total system energy consumption, wherein the target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve, includes: The predicted heat load demand values ​​of each indoor unit are summed to obtain the total heat load demand of the multi-split air conditioning system in the future preset period, and the minimum cooling capacity boundary that the outdoor unit compressor needs to provide is determined based on the total heat load demand. Based on the energy consumption characteristic curve, under the premise of the minimum cooling capacity boundary, all candidate compressor frequency operating points that can provide sufficient cooling capacity are selected, and the energy efficiency coefficient and power consumption value corresponding to each candidate frequency operating point are extracted to form a set of candidate frequency operating points. With minimizing the total energy consumption of the system as the primary optimization objective and minimizing the temperature deviation of each indoor unit as the secondary optimization objective, a multi-objective optimization model is established. The decision variables of the multi-objective optimization model include the operating frequency of the outdoor unit compressor and the opening value of the electronic expansion valve of each indoor unit. The constraints include the minimum cooling capacity boundary, the allowable range of compressor frequency, and the physical limitation of the opening of the electronic expansion valve. A multi-objective optimization algorithm is used to solve the multi-objective optimization model. During the solution process, the target operating frequency is searched from the candidate frequency operating point set. At the same time, the opening distribution scheme of the electronic expansion valve of each indoor unit is optimized. Through iterative calculation, the total energy consumption of the system is minimized, and it is ensured that the target operating frequency corresponds to the optimal energy efficiency operating point with the highest energy efficiency coefficient on the energy consumption characteristic curve. The target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit's electronic expansion valve are obtained from the solution and output as optimized control commands. The optimized control commands are sent to the outdoor unit compressor frequency converter and each indoor unit's electronic expansion valve actuator to achieve coordinated control of the multi-split air conditioning system.

[0008] Optionally, the step of using image recognition modules installed in each indoor unit to identify the indoor occupant distribution density and occupant activity status in real time, generating a real-time load correction coefficient for each indoor unit, dynamically correcting the predicted heat load demand value for each indoor unit based on the real-time load correction coefficient, and using the corrected heat load demand value for optimization calculation to update the target operating frequency and the target opening value includes: The image recognition module installed in each indoor unit continuously collects real-time image data of the area covered by the indoor unit, and performs preprocessing operations on the collected image data to form standardized image data suitable for subsequent recognition and analysis. The processed standardized image data is input into a pre-trained personnel detection algorithm to identify and count the actual number of people in the coverage area of ​​each indoor unit, and obtain the real-time personnel distribution density of each indoor unit; at the same time, by analyzing the changes in personnel position between consecutive image frames, the movement speed and activity amplitude of each person are calculated to determine the type of personnel activity, including sitting, walking and vigorous exercise. The deviation value of the number of people is calculated by comparing the real-time personnel distribution density with the number of people, and the activity intensity correction amount is determined based on the identified personnel activity status type. By combining the personnel number deviation value and the activity intensity correction amount, a real-time load correction coefficient for each indoor unit is generated. The real-time load correction coefficient reflects the degree of deviation of the actual load from the predicted load. The real-time load correction coefficient is applied to the predicted heat load demand value, and the corrected heat load demand value of each indoor unit is obtained by multiplication. The corrected heat load demand value can more accurately reflect the current actual heat load status of each indoor unit. The revised heat load demand values ​​for each indoor unit are used to replace the original predicted heat load demand values ​​for optimization calculation. The updated target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit are obtained by solving the problem. The updated control commands are then sent to the corresponding actuators to achieve dynamic control adjustment based on actual load changes.

[0009] Optionally, the step of continuously acquiring the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to reach the temperature and humidity standard as feedback data, calculating the heat load deviation between the predicted heat load demand value and the actual heat load, and the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency, and correcting the weight parameters of the neural network and the energy consumption characteristic curve of the digital twin model based on the heat load deviation value and the energy efficiency coefficient deviation value, includes: The system continuously collects the actual operating energy consumption of the system, the actual temperature and humidity values ​​of each indoor unit, and the time required for each indoor unit to reach the set temperature and humidity target from the issuance of the control command by energy consumption metering devices and smart sensors deployed in the multi-split air conditioning system. The collected data is stored as feedback data. Based on the actual operating energy consumption, the actual cooling capacity of the outdoor unit, and the actual temperature and humidity changes of each indoor unit, the true heat load value of each indoor unit during actual operation is calculated in reverse. The actual heat load value is compared with the predicted heat load demand value and the corrected heat load demand value to calculate the heat load deviation value between the predicted heat load demand value and the actual heat load, and the first influencing factor of the deviation is identified through data analysis technology. Based on the actual operating energy consumption and the actual cooling capacity of the outdoor unit at the target operating frequency, calculate the actual energy efficiency coefficient of the outdoor unit at that frequency. The actual energy efficiency coefficient is compared with the predicted energy efficiency coefficient predicted by the digital twin model under the same frequency and operating conditions. The deviation between the actual and predicted energy efficiency coefficients is calculated, and the influence of secondary influencing factors such as ambient temperature and operating time on the energy efficiency deviation is analyzed. Using the heat load deviation value and the first influencing factor as training samples, an online learning algorithm is used to incrementally adjust the weight parameters of the neural network. The weight parameters of each layer of the neural network are updated through the backpropagation mechanism, so that the load prediction model can gradually adapt to the changing characteristics of the actual operating environment and reduce the deviation of future predictions. The energy efficiency coefficient deviation, the second influencing factor, and the degree of influence are used as calibration data to correct the energy consumption characteristic curve generated by the digital twin model, and the predicted energy efficiency coefficient and power consumption value corresponding to each frequency level on the curve are adjusted.

[0010] Optionally, the specific steps for establishing the heat load prediction model for each indoor unit in future time periods include: Environmental characteristic data of each indoor unit within a historical time window are collected, and a neural network optimized using the quantum annealing algorithm is used for training to obtain the predicted heat load demand value of the m-th indoor unit at time t in the future. The calculation formula is as follows:

[0011] Where m is the indoor unit number, ranging from 1 to M, and M is the total number of indoor units in the system; t is the future prediction time. is the base load factor of the m-th indoor unit, reflecting the building thermal characteristics of the area where the indoor unit is located; N is the number of hidden layer nodes in the neural network; The weight of the nth hidden layer node after optimization by the quantum annealing algorithm; Let be the activation function for the nth hidden layer node, where Input for the node; Let m be the temperature difference load coefficient of the m-th indoor unit; Let be the temperature difference between the set temperature of the m-th indoor unit and the outdoor temperature at time t; Let m be the occupancy load factor for the m-th indoor unit; Let m be the predicted population density of the m-th indoor unit at time t; Let be the predicted human activity intensity coefficient for the m-th indoor unit at time t.

[0012] Optionally, the specific steps for the digital twin model to calculate the current real-time energy efficiency coefficient of the outdoor unit and the energy consumption characteristic curves corresponding to each frequency level based on the operating status data include: Based on real-time collected operating status data, a dynamic mapping relationship between the outdoor unit's energy efficiency coefficient and operating parameters is established, and the real-time energy efficiency coefficient of the outdoor unit under the current operating conditions is calculated. The calculation formula is as follows:

[0013]

[0014] in, This is the current real-time energy efficiency coefficient of the outdoor unit; The total cooling capacity of the system is calculated using the enthalpy difference and flow rate of the refrigerant at the inlet and outlet of the outdoor unit's heat exchanger. This represents the current electrical power of the compressor. This represents the current electrical power of the outdoor fan. This is the deviation correction factor for operating conditions, used to correct energy efficiency assessment errors under non-standard operating conditions; The penalty coefficient for deviation of condensation temperature; This is the current temperature at the condenser outlet. This is the optimal outlet temperature of the condenser under the current outdoor ambient temperature; This is the compressor frequency deviation penalty coefficient; This refers to the compressor's current operating frequency. This is the compressor's rated frequency.

[0015] Optionally, the specific steps for calculating the target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit's electronic expansion valve using a multi-objective optimization algorithm include: With the dual objectives of minimizing total system energy consumption and minimizing indoor unit temperature deviation, a multi-objective optimization function is established, and the target operating frequency of the outdoor unit compressor is obtained by solving the function. Its optimization model is:

[0016]

[0017]

[0018] Constraints: , ; in, A comprehensive objective function for multi-objective optimization; The weighting coefficient for energy consumption targets; For comfort target weighting coefficients, and ; Let the system energy consumption objective function be... The objective function is the comfort deviation. This refers to the compressor's operating frequency. The compressor's electrical power is a function of frequency; The outdoor fan's power is a function of frequency; The outdoor unit's coefficient of performance (COP) is a function of frequency, provided by the energy consumption characteristic curve; To control cycle duration; Let m be the comfort weighting factor for the m-th indoor unit; The actual temperature of the m-th indoor unit; The target temperature for the m-th indoor unit; This is the humidity deviation conversion factor; The actual relative humidity of the m-th indoor unit; Let m be the target relative humidity of the m-th indoor unit; and These are the minimum and maximum allowable operating frequencies of the compressor, respectively. This represents the actual cooling capacity of the m-th indoor unit.

[0019] Another aspect of the present invention provides an energy-saving control system for a multi-split air conditioner, for executing an energy-saving control method for a multi-split air conditioner, comprising: an indoor unit; intelligent sensors and an image recognition module deployed in each indoor unit; and a server; The smart sensor is configured to collect indoor temperature, humidity, number of people, and intensity of people's activities as environmental characteristic data. The server is configured as follows: The weight parameters of the neural network are optimized using the quantum annealing algorithm to establish a heat load prediction model for each indoor unit in the future preset time period, and output the predicted heat load demand value of each indoor unit. Based on IoT technology, the compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow of the outdoor unit are acquired in real time as operating status data. A digital twin model of the outdoor unit is established through 3D modeling technology. The digital twin model calculates the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data. The predicted heat load demand value is fused with the real-time energy efficiency coefficient and the energy consumption characteristic curve. With the minimization of total system energy consumption as the optimization objective, a multi-objective optimization algorithm is used to calculate the target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit. The target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve. The image recognition module is configured to: identify the indoor personnel distribution density and personnel activity status in real time, and generate real-time load correction coefficients for each indoor unit; The server is also configured to: The predicted heat load demand value of each indoor unit is dynamically corrected according to the real-time load correction coefficient, and the corrected heat load demand value is used for optimization calculation to update the target operating frequency and the target opening value. The system continuously acquires the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to reach the temperature and humidity standard as feedback data. Through data analysis technology, it calculates the heat load deviation between the predicted heat load demand value and the actual heat load, as well as the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency. Based on the heat load deviation value and the energy efficiency coefficient deviation value, it corrects the weight parameters of the neural network and the energy consumption characteristic curve, respectively.

[0020] The energy-saving control method for multi-split air conditioners using the technical solution of this invention includes: collecting indoor temperature, humidity, number of people, and activity intensity as environmental characteristic data through intelligent sensors deployed in each indoor unit; optimizing the weight parameters of the neural network using a quantum annealing algorithm; establishing a heat load prediction model for each indoor unit within a preset future time period; and outputting the predicted heat load demand value for each indoor unit; acquiring real-time compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow rate of the outdoor unit as operating status data based on Internet of Things (IoT) technology; establishing a digital twin model of the outdoor unit using 3D modeling technology; calculating the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data; fusing the predicted heat load demand value with the real-time energy efficiency coefficient and the energy consumption characteristic curves; and using a multi-objective optimization algorithm to calculate the target operating temperature of the outdoor unit compressor with the goal of minimizing the total system energy consumption. The system uses frequency and target opening values ​​of electronic expansion valves in each indoor unit, with the target operating frequency corresponding to the optimal energy efficiency operating point on the energy consumption characteristic curve. It identifies the indoor occupant density and activity status in real time using image recognition modules installed in each indoor unit, generating real-time load correction coefficients for each indoor unit. Based on these real-time load correction coefficients, it dynamically corrects the predicted heat load demand values ​​for each indoor unit and uses the corrected heat load demand values ​​for optimization calculations to update the target operating frequency and target opening value. It continuously acquires the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to achieve temperature and humidity standards as feedback data. Through data analysis technology, it calculates the heat load deviation between the predicted heat load demand value and the actual heat load, as well as the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency. Based on the heat load deviation and the energy efficiency coefficient deviation, it corrects the weight parameters of the neural network and the energy consumption characteristic curve. By establishing a load forecasting model to output predicted heat load demand values, and then combining the energy efficiency coefficient and energy consumption characteristic curves output by the energy efficiency status assessment model, global optimization decisions are made. At the same time, the control parameters are dynamically adjusted using real-time load correction coefficients obtained by image recognition, and the feedback data from actual operation is used for adaptive optimization of the model, forming a closed-loop control system of prediction-evaluation-decision-correction-feedback. Under the premise of ensuring the comfort of each area, the overall energy consumption of the multi-split air conditioning system is significantly reduced, and the system's response accuracy to dynamic load changes and long-term energy-saving stability are improved. Attached Figure Description

[0021] Figure 1 This is a flowchart of an energy-saving control method for a multi-split air conditioner provided in one embodiment of the present invention; Figure 2 This is a schematic block diagram of an energy-saving control system for a multi-split air conditioner provided in one embodiment of the present invention. Detailed Implementation

[0022] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0024] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0025] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0026] The following reference Figures 1 to 2 This invention describes an energy-saving control method and system for multi-split air conditioners, provided by some embodiments of the present invention.

[0027] like Figure 1 As shown, one embodiment of the present invention provides an energy-saving control method for multi-split air conditioners, comprising: By collecting indoor temperature, humidity, number of people and intensity of people's activities as environmental feature data through smart sensors deployed in each indoor unit, the weight parameters of the neural network are optimized using the quantum annealing algorithm, and a heat load prediction model for each indoor unit in the future preset period is established, and the predicted heat load demand value of each indoor unit is output. Based on IoT technology, the compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow of the outdoor unit are acquired in real time as operating status data. A digital twin model of the outdoor unit is established through 3D modeling technology. The digital twin model calculates the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data. The predicted heat load demand value is fused with the real-time energy efficiency coefficient and the energy consumption characteristic curve. With the minimization of total system energy consumption as the optimization objective, a multi-objective optimization algorithm is used to calculate the target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit. The target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve. The image recognition module installed in each indoor unit identifies the distribution density and activity status of people in the room in real time, generates a real-time load correction coefficient for each indoor unit, dynamically corrects the predicted heat load demand value of each indoor unit based on the real-time load correction coefficient, and uses the corrected heat load demand value for optimization calculation to update the target operating frequency and the target opening value. The system continuously acquires the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to reach the temperature and humidity standard as feedback data. Through data analysis technology, it calculates the heat load deviation between the predicted heat load demand value and the actual heat load, as well as the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency. Based on the heat load deviation value and the energy efficiency coefficient deviation value, it corrects the weight parameters of the neural network and the energy consumption characteristic curve of the digital twin model, thereby achieving continuous adaptive optimization of the load prediction model and the energy efficiency assessment model.

[0028] The technical solution adopted in this embodiment establishes a load prediction model to output predicted heat load demand values, and then combines the energy efficiency coefficient and energy consumption characteristic curve output by the energy efficiency status assessment model to make global optimization decisions. At the same time, the control parameters are dynamically adjusted by using real-time load correction coefficients obtained by image recognition, and the feedback data from actual operation is used for adaptive optimization of the model, forming a closed-loop control system of prediction-evaluation-decision-correction-feedback. Under the premise of ensuring the comfort of each area, the overall energy consumption of the multi-split air conditioning system is significantly reduced, and the system's response accuracy to dynamic load changes and long-term energy-saving stability are improved.

[0029] In some possible embodiments of the present invention, the step of collecting indoor temperature, humidity, number of people, and activity intensity as environmental characteristic data through intelligent sensors deployed in each indoor unit, optimizing the weight parameters of the neural network using the quantum annealing algorithm, establishing a heat load prediction model for each indoor unit within a preset future time period, and outputting the predicted heat load demand value for each indoor unit includes: The collected environmental feature data are organized according to time series to construct a training dataset containing input feature vectors and corresponding actual heat load values; A first neural network is established, comprising an input layer, a hidden layer, and an output layer. The number of nodes in the input layer corresponds to the dimension of the environmental feature data, and the nodes in the output layer correspond to the predicted heat load demand value of each indoor unit. Initial random values ​​are assigned to the weight parameters of each layer of the first neural network. The optimization problem of the weight parameters of the first neural network is transformed into an energy minimization problem. The quantum annealing algorithm is used to perform a global search in the solution space of the weight parameters. By avoiding getting trapped in local optima through the quantum tunneling effect, the optimal combination of weight parameters that minimizes the prediction error on the training dataset is obtained through iterative optimization. The optimal weight parameter combination is loaded into the first neural network to form a heat load prediction model for each indoor unit. The heat load prediction model takes current and historical environmental characteristic data as input and outputs the predicted heat load demand value of each indoor unit in the future preset period. During system operation, the latest environmental characteristic data collected by the intelligent sensors are acquired in real time, input into the heat load prediction model, and the predicted heat load demand value of each indoor unit in the future preset time period is calculated and output.

[0030] In this embodiment, the global optimization capability of the quantum annealing algorithm overcomes the shortcomings of the traditional gradient descent algorithm, which is prone to getting trapped in local optima. This enables the neural network to find a better weight parameter configuration, significantly improving the learning ability and prediction accuracy of the heat load prediction model for complex nonlinear environmental factors. At the same time, by combining the collected multidimensional environmental feature data with quantum optimization technology, the prediction model can accurately capture the load change patterns under the coupled effects of multiple factors such as temperature, humidity, number of people, and activity intensity. This provides a more reliable and accurate load prediction basis for subsequent global optimization decisions, thereby improving the energy-saving effect and response accuracy of the overall control system.

[0031] In some possible embodiments of the present invention, the step of acquiring real-time compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow rate of the outdoor unit based on Internet of Things technology as operating status data, establishing a digital twin model of the outdoor unit through three-dimensional modeling technology, and calculating the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data includes: Sensor nodes are deployed at key locations of the compressor, heat exchanger, and refrigerant piping of the outdoor unit. The compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow rate are collected in real time via the Internet of Things communication protocol as operating status data, and the operating status data is transmitted to the server (equipped with a control system). Based on the structural parameters and spatial layout of the outdoor unit's physical components, such as the compressor, heat exchanger, fan, and refrigerant piping, a three-dimensional geometric model of the outdoor unit is established using three-dimensional modeling technology. This three-dimensional geometric model accurately maps the spatial positional relationships and geometric dimensions of each component. Based on the three-dimensional geometric model, energy efficiency simulation sub-models are established for components such as compressors, heat exchangers and fans. Each energy efficiency simulation sub-model calculates the current power consumption and heat transfer performance parameters of the corresponding component based on the component's operating status data (extracted from the operating status data). The energy efficiency simulation sub-models of each component are integrated according to the energy flow and mass flow relationship of the actual system to form a digital twin model of the outdoor unit. The digital twin model calculates the total cooling capacity and total power consumption of the outdoor unit synchronously based on the operating status data, thereby obtaining the current real-time energy efficiency coefficient of the outdoor unit. In the digital twin model, the compressor frequency is set as a variable, and the simulation is performed at preset frequency intervals within the allowed frequency range. The operating status data other than the compressor frequency is kept unchanged. The energy efficiency coefficient and power consumption value corresponding to each frequency level are calculated, and the simulation results are plotted as an energy consumption characteristic curve with the compressor frequency as the horizontal axis and the energy efficiency coefficient and power consumption as the vertical axis.

[0032] In this embodiment, IoT technology is used to collect multi-dimensional operating status data of the outdoor unit in real time. Combined with the digital twin model established by 3D modeling technology, it can accurately map the energy conversion process and interaction relationship of each component of the outdoor unit under actual operating conditions. The digital twin model can not only calculate the energy efficiency coefficient of the current operating condition in real time, but also predict the energy consumption characteristics at different frequency levels through virtual simulation. It provides energy efficiency assessment information across the entire frequency range for optimization decisions, enabling the control system to accurately identify the optimal energy efficiency operating point. This avoids the assessment errors caused by relying on experience curves or static data in traditional methods, and significantly improves the real-time performance, accuracy and adaptability to changing operating conditions of energy efficiency status assessment.

[0033] In some possible embodiments of the present invention, the step of fusing the predicted heat load demand value with the real-time energy efficiency coefficient and the energy consumption characteristic curve, using the minimization of total system energy consumption as the optimization objective, and employing a multi-objective optimization algorithm to calculate the target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit electronic expansion valve, wherein the target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve, includes: The predicted heat load demand values ​​of each indoor unit are summed to obtain the total heat load demand of the multi-split air conditioning system in the future preset period, and the minimum cooling capacity boundary that the outdoor unit compressor needs to provide is determined based on the total heat load demand. Based on the energy consumption characteristic curve, under the premise of the minimum cooling capacity boundary, all candidate compressor frequency operating points that can provide sufficient cooling capacity are selected, and the energy efficiency coefficient and power consumption value corresponding to each candidate frequency operating point are extracted to form a set of candidate frequency operating points. With minimizing the total energy consumption of the system as the primary optimization objective and minimizing the temperature deviation of each indoor unit as the secondary optimization objective, a multi-objective optimization model is established. The decision variables of the multi-objective optimization model include the operating frequency of the outdoor unit compressor and the opening value of the electronic expansion valve of each indoor unit. The constraints include the minimum cooling capacity boundary, the allowable range of compressor frequency, and the physical limitation of the opening of the electronic expansion valve. A multi-objective optimization algorithm is used to solve the multi-objective optimization model. During the solution process, the target operating frequency is searched from the candidate frequency operating point set. At the same time, the opening distribution scheme of the electronic expansion valve of each indoor unit is optimized. Through iterative calculation, the total energy consumption of the system is minimized, and it is ensured that the target operating frequency corresponds to the optimal energy efficiency operating point with the highest energy efficiency coefficient on the energy consumption characteristic curve. The target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit's electronic expansion valve are obtained from the solution and output as optimized control commands. The optimized control commands are sent to the outdoor unit compressor frequency converter and each indoor unit's electronic expansion valve actuator to achieve coordinated control of the multi-split air conditioning system.

[0034] In this embodiment, by deeply integrating load forecasting results with real-time energy efficiency assessment information, precise matching between supply and demand is achieved, avoiding energy efficiency losses caused by the disconnect between load forecasting and equipment control in traditional methods. The multi-objective optimization algorithm can intelligently search for the optimal energy efficiency operating point from the energy consumption characteristic curve under the constraint of meeting the total heat load demand, ensuring that the outdoor unit compressor operates at the frequency level with the highest energy efficiency coefficient. At the same time, it achieves reasonable load distribution by optimizing the opening of the electronic expansion valves of each indoor unit. This global collaborative optimization mechanism breaks the limitations of independent control of each component, enabling compressor frequency regulation and expansion valve opening control to work together, significantly reducing the overall energy consumption of the system when meeting comfort requirements, and improving the operating efficiency and energy-saving effect of the multi-split air conditioning system.

[0035] In some possible embodiments of the present invention, the step of identifying the indoor occupant distribution density and occupant activity status in real time through the image recognition module installed in each indoor unit, generating a real-time load correction coefficient for each indoor unit, dynamically correcting the predicted heat load demand value of each indoor unit based on the real-time load correction coefficient, and using the corrected heat load demand value for optimization calculation to update the target operating frequency and the target opening value includes: The image recognition module installed in each indoor unit continuously collects real-time image data of the area covered by the indoor unit, and performs preprocessing operations on the collected image data to form standardized image data suitable for subsequent recognition and analysis. The processed standardized image data is input into a pre-trained personnel detection algorithm to identify and count the actual number of people in the coverage area of ​​each indoor unit, and obtain the real-time personnel distribution density of each indoor unit; at the same time, by analyzing the changes in personnel position between consecutive image frames, the movement speed and activity amplitude of each person are calculated to determine the type of personnel activity, including sitting, walking and vigorous exercise. The deviation value of the number of people is calculated by comparing the real-time personnel distribution density with the number of people, and the activity intensity correction amount is determined based on the identified personnel activity status type. By combining the personnel number deviation value and the activity intensity correction amount, a real-time load correction coefficient for each indoor unit is generated. The real-time load correction coefficient reflects the degree of deviation of the actual load from the predicted load. The real-time load correction coefficient is applied to the predicted heat load demand value, and the corrected heat load demand value of each indoor unit is obtained by multiplication. The corrected heat load demand value can more accurately reflect the current actual heat load status of each indoor unit. The revised heat load demand values ​​for each indoor unit are used to replace the original predicted heat load demand values ​​for optimization calculation. The updated target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit are obtained by solving the problem. The updated control commands are then sent to the corresponding actuators to achieve dynamic control adjustment based on actual load changes.

[0036] In this embodiment, image recognition technology is used to capture real-time information on sudden changes in the distribution and activity status of people indoors, compensating for the shortcomings of traditional sensors in accurately sensing the dynamic characteristics of people's activities. This enables the control system to promptly detect deviations between the actual load and the predicted load. The introduction of a real-time load correction coefficient establishes a rapid response path from image recognition to load correction and then to control updates, allowing the system to quickly adjust its control strategy in scenarios such as sudden increases in people or drastic changes in activity status, avoiding response lag caused by relying on periodic sensor sampling. This image recognition-assisted dynamic correction mechanism significantly improves the tracking accuracy and response speed of the control system for actual load fluctuations, ensuring efficient and energy-saving operation even in complex usage scenarios with frequent personnel movement, while ensuring that the comfort needs of each area are met in a timely manner.

[0037] In some possible embodiments of the present invention, the step of continuously acquiring the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time for temperature and humidity to reach the standard as feedback data, calculating the heat load deviation between the predicted heat load demand value and the actual heat load, and the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency, and correcting the weight parameters of the neural network and the energy consumption characteristic curve of the digital twin model based on the heat load deviation value and the energy efficiency coefficient deviation value, respectively, includes: The system continuously collects the actual operating energy consumption of the system, the actual temperature and humidity values ​​of each indoor unit, and the time required for each indoor unit to reach the set temperature and humidity target from the issuance of the control command by energy consumption metering devices and smart sensors deployed in the multi-split air conditioning system. The collected data is stored as feedback data. Based on the actual operating energy consumption, the actual cooling capacity of the outdoor unit, and the actual temperature and humidity changes of each indoor unit, the true heat load value of each indoor unit during actual operation is calculated in reverse. The actual heat load value is compared with the predicted heat load demand value and the corrected heat load demand value to calculate the heat load deviation value between the predicted heat load demand value and the actual heat load, and the first influencing factor of the deviation is identified through data analysis technology. Based on the actual operating energy consumption and the actual cooling capacity of the outdoor unit at the target operating frequency, calculate the actual energy efficiency coefficient of the outdoor unit at that frequency. The actual energy efficiency coefficient is compared with the predicted energy efficiency coefficient predicted by the digital twin model under the same frequency and operating conditions. The deviation between the actual and predicted energy efficiency coefficients is calculated, and the influence of secondary influencing factors such as ambient temperature and operating time on the energy efficiency deviation is analyzed. Using the heat load deviation value and the first influencing factor as training samples, an online learning algorithm is used to incrementally adjust the weight parameters of the neural network. The weight parameters of each layer of the neural network are updated through the backpropagation mechanism, so that the load prediction model can gradually adapt to the changing characteristics of the actual operating environment and reduce the deviation of future predictions. The energy efficiency coefficient deviation, the second influencing factor, and the degree of influence are used as calibration data to correct the energy consumption characteristic curve generated by the digital twin model. The predicted energy efficiency coefficient and power consumption value corresponding to each frequency level on the curve are adjusted so that the energy consumption characteristic curve can more accurately reflect the energy efficiency characteristics of the outdoor unit under actual working conditions.

[0038] In this embodiment, by establishing a closed-loop learning mechanism from control execution to data feedback and then to model correction, the control system possesses the ability to continuously self-evolve, adapting to dynamic factors during long-term operation such as equipment aging, environmental changes, and evolving usage patterns. The dual feedback mechanism of actual heat load deviation and energy efficiency coefficient deviation ensures that the load prediction model and energy efficiency assessment model can be optimized synchronously, avoiding the accumulation of systematic deviations caused by single model correction. The application of online learning algorithms enables model parameters to be incrementally updated based on actual operating data without interrupting system operation or retraining, ensuring the continuity and stability of the control system. This adaptive optimization mechanism continuously improves the prediction accuracy and control effect of the multi-split air conditioning system during long-term operation, significantly enhancing the system's robustness in dealing with complex and changing usage scenarios, and ensuring that energy-saving effects remain at the optimal level throughout the entire life cycle.

[0039] In some possible embodiments of the present invention, the specific steps for establishing the heat load prediction model for each indoor unit for future periods include: Environmental characteristic data of each indoor unit within a historical time window are collected, and a neural network optimized using the quantum annealing algorithm is used for training to obtain the predicted heat load demand value of the m-th indoor unit at time t in the future. The calculation formula is as follows:

[0040] Where m is the indoor unit number, ranging from 1 to M, and M is the total number of indoor units in the system; t is the future prediction time. is the base load factor of the m-th indoor unit, reflecting the building thermal characteristics of the area where the indoor unit is located; N is the number of hidden layer nodes in the neural network; The weight of the nth hidden layer node after optimization by the quantum annealing algorithm; Let be the activation function for the nth hidden layer node, where Input for the node; Let m be the temperature difference load coefficient of the m-th indoor unit; Let be the temperature difference between the set temperature of the m-th indoor unit and the outdoor temperature at time t; Let m be the occupancy load factor for the m-th indoor unit; Let m be the predicted population density of the m-th indoor unit at time t; Let be the predicted human activity intensity coefficient for the m-th indoor unit at time t.

[0041] In this embodiment, a quantum-optimized neural network is used to capture complex nonlinear characteristics. At the same time, the physical mechanisms of temperature difference and human factors are explicitly introduced, which improves the accuracy and robustness of the predicted heat load demand value in dynamic environments, thereby providing a more reliable load prediction basis for subsequent optimization decisions.

[0042] In some possible embodiments of the present invention, the specific steps by which the digital twin model calculates the current real-time energy efficiency coefficient of the outdoor unit and the energy consumption characteristic curves corresponding to each frequency level based on the operating status data include: Based on real-time collected operating status data, a dynamic mapping relationship between the outdoor unit's energy efficiency coefficient and operating parameters is established, and the real-time energy efficiency coefficient of the outdoor unit under the current operating conditions is calculated. The calculation formula is as follows:

[0043]

[0044] in, This is the current real-time energy efficiency coefficient of the outdoor unit; The total cooling capacity of the system is calculated using the enthalpy difference and flow rate of the refrigerant at the inlet and outlet of the outdoor unit's heat exchanger. This represents the current electrical power of the compressor. This represents the current electrical power of the outdoor fan. This is the deviation correction factor for operating conditions, used to correct energy efficiency assessment errors under non-standard operating conditions; The penalty coefficient for deviation of condensation temperature; This is the current temperature at the condenser outlet. This is the optimal outlet temperature of the condenser under the current outdoor ambient temperature; This is the compressor frequency deviation penalty coefficient; This refers to the compressor's current operating frequency. This is the compressor's rated frequency.

[0045] This embodiment accurately quantifies the nonlinear impact of condensing temperature deviation and compressor frequency deviation on energy efficiency by introducing a working condition deviation correction coefficient. This enables the digital twin model to accurately reflect the true energy efficiency status of the outdoor unit under actual variable working conditions, providing an accurate energy efficiency assessment basis for global optimization decisions.

[0046] In some possible embodiments of the present invention, the specific steps of calculating the target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit electronic expansion valve using a multi-objective optimization algorithm include: With the dual objectives of minimizing total system energy consumption and minimizing indoor unit temperature deviation, a multi-objective optimization function is established, and the target operating frequency of the outdoor unit compressor is obtained by solving the function. Its optimization model is:

[0047]

[0048]

[0049] Constraints: , ; in, A comprehensive objective function for multi-objective optimization; The weighting coefficient for energy consumption targets; For comfort target weighting coefficients, and ; Let the system energy consumption objective function be... The objective function is the comfort deviation. This refers to the compressor's operating frequency. The compressor's electrical power is a function of frequency; The outdoor fan's power is a function of frequency; The outdoor unit's coefficient of performance (COP) is a function of frequency, provided by the energy consumption characteristic curve; To control cycle duration; Let m be the comfort weighting factor for the m-th indoor unit; The actual temperature of the m-th indoor unit; The target temperature for the m-th indoor unit; This is the humidity deviation conversion factor; The actual relative humidity of the m-th indoor unit; Let m be the target relative humidity of the m-th indoor unit; and These are the minimum and maximum allowable operating frequencies of the compressor, respectively. This represents the actual cooling capacity of the m-th indoor unit.

[0050] The optimization model in this embodiment achieves synergistic optimization of energy saving and comfort by weighting energy consumption and introducing a comfort function that jointly evaluates temperature and humidity. This ensures that the comfort requirements of each indoor unit are met while reducing system energy consumption.

[0051] In some possible embodiments of the present invention, the specific steps of generating a real-time load correction coefficient for each indoor unit by real-time identification of indoor personnel distribution density and personnel activity status through image recognition modules installed in each indoor unit include: The image recognition module identifies the population density and activity status in each indoor unit area, and calculates the real-time load correction coefficient for the m-th indoor unit. The predicted heat load demand is dynamically corrected, and the calculation formula is as follows:

[0052]

[0053]

[0054] in, This is the real-time load correction factor for the m-th indoor unit; Correction gain for personnel number deviation; The m-th indoor unit represents the actual number of people detected by image recognition. Let m be the number of people predicted by the m-th indoor unit; To prevent tiny positive numbers with a denominator of zero; Adjust the gain for activity intensity; The movement speed of the p-th person within the m-th indoor unit area is calculated using the inter-frame displacement of continuous images. The speed of movement is used as a reference, representing moderate activity intensity; This represents the corrected heat load demand value for the m-th indoor unit; This is a spatial distribution correction factor, reflecting the concentration of people in the indoor unit's coverage area; This is a spatial concentration correction factor; The distance between the detected center of gravity of the person and the center of the indoor unit's air supply; Let m be the effective air supply radius of the m-th indoor unit.

[0055] This embodiment comprehensively considers the deviation in the number of people, the intensity of their activities, and the spatial distribution characteristics to accurately correct the predicted heat load demand value in real time. This enables the control system to quickly respond to sudden changes in the actual indoor load, improving the accuracy of load allocation and the real-time performance of the system response.

[0056] Please refer to Figure 2 Another embodiment of the present invention provides an energy-saving control system for a multi-split air conditioner, for executing an energy-saving control method for a multi-split air conditioner, comprising: an indoor unit; intelligent sensors and image recognition modules deployed in each indoor unit; and a server; The smart sensor is configured to collect indoor temperature, humidity, number of people, and intensity of people's activities as environmental characteristic data. The server is configured as follows: The weight parameters of the neural network are optimized using the quantum annealing algorithm to establish a heat load prediction model for each indoor unit in the future preset time period, and output the predicted heat load demand value of each indoor unit. Based on IoT technology, the compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow of the outdoor unit are acquired in real time as operating status data. A digital twin model of the outdoor unit is established through 3D modeling technology. The digital twin model calculates the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data. The predicted heat load demand value is fused with the real-time energy efficiency coefficient and the energy consumption characteristic curve. With the minimization of total system energy consumption as the optimization objective, a multi-objective optimization algorithm is used to calculate the target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit. The target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve. The image recognition module is configured to: identify the indoor personnel distribution density and personnel activity status in real time, and generate real-time load correction coefficients for each indoor unit; The server is also configured to: The predicted heat load demand value of each indoor unit is dynamically corrected according to the real-time load correction coefficient, and the corrected heat load demand value is used for optimization calculation to update the target operating frequency and the target opening value. The system continuously acquires the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to reach the temperature and humidity standard as feedback data. Through data analysis technology, it calculates the heat load deviation between the predicted heat load demand value and the actual heat load, as well as the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency. Based on the heat load deviation value and the energy efficiency coefficient deviation value, it corrects the weight parameters of the neural network and the energy consumption characteristic curve of the digital twin model, thereby achieving continuous adaptive optimization of the load prediction model and the energy efficiency assessment model.

[0057] It should be known that, Figure 2 The block diagram of the energy-saving control system for multi-split air conditioners shown is for illustrative purposes only, and the number of modules shown does not limit the scope of protection of this invention. The energy-saving control system for multi-split air conditioners provided in this embodiment can be used to execute various embodiments of the corresponding energy-saving control method for multi-split air conditioners. For specific implementation details, please refer to the descriptions of the respective method embodiments, which will not be repeated here.

[0058] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0059] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0060] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0061] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0062] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0063] If the integrated units described above are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0064] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0065] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0066] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can easily conceive of variations or substitutions without departing from the spirit and scope of the present invention, and various modifications and alterations can be made, including combinations of the different functions and implementation steps described above, as well as software and hardware implementation methods, all of which are within the protection scope of the present invention.

Claims

1. An energy-saving control method for multi-split air conditioners, characterized in that, include: By collecting indoor temperature, humidity, number of people and intensity of people's activities as environmental feature data through smart sensors deployed in each indoor unit, the weight parameters of the neural network are optimized using the quantum annealing algorithm, and a heat load prediction model for each indoor unit in the future preset period is established, and the predicted heat load demand value of each indoor unit is output. Based on IoT technology, the compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow of the outdoor unit are acquired in real time as operating status data. A digital twin model of the outdoor unit is established through 3D modeling technology. The digital twin model calculates the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data. The predicted heat load demand value is fused with the real-time energy efficiency coefficient and the energy consumption characteristic curve. With the minimization of total system energy consumption as the optimization objective, a multi-objective optimization algorithm is used to calculate the target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit. The target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve. The image recognition module installed in each indoor unit identifies the distribution density and activity status of people in the room in real time, generates a real-time load correction coefficient for each indoor unit, dynamically corrects the predicted heat load demand value of each indoor unit based on the real-time load correction coefficient, and uses the corrected heat load demand value for optimization calculation to update the target operating frequency and the target opening value. The system continuously acquires the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to reach the temperature and humidity standard as feedback data. Through data analysis technology, it calculates the heat load deviation between the predicted heat load demand value and the actual heat load, as well as the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency. Based on the heat load deviation value and the energy efficiency coefficient deviation value, it corrects the weight parameters of the neural network and the energy consumption characteristic curve, respectively.

2. The energy-saving control method for multi-split air conditioners according to claim 1, characterized in that, The steps of collecting indoor temperature, humidity, number of people, and activity intensity as environmental characteristic data through intelligent sensors deployed in each indoor unit, optimizing the weight parameters of the neural network using the quantum annealing algorithm, establishing a heat load prediction model for each indoor unit within a preset future time period, and outputting the predicted heat load demand value for each indoor unit include: The collected environmental feature data are organized according to time series to construct a training dataset containing input feature vectors and corresponding actual heat load values; A first neural network is established, comprising an input layer, a hidden layer, and an output layer. The number of nodes in the input layer corresponds to the dimension of the environmental feature data, and the nodes in the output layer correspond to the predicted heat load demand value of each indoor unit. Initial random values ​​are assigned to the weight parameters of each layer of the first neural network. The optimization problem of the weight parameters of the first neural network is transformed into an energy minimization problem. The quantum annealing algorithm is used to perform a global search in the solution space of the weight parameters. By avoiding getting trapped in local optima through the quantum tunneling effect, the optimal combination of weight parameters that minimizes the prediction error on the training dataset is obtained through iterative optimization. The optimal weight parameter combination is loaded into the first neural network to form a heat load prediction model for each indoor unit. The heat load prediction model takes current and historical environmental characteristic data as input and outputs the predicted heat load demand value of each indoor unit in the future preset period. During system operation, the latest environmental characteristic data collected by the intelligent sensors are acquired in real time, input into the heat load prediction model, and the predicted heat load demand value of each indoor unit in the future preset time period is calculated and output.

3. The energy-saving control method for multi-split air conditioners according to claim 2, characterized in that, The steps of acquiring real-time compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow rate of the outdoor unit based on IoT technology as operating status data, establishing a digital twin model of the outdoor unit using 3D modeling technology, and calculating the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data include: Sensor nodes are deployed at key locations of the compressor, heat exchanger, and refrigerant piping of the outdoor unit. The compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow rate are collected in real time via IoT communication protocols as operating status data, and the operating status data is transmitted to the server. Based on the structural parameters and spatial layout of the outdoor unit's physical components, such as the compressor, heat exchanger, fan, and refrigerant piping, a three-dimensional geometric model of the outdoor unit is established using three-dimensional modeling technology. This three-dimensional geometric model accurately maps the spatial positional relationships and geometric dimensions of each component. Based on the aforementioned three-dimensional geometric model, energy efficiency simulation sub-models are established for components such as compressors, heat exchangers, and fans. Each energy efficiency simulation sub-model calculates the current power consumption and heat transfer performance parameters of the corresponding component based on the component's operating status data. The energy efficiency simulation sub-models of each component are integrated according to the energy flow and mass flow relationship of the actual system to form a digital twin model of the outdoor unit. The digital twin model calculates the total cooling capacity and total power consumption of the outdoor unit synchronously based on the operating status data, thereby obtaining the current real-time energy efficiency coefficient of the outdoor unit. In the digital twin model, the compressor frequency is set as a variable, and the simulation is performed at preset frequency intervals within the allowed frequency range. The operating status data other than the compressor frequency is kept unchanged. The energy efficiency coefficient and power consumption value corresponding to each frequency level are calculated, and the simulation results are plotted as an energy consumption characteristic curve with the compressor frequency as the horizontal axis and the energy efficiency coefficient and power consumption as the vertical axis.

4. The energy-saving control method for multi-split air conditioners according to claim 3, characterized in that, The step of fusing the predicted heat load demand with the real-time energy efficiency coefficient and the energy consumption characteristic curve, using the minimization of total system energy consumption as the optimization objective, and employing a multi-objective optimization algorithm to calculate the target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit's electronic expansion valve, wherein the target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve, includes: The predicted heat load demand values ​​of each indoor unit are summed to obtain the total heat load demand of the multi-split air conditioning system in the future preset period, and the minimum cooling capacity boundary that the outdoor unit compressor needs to provide is determined based on the total heat load demand. Based on the energy consumption characteristic curve, under the premise of the minimum cooling capacity boundary, all candidate compressor frequency operating points that can provide sufficient cooling capacity are selected, and the energy efficiency coefficient and power consumption value corresponding to each candidate frequency operating point are extracted to form a set of candidate frequency operating points. With minimizing the total energy consumption of the system as the primary optimization objective and minimizing the temperature deviation of each indoor unit as the secondary optimization objective, a multi-objective optimization model is established. The decision variables of the multi-objective optimization model include the operating frequency of the outdoor unit compressor and the opening value of the electronic expansion valve of each indoor unit. The constraints include the minimum cooling capacity boundary, the allowable range of compressor frequency, and the physical limitation of the opening of the electronic expansion valve. A multi-objective optimization algorithm is used to solve the multi-objective optimization model. During the solution process, the target operating frequency is searched from the candidate frequency operating point set. At the same time, the opening distribution scheme of the electronic expansion valve of each indoor unit is optimized. Through iterative calculation, the total energy consumption of the system is minimized, and it is ensured that the target operating frequency corresponds to the optimal energy efficiency operating point with the highest energy efficiency coefficient on the energy consumption characteristic curve. The target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit's electronic expansion valve are obtained from the solution and output as optimized control commands. The optimized control commands are sent to the outdoor unit compressor frequency converter and each indoor unit's electronic expansion valve actuator to achieve coordinated control of the multi-split air conditioning system.

5. The energy-saving control method for multi-split air conditioners according to claim 4, characterized in that, The steps of using image recognition modules installed in each indoor unit to identify the indoor occupant density and activity status in real time, generating a real-time load correction coefficient for each indoor unit, dynamically correcting the predicted heat load demand value for each indoor unit based on the real-time load correction coefficient, and using the corrected heat load demand value for optimization calculation to update the target operating frequency and the target opening value include: The image recognition module installed in each indoor unit continuously collects real-time image data of the area covered by the indoor unit, and performs preprocessing operations on the collected image data to form standardized image data suitable for subsequent recognition and analysis. The processed standardized image data is input into a pre-trained personnel detection algorithm to identify and count the actual number of people in the coverage area of ​​each indoor unit, and obtain the real-time personnel distribution density of each indoor unit; at the same time, by analyzing the changes in personnel position between consecutive image frames, the movement speed and activity amplitude of each person are calculated to determine the type of personnel activity, including sitting, walking and vigorous exercise. The deviation value of the number of people is calculated by comparing the real-time personnel distribution density with the number of people, and the activity intensity correction amount is determined based on the identified personnel activity status type. By combining the personnel number deviation value and the activity intensity correction amount, a real-time load correction coefficient for each indoor unit is generated. The real-time load correction coefficient reflects the degree of deviation of the actual load from the predicted load. The real-time load correction coefficient is applied to the predicted heat load demand value, and the corrected heat load demand value of each indoor unit is obtained by multiplication. The corrected heat load demand value can more accurately reflect the current actual heat load status of each indoor unit. The revised heat load demand values ​​for each indoor unit are used to replace the original predicted heat load demand values ​​for optimization calculation. The updated target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit are obtained by solving the problem. The updated control commands are then sent to the corresponding actuators to achieve dynamic control adjustment based on actual load changes.

6. The energy-saving control method for multi-split air conditioners according to claim 5, characterized in that, The steps of continuously acquiring the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to reach the temperature and humidity standard as feedback data, calculating the heat load deviation between the predicted heat load demand value and the actual heat load, and the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency, and correcting the weight parameters of the neural network and the energy consumption characteristic curve of the digital twin model based on the heat load deviation value and the energy efficiency coefficient deviation value, respectively, include: The system continuously collects the actual operating energy consumption of the system, the actual temperature and humidity values ​​of each indoor unit, and the time required for each indoor unit to reach the set temperature and humidity target from the issuance of the control command by energy consumption metering devices and smart sensors deployed in the multi-split air conditioning system. The collected data is stored as feedback data. Based on the actual operating energy consumption, the actual cooling capacity of the outdoor unit, and the actual temperature and humidity changes of each indoor unit, the true heat load value of each indoor unit during actual operation is calculated in reverse. The actual heat load value is compared with the predicted heat load demand value and the corrected heat load demand value to calculate the heat load deviation value between the predicted heat load demand value and the actual heat load, and the first influencing factor of the deviation is identified through data analysis technology. Based on the actual operating energy consumption and the actual cooling capacity of the outdoor unit at the target operating frequency, calculate the actual energy efficiency coefficient of the outdoor unit at that frequency. The actual energy efficiency coefficient is compared with the predicted energy efficiency coefficient predicted by the digital twin model under the same frequency and operating conditions. The deviation between the actual and predicted energy efficiency coefficients is calculated, and the influence of secondary influencing factors such as ambient temperature and operating time on the energy efficiency deviation is analyzed. Using the heat load deviation value and the first influencing factor as training samples, an online learning algorithm is used to incrementally adjust the weight parameters of the neural network. The weight parameters of each layer of the neural network are updated through the backpropagation mechanism, so that the load prediction model can gradually adapt to the changing characteristics of the actual operating environment and reduce the deviation of future predictions. The energy efficiency coefficient deviation, the second influencing factor, and the degree of influence are used as calibration data to correct the energy consumption characteristic curve generated by the digital twin model, and the predicted energy efficiency coefficient and power consumption value corresponding to each frequency level on the curve are adjusted.

7. The energy-saving control method for multi-split air conditioners according to claim 6, characterized in that, The specific steps for establishing the heat load prediction model for each indoor unit in future time periods include: Environmental characteristic data of each indoor unit within a historical time window are collected, and a neural network optimized using the quantum annealing algorithm is used for training to obtain the predicted heat load demand value of the m-th indoor unit at time t in the future. The calculation formula is as follows: Where m is the indoor unit number, ranging from 1 to M, and M is the total number of indoor units in the system; t is the future prediction time. is the base load factor of the m-th indoor unit, reflecting the building thermal characteristics of the area where the indoor unit is located; N is the number of hidden layer nodes in the neural network; The weight of the nth hidden layer node after optimization by the quantum annealing algorithm; Let be the activation function for the nth hidden layer node, where Input for the node; Let m be the temperature difference load coefficient of the m-th indoor unit; Let be the temperature difference between the set temperature of the m-th indoor unit and the outdoor temperature at time t; Let m be the occupancy load factor for the m-th indoor unit; Let m be the predicted population density of the m-th indoor unit at time t; Let be the predicted human activity intensity coefficient for the m-th indoor unit at time t.

8. The energy-saving control method for multi-split air conditioners according to claim 7, characterized in that, The specific steps for the digital twin model to calculate the current real-time energy efficiency coefficient of the outdoor unit and the energy consumption characteristic curves corresponding to each frequency level based on the operating status data include: Based on real-time collected operating status data, a dynamic mapping relationship between the outdoor unit's energy efficiency coefficient and operating parameters is established, and the real-time energy efficiency coefficient of the outdoor unit under the current operating conditions is calculated. The calculation formula is as follows: in, This is the current real-time energy efficiency coefficient of the outdoor unit; The total cooling capacity of the system is calculated using the enthalpy difference and flow rate of the refrigerant at the inlet and outlet of the outdoor unit's heat exchanger. This represents the current electrical power of the compressor. This represents the current electrical power of the outdoor fan. This is the deviation correction factor for operating conditions, used to correct energy efficiency assessment errors under non-standard operating conditions; The penalty coefficient for deviation of condensation temperature; This is the current temperature at the condenser outlet. This is the optimal outlet temperature of the condenser under the current outdoor ambient temperature; This is the compressor frequency deviation penalty coefficient; This refers to the compressor's current operating frequency. This is the compressor's rated frequency.

9. The energy-saving control method for multi-split air conditioners according to claim 8, characterized in that, The specific steps for calculating the target operating frequency of the outdoor unit compressor and the target opening value of each indoor unit's electronic expansion valve using a multi-objective optimization algorithm include: With the dual objectives of minimizing total system energy consumption and minimizing indoor unit temperature deviation, a multi-objective optimization function is established, and the target operating frequency of the outdoor unit compressor is obtained by solving the function. Its optimization model is: Constraints: , ; in, A comprehensive objective function for multi-objective optimization; The weighting coefficient for energy consumption targets; For comfort target weighting coefficients, and ; Let the system energy consumption objective function be... The objective function is the comfort deviation. This refers to the compressor's operating frequency. The compressor's electrical power is a function of frequency; The outdoor fan's power is a function of frequency; The outdoor unit's coefficient of performance (COP) is a function of frequency, provided by the energy consumption characteristic curve; To control cycle duration; Let m be the comfort weighting factor for the m-th indoor unit; The actual temperature of the m-th indoor unit; The target temperature for the m-th indoor unit; This is the humidity deviation conversion factor; The actual relative humidity of the m-th indoor unit; Let m be the target relative humidity of the m-th indoor unit; and These are the minimum and maximum allowable operating frequencies of the compressor, respectively. This represents the actual cooling capacity of the m-th indoor unit.

10. An energy-saving control system for a multi-split air conditioner, used to execute the energy-saving control method for a multi-split air conditioner as described in any one of claims 1 to 9, characterized in that, include: Indoor unit; Intelligent sensors and image recognition modules deployed in each indoor unit; server; The smart sensor is configured to collect indoor temperature, humidity, number of people, and intensity of people's activities as environmental characteristic data. The server is configured as follows: The weight parameters of the neural network are optimized using the quantum annealing algorithm to establish a heat load prediction model for each indoor unit in the future preset time period, and output the predicted heat load demand value of each indoor unit. Based on IoT technology, the compressor frequency, compressor current, heat exchanger inlet and outlet temperatures, and refrigerant flow of the outdoor unit are acquired in real time as operating status data. A digital twin model of the outdoor unit is established through 3D modeling technology. The digital twin model calculates the current real-time energy efficiency coefficient and energy consumption characteristic curves corresponding to each frequency level of the outdoor unit based on the operating status data. The predicted heat load demand value is fused with the real-time energy efficiency coefficient and the energy consumption characteristic curve. With the minimization of total system energy consumption as the optimization objective, a multi-objective optimization algorithm is used to calculate the target operating frequency of the outdoor unit compressor and the target opening value of the electronic expansion valve of each indoor unit. The target operating frequency corresponds to the optimal energy efficiency operating point on the energy consumption characteristic curve. The image recognition module is configured to: identify the indoor personnel distribution density and personnel activity status in real time, and generate real-time load correction coefficients for each indoor unit; The server is also configured to: The predicted heat load demand value of each indoor unit is dynamically corrected according to the real-time load correction coefficient, and the corrected heat load demand value is used for optimization calculation to update the target operating frequency and the target opening value. The system continuously acquires the actual operating energy consumption of the multi-split air conditioning system, the actual temperature and humidity values ​​of each indoor unit, and the time to reach the temperature and humidity standard as feedback data. Through data analysis technology, it calculates the heat load deviation between the predicted heat load demand value and the actual heat load, as well as the energy efficiency coefficient deviation between the actual energy efficiency coefficient and the predicted energy efficiency coefficient at the target operating frequency. Based on the heat load deviation value and the energy efficiency coefficient deviation value, it corrects the weight parameters of the neural network and the energy consumption characteristic curve, respectively.

Images