Magnetic levitation direct expansion air handling unit control optimization method based on MPC algorithm

Abstract

The application discloses a magnetic suspension direct expansion air handling unit control optimization method based on an MPC algorithm. The method collects unit operation state parameters, external environment parameters and target control parameters in real time; a unit dynamic prediction model is constructed based on a mechanism modeling and data driving fusion mode, a mapping relationship between control variables, disturbance variables and output variables is established; a multi-objective rolling optimization function and a device safe operation constraint condition are set; the prediction model is used to predict future working conditions, and an optimal control sequence is solved under the condition of meeting the constraint condition; the first control instruction of the optimal control sequence is executed, actual output feedback is collected in the next period, the prediction model is corrected online based on the actual and prediction deviation, and closed-loop rolling optimization control is realized. Through the construction of the accurate unit prediction model, the application realizes the collaborative optimization adjustment of multiple control variables by combining the rolling optimization strategy, ensures the stable operation of the unit under complex working conditions, and maximally reduces the operation energy consumption and improves the energy efficiency.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology for air conditioning systems, and in particular to a control optimization method for magnetic levitation direct expansion air conditioning units based on the MPC algorithm. Background Technology

[0002] Magnetic levitation direct expansion air conditioning units are widely used in temperature control scenarios of various buildings due to their advantages such as high efficiency, energy saving, and stable operation. The unit mainly consists of a magnetic levitation compressor condensing unit (including core components such as a magnetic levitation compressor, condenser, and electronic expansion valve), an evaporator, a self-cleaning electronic filter section, and a variable frequency fan. It uses water cooling, and its supply air temperature and humidity need to be dynamically adjusted according to the target return air temperature and humidity, changes in fresh air volume, and changes in ambient temperature and humidity.

[0003] Currently, the control of magnetic levitation direct expansion air conditioning units generally adopts the traditional PID control algorithm. PID control adjusts the control quantity through the coordinated action of proportional, integral, and derivative components to achieve tracking of the target parameter. However, in practical applications, the operation of magnetic levitation direct expansion air conditioning units exhibits significant multivariate coupling, large inertia, and nonlinear characteristics. They are also susceptible to various disturbances such as fluctuations in outdoor temperature and humidity, changes in indoor load, and fluctuations in fresh air volume. Traditional PID control algorithms have gradually revealed numerous shortcomings: First, PID control parameter tuning relies on experience, making it difficult to adapt to complex and changing operating conditions. PID controllers with fixed parameters are prone to overshoot and oscillation when operating conditions change abruptly, resulting in poor unit stability and an inability to accurately maintain the supply air temperature and humidity within the set range. Second, PID control is a feedback control system, only able to adjust based on historical and current error signals, lacking the ability to predict future changes in operating conditions and failing to proactively avoid the adverse effects of disturbances, leading to control lag. Third, PID control focuses on minimizing the steady-state error of temperature and humidity, failing to fully consider the synergistic optimization among control variables such as the magnetic levitation compressor speed, fan frequency, and electronic expansion valve opening. This makes it difficult to achieve the globally optimal goal of minimizing unit energy consumption and maximizing energy efficiency, and fails to fully leverage the high energy efficiency advantages of magnetic levitation compressors under partial load conditions.

[0004] Model predictive control (MPC), as an advanced multivariable control strategy, possesses the characteristics of rolling optimization and online feedback correction. It can predict the system output over a future period based on a predictive model, considering system constraints, and solve for the optimal control sequence through optimization algorithms. This effectively addresses the control challenges of multivariable coupled, large-inertia, and nonlinear systems. Applying the MPC algorithm to the control optimization of magnetic levitation direct expansion air conditioning units is expected to overcome the limitations of traditional PID control, achieving a synergistic improvement in unit operational stability and energy efficiency. This represents an important development direction in the field of intelligent control for air conditioning systems. Summary of the Invention

[0005] Based on this, this application provides a control optimization method for magnetic levitation direct expansion air conditioning units based on the MPC algorithm. By constructing an accurate unit prediction model and combining it with a rolling optimization strategy, the method achieves coordinated optimization and adjustment of multiple control variables, ensuring stable operation of the unit under complex operating conditions while minimizing operating energy consumption and improving energy efficiency. This solves the problems of poor operating stability, control lag, and low energy efficiency that exist when traditional PID control algorithms are applied to magnetic levitation direct expansion air conditioning units.

[0006] Firstly, a control optimization method for magnetic levitation direct expansion air conditioning units based on the MPC algorithm is provided, the method comprising:

[0007] Real-time acquisition of unit operating status parameters, external environmental parameters, and target control parameters;

[0008] A dynamic prediction model for the unit is constructed based on a fusion of mechanism modeling and data-driven approach, and a mapping relationship between control variables, disturbance variables and output variables is established.

[0009] Define the multi-objective rolling optimization function and the constraints for safe operation of the equipment;

[0010] The prediction model is used to predict future operating conditions, and the optimal control sequence is solved under the constraints; wherein the constraints include control variable constraints and output variable constraints.

[0011] The first control command of the optimal control sequence is executed, and the actual output feedback is collected in the next cycle. The prediction model is corrected online based on the deviation between the actual and the prediction to achieve closed-loop rolling optimization control.

[0012] Optionally, the real-time acquisition of unit operating status parameters, external environmental parameters, and target control parameters includes:

[0013] Temperature and humidity sensors are installed on the outdoor side of the unit to collect outdoor ambient temperature and humidity. Temperature and humidity sensors are installed in the return air duct to collect indoor return air temperature and humidity. Air volume sensors are installed in the fresh air duct to collect fresh air volume. Load sensors are installed indoors to collect indoor cooling load. Temperature sensors and pressure sensors are installed at the inlet and outlet of the evaporator and condenser, respectively.

[0014] The unit's electrical control system collects data on the magnetic levitation compressor speed, variable frequency fan operating frequency, electronic expansion valve opening, and real-time energy consumption, and sets target values ​​for supply air temperature and humidity.

[0015] Optionally, the method of constructing a dynamic prediction model for the unit based on the fusion of mechanism modeling and data-driven approaches includes:

[0016] A unit mechanism model is established based on thermodynamic and fluid mechanics principles. The parameters of the mechanism model are corrected and optimized using the least squares method, support vector machine, or neural network with historical operating data to obtain the prediction model y(k+i|k) =f[u(k+i-1|k), d(k+i|k), y(k|k)];

[0017] Where i = 1, 2, ..., N_p, N_p is the prediction time domain length; y(k+i|k) represents the output variable value at time k+i predicted based on the measurement value at time k; u(k+i-1|k) represents the control variable value at time k+i-1 for the decision at time k; d(k+i|k) represents the disturbance variable value at time k+i predicted at time k.

[0018] Optionally, a multi-objective rolling optimization function and equipment safety operation constraints are defined, including constructing the optimization objective function:

[0019] J=ω1·Σ(T_supply(k+i|k)-T_supply_ref)²+ω2·Σ(RH_supply(k+i|k)-RH_supply_ref)²+ω3·ΣP_energy(k+i|k)

[0020] Where i = 1, 2, ..., N_p; ω1, ω2, ω3 are the weighting coefficients for supply air temperature deviation, supply air humidity deviation, and operating energy consumption, respectively; Σ represents the summation of various deviations in the prediction time domain; T_supply(k+i|k) represents the supply air temperature predicted at time k+i based on the measurement value at time k; T_supply_ref represents the preset target value for supply air temperature; RH_supply(k+i|k) represents the supply air humidity predicted at time k+i based on the measurement value at time k; RH_supply_ref represents the preset target value for supply air humidity; P_energy(k+i|k) represents the unit operating energy consumption predicted at time k+i based on the measurement value at time k.

[0021] Optionally, the step of using a predictive model to predict future operating conditions and solving for the optimal control sequence under satisfied constraints includes:

[0022] Based on the actual output and predicted disturbance values ​​at the current moment, the output variables at the next N_p moments are predicted using a prediction model. Sequence quadratic programming or particle swarm optimization algorithms are then used to solve for the optimal control sequence at the next N_c moments, provided that the constraints are met. The control time domain length N_c does not exceed the prediction time domain length N_p.

[0023] Optionally, the step of executing the first control command of the optimal control sequence and collecting actual output feedback in the next cycle, and correcting the prediction model online based on the deviation between the actual and predicted outputs, includes:

[0024] The first control quantity of the optimal control sequence is output to the actuators of the magnetic levitation compressor, variable frequency fan and electronic expansion valve. In the next control cycle, the actual operating parameters are collected again, the deviation between the actual output and the predicted output is calculated, and the deviation is used to perform online parameter correction of the prediction model.

[0025] Optionally, the control variable constraints include magnetic levitation compressor speed constraints, variable frequency fan frequency constraints, and electronic expansion valve opening constraints; the output variable constraints include allowable fluctuation range constraints for supply air temperature and allowable fluctuation range constraints for supply air humidity.

[0026] Optionally, the method further includes a feedforward control step:

[0027] Based on the historical data and trends of the external environmental parameters and cooling load, the trajectory of the change of disturbance variables in the future prediction time domain is predicted.

[0028] The predicted change trajectory of the disturbance variable is input into the unit dynamic prediction model, so that the prediction model can respond in advance to the changes in the external environmental parameters and cooling load when solving the optimal control sequence, thereby realizing feedforward compensation control.

[0029] Secondly, a control optimization system for a magnetic levitation direct expansion air conditioning unit based on the MPC algorithm is provided. This system includes:

[0030] The data acquisition module is used to collect unit operating status parameters, external environmental parameters, and target control parameters in real time.

[0031] The module is used to build a dynamic prediction model for the unit based on the fusion of mechanism modeling and data-driven approach, and to establish the mapping relationship between control variables, disturbance variables and output variables;

[0032] The configuration module is used to set the multi-objective rolling optimization function and the constraints for safe operation of the equipment; the constraints include control variable constraints and output variable constraints.

[0033] The prediction module is used to predict future operating conditions using the prediction model and solve for the optimal control sequence under the constraints.

[0034] The output correction module is used to execute the first control command of the optimal control sequence and collect the actual output feedback in the next cycle. Based on the deviation between the actual and the prediction, the prediction model is corrected online to realize closed-loop rolling optimization control.

[0035] Thirdly, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the methods described in the first aspect above.

[0036] The beneficial effects of the technical solutions provided in this application include at least the following:

[0037] (1) Improve the stability of unit operation: This application adopts the rolling optimization and feedback correction mechanism of MPC algorithm, which can predict changes in outdoor environment, indoor load and other interference factors in advance, dynamically adjust the control strategy, effectively solve the overshoot and oscillation problems that are easy to occur in traditional PID control, ensure that the supply air temperature and humidity are stable within the set range, reduce the fluctuation range by more than 50%, and significantly improve the stability of unit operation.

[0038] (2) Achieving optimal global energy efficiency: The optimization objective function constructed in this application comprehensively considers the temperature and humidity control accuracy and operating energy consumption. By coordinating and optimizing the speed of the magnetic levitation compressor, the frequency of the fan and the opening of the electronic expansion valve through multi-variable collaborative optimization, the high efficiency and energy saving advantages of the magnetic levitation compressor under partial load conditions are fully utilized. Compared with traditional PID control, the unit's operating energy consumption is reduced by more than 20%.

[0039] (3) Enhanced adaptability to operating conditions: This application constructs a prediction model by combining mechanism and data-driven methods and has an online correction function. It can accurately adapt to the nonlinear and large inertia characteristics of the unit, and adapt to changes in outdoor temperature and humidity, fresh air volume and indoor load. It has strong adaptability to operating conditions and does not require frequent manual adjustment of control parameters.

[0040] (4) Ensure safe operation of equipment: In the optimization process, this application sets constraints on control variables and output variables to avoid core components such as magnetic levitation compressor, fan, and electronic expansion valve from operating under extreme conditions, thereby extending the service life of the equipment and reducing maintenance costs. Attached Figure Description

[0041] To more clearly illustrate the embodiments of this application or the technical solutions in the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0042] Figure 1 A flowchart illustrating the steps of a magnetic levitation direct expansion air conditioning unit control optimization method based on the MPC algorithm provided in this application embodiment;

[0043] Figure 2A system block diagram of a magnetic levitation direct expansion air conditioning unit control optimization method based on MPC algorithm provided in this application embodiment;

[0044] Figure 3 This is a schematic diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0046] In the description of this application, the terms "comprising," "having," and any variations thereof are intended to cover non-exclusive inclusion, such as a process, method, system, product, or apparatus that includes a series of steps or units, not necessarily limited to those steps or units that are expressly listed, but may also include other steps or units that are not expressly listed but are inherent to these processes, methods, products, or apparatuses, or steps or units added based on further optimizations conceived in this application.

[0047] To address the problems of poor operational stability, control lag, and low energy efficiency when traditional PID control algorithms are applied to magnetic levitation direct expansion air conditioning units, this application provides a control optimization method for magnetic levitation direct expansion air conditioning units based on the MPC algorithm. By constructing an accurate unit prediction model and combining it with a rolling optimization strategy, the method achieves coordinated optimization and adjustment of multiple control variables, ensuring stable operation of the unit under complex operating conditions while minimizing operating energy consumption and improving energy efficiency.

[0048] Please refer to Figure 1 The document illustrates a flowchart of a control optimization method for a magnetic levitation direct expansion air conditioning unit based on the MPC algorithm, provided in an embodiment of this application. This method may include the following steps:

[0049] S1 collects unit operating status parameters, external environmental parameters, and target control parameters in real time.

[0050] This step involves parameter acquisition. Specifically, sensors are deployed at key locations within the magnetic levitation direct expansion air conditioning unit to collect the following parameters in real time and transmit them to the MPC controller:

[0051] Environmental and load parameters: outdoor ambient temperature T_out, outdoor ambient humidity RH_out, indoor return air temperature T_in, indoor return air humidity RH_in, fresh air volume Q_f, cooling load P_load (obtained through power sensor or load calculation model);

[0052] Unit operating status parameters: Evaporator inlet and outlet temperatures (i.e., evaporation temperatures) are T_eva_in and T_eva_out respectively; condenser inlet and outlet temperatures (i.e., condenser temperatures) are T_cond_in and T_cond_out respectively; evaporator inlet and outlet pressures are P_eva_in and P_eva_out; condenser inlet and outlet pressures are P_cond_in and P_cond_out respectively; magnetic levitation compressor speed is f_comp; variable frequency fan operating frequency is f_fan; electronic expansion valve opening is K_valve; and unit real-time energy consumption is P_energy.

[0053] Control target parameters: preset supply air temperature target value T_supply_ref, supply air humidity target value RH_supply_ref.

[0054] In optional embodiments of this application, the unit's operating parameters, environmental parameters, and control target parameters are collected in real time by sensors. The unit operating parameters include evaporator temperature, condenser temperature, evaporator inlet and outlet pressure, condenser inlet and outlet pressure, magnetic levitation compressor speed, variable frequency fan operating frequency, electronic expansion valve opening, and real-time unit energy consumption. The environmental parameters include outdoor ambient temperature, outdoor ambient humidity, indoor return air temperature, indoor return air humidity, fresh air volume, and indoor cooling load. The control target parameters include preset target values ​​for supply air temperature and supply air humidity.

[0055] S2, based on the fusion of mechanism modeling and data-driven approach, constructs a dynamic prediction model for the unit and establishes a mapping relationship between control variables, disturbance variables and output variables.

[0056] This step involves constructing a predictive model for a magnetic levitation direct expansion air conditioning unit. Based on a combination of mechanism analysis and data-driven methods, a predictive model is built to describe the dynamic mapping relationship between control variables and system output variables.

[0057] S21, Define variables:

[0058] The control variable u(k) = [f_comp(k), f_fan(k), K_valve(k)]^T, where f_comp(k) is the speed of the magnetic levitation compressor at time k, f_fan(k) is the frequency of the variable frequency fan at time k, and K_valve(k) is the opening degree of the electronic expansion valve at time k;

[0059] The interference variable d(k) = [T_out(k),RH_out(k),Q_f(k),P_load(k)]^T, where each parameter is defined in the same way as in step S1;

[0060] The output variable y(k) = [T_supply(k), RH_supply(k), P_energy(k)]^T, where T_supply(k) is the supply air temperature at time k, RH_supply(k) is the supply air humidity at time k, and P_energy(k) is the unit's operating energy consumption at time k.

[0061] S22, Model Building:

[0062] Based on thermodynamic and fluid mechanics principles, a mechanistic model of a magnetic levitation direct expansion air conditioning unit is established to describe the variation patterns of various parameters during refrigerant circulation and air handling. Simultaneously, using historical operating data collected in step S1, data-driven methods such as least squares, support vector machines, or neural networks are employed to correct and optimize the parameters of the mechanistic model, resulting in a high-precision unit prediction model: y(k+i|k)=f[u(k+i-1|k),d(k+i|k),y(k|k)], where i=1,2,...,N_p, and N_p is the prediction time domain length; y(k+i|k) represents the output variable value predicted at time k+i based on the measurement value at time k; u(k+i-1|k) represents the control variable value at time k+i-1 based on the decision made at time k; and d(k+i|k) represents the disturbance variable value predicted at time k+i based on the value made at time k.

[0063] In an optional embodiment of this application, a predictive model for a magnetic levitation direct expansion air conditioning unit is constructed based on a method combining mechanism analysis and data-driven approaches. The predictive model is used to describe the dynamic mapping relationship between control variables, disturbance variables, and output variables. The control variables are the magnetic levitation compressor speed, variable frequency fan frequency, and electronic expansion valve opening. The disturbance variables are the outdoor ambient temperature, outdoor ambient humidity, fresh air volume, and indoor cooling load. The output variables are the supply air temperature, supply air humidity, and unit operating energy consumption.

[0064] The process of constructing the unit prediction model is as follows: first, a unit mechanism model is established based on thermodynamic and fluid mechanics principles; then, the parameters of the mechanism model are corrected and optimized using collected historical operating data through a data-driven method, which is one of least squares method, support vector machine, or neural network.

[0065] S3 sets the multi-objective rolling optimization function and the constraints for safe operation of the equipment.

[0066] This step involves setting the MPC optimization objective and constraints, including:

[0067] S31, Optimize the objective function:

[0068] With the core objectives of minimizing unit operating energy consumption, maximizing energy efficiency, and accurately tracking supply air temperature and humidity, a multi-objective optimization function is constructed. A weighted summation method is then used to transform the multi-objective problem into a single-objective optimization problem.

[0069] J=ω1·Σ(T_supply(k+i|k)-T_supply_ref)²+ω2·Σ(RH_supply(k+i|k)-RH_supply_ref)²+ω3·ΣP_energy(k+i|k)

[0070] Where i = 1, 2, ..., N_p; ω1, ω2, and ω3 are the weighting coefficients for supply air temperature deviation, supply air humidity deviation, and operating energy consumption, respectively, which can be dynamically adjusted according to actual operating needs; Σ represents the summation of various deviations in the prediction time domain.

[0071] S32, Constraints:

[0072] To ensure the safe and stable operation of the unit, control variable constraints and output variable constraints are set:

[0073] Control variable constraints: f_comp_min≤f_comp(k+i|k)≤f_comp_max, f_fan_min≤f_fan(k+i|k)≤f_fan_max, K_valve_min≤K_valve(k+i|k)≤K_valve_max, where f_comp_min and f_comp_max are the minimum and maximum speeds of the magnetic levitation compressor, respectively; f_fan_min and f_fan_max are the minimum and maximum operating frequencies of the variable frequency fan, respectively; and K_valve_min and K_valve_max are the minimum and maximum opening degrees of the electronic expansion valve, respectively.

[0074] Output variable constraints: T_supply_min≤T_supply(k+i|k)≤T_supply_max, RH_supply_min≤RH_supply(k+i|k)≤RH_supply_max, where T_supply_min and T_supply_max are the allowable fluctuation ranges of supply air temperature, and RH_supply_min and RH_supply_max are the allowable fluctuation ranges of supply air humidity.

[0075] In this embodiment, an MPC optimization objective function and constraints are defined. The objective function is a weighted summation multi-objective function, and the weight coefficients can be dynamically adjusted. The constraints include control variable constraints and output variable constraints. Control variable constraints include magnetic levitation compressor speed constraints, variable frequency fan frequency constraints, and electronic expansion valve opening constraints. Output variable constraints include allowable fluctuation range constraints for supply air temperature and allowable fluctuation range constraints for supply air humidity.

[0076] S4 uses a predictive model to predict future operating conditions and solves for the optimal control sequence under the constraints.

[0077] In this step, the MPC rolling optimization solution is implemented. Based on the actual output y(k) and the predicted disturbance value d(k+i|k) collected at time k, the MPC controller uses the prediction model constructed in step S2 to predict the output variable y(k+i|k) for the next N_p times. Using the optimization objective function set in step S3 as the criterion, and under the premise of satisfying the constraints, the optimal control sequence u for the next N_c times is obtained by solving optimization algorithms such as sequential quadratic programming and particle swarm optimization. (k|k),u (k+1|k),...,u (k+N_c-1|k), where N_c is the control time domain length and N_c≤N_p.

[0078] In an optional embodiment of this application, based on the actual parameters collected at the current moment and the prediction model, the output variables in the future prediction time domain are predicted, and under the premise of satisfying the constraints, the optimal control sequence in the future control time domain is obtained by solving through an optimization algorithm.

[0079] S5 executes the first control command of the optimal control sequence and collects actual output feedback in the next cycle. Based on the deviation between the actual and predicted outputs, the prediction model is corrected online to achieve closed-loop rolling optimization control.

[0080] This step implements control command output and feedback correction, specifically employing a rolling implementation strategy, where only the first control variable u of the optimal control sequence is applied. (k|k) serves as the control command for the current moment, outputting to the actuators of the magnetic levitation compressor, variable frequency fan, and electronic expansion valve to adjust the speed of the magnetic levitation compressor, the frequency of the fan, and the opening of the electronic expansion valve. After entering the next control cycle k+1, the actual operating parameters and environmental parameters at that moment are re-acquired, the deviation between the actual output y(k+1) and the predicted output y(k+1|k) is calculated, and the prediction model is corrected online using this deviation to update the prediction model parameters. Then, the optimization solution process in step S4 is repeated to obtain a new optimal control sequence, thus realizing closed-loop rolling optimization control.

[0081] This method also includes a feedforward control step:

[0082] Based on the historical data and trends of the external environmental parameters and cooling load, the trajectory of the change of disturbance variables in the future prediction time domain is predicted.

[0083] The predicted change trajectory of the disturbance variable is input into the unit dynamic prediction model, so that the prediction model can respond in advance to the changes in the external environmental parameters and cooling load when solving the optimal control sequence, thereby realizing feedforward compensation control.

[0084] The following is another optional embodiment of this application. To further illustrate the implementation process of this application, a specific description is given in conjunction with a magnetic levitation direct expansion air conditioning unit for a public building:

[0085] 1. Parameter Acquisition Configuration: Temperature and humidity sensors are installed on the outdoor side of the unit to collect outdoor ambient temperature T_out and humidity RH_out; temperature and humidity sensors are installed in the return air duct to collect return air temperature T_in and humidity RH_in; air volume sensors are installed in the fresh air duct to collect fresh air volume Q_f; load sensors are installed indoors to collect indoor cooling load P_load; temperature sensors and pressure sensors are installed at the inlet and outlet of the evaporator and condenser respectively to collect T_eva_in, T_eva_out, T_cond_in, T_cond_out, P_eva_in, P_eva_out, P_cond_in, P_cond_out; the unit's electrical control system collects the magnetic levitation compressor speed f_comp, variable frequency fan frequency f_fan, electronic expansion valve opening K_valve, and real-time energy consumption P_energy; the target value for supply air temperature T_supply_ref=20℃ and the target value for supply air humidity RH_supply_ref=50%RH.

[0086] 2. Predictive Model Construction: Based on the thermodynamic equation of the refrigerant cycle and the mechanism of the air handling process, an initial mechanism model was established. Historical operating data of the unit under different operating conditions (covering different outdoor temperatures and humidity, different fresh air volumes, and different indoor loads) were collected, totaling 1000 sets of valid data. Of these, 800 sets were used for model training and 200 sets were used for model verification. The support vector machine algorithm was used to correct the parameters of the mechanism model, resulting in the final unit prediction model. Verification showed that the prediction errors of this model for supply air temperature, humidity, and energy consumption were all less than 3%, meeting the control accuracy requirements.

[0087] 3. MPC parameter settings: The prediction time domain N_p is set to 10 control cycles, and the control time domain N_c is set to 3 control cycles; the weight coefficients of the optimization objective function are set to ω1=0.4, ω2=0.3, and ω3=0.3 to ensure the balance between temperature and humidity control accuracy and energy-saving target; the control variable constraints are set as follows: magnetic levitation compressor speed f_comp∈[10000rpm,40000rpm], variable frequency fan frequency f_fan∈[20Hz,50Hz], and electronic expansion valve opening K_valve∈[10%,90%]; the output variable constraints are set as follows: supply air temperature T_supply∈[24℃,26℃], and supply air humidity RH_supply∈[45%,55%].

[0088] 4. Operation and Control Process: After the unit starts up, the sensors collect various parameters every 10 seconds and transmit them to the MPC controller. Based on the parameters at the current moment, the MPC controller uses a predictive model to predict the supply air temperature, humidity, and energy consumption for the next 10 control cycles (a total of 100 seconds). The optimal control sequence is solved using a particle swarm optimization algorithm to obtain the compressor speed, fan frequency, and electronic expansion valve opening for the next 3 control cycles. The optimal control command for the first control cycle is output to the actuator to adjust the operating parameters of the corresponding equipment. After 10 seconds, the next control cycle begins, new actual parameters are collected, the prediction deviation is calculated, and the model is corrected. The above optimization and control process is repeated to achieve dynamic optimized operation of the unit.

[0089] In an optional embodiment of this application, the first control quantity of the optimal control sequence is output to the actuator as the control command at the current moment to adjust the speed of the magnetic levitation compressor, the frequency of the fan, and the opening of the electronic expansion valve; after entering the next control cycle, steps 1-4 are repeated to achieve rolling optimization control.

[0090] In another optional embodiment of this application, the MPC controller is used to acquire sensor signals from the information acquisition device and run a neural network (such as LSTM) based on the sensor signals to generate load forecast data; and the controller is configured to execute a global optimization algorithm based on the load forecast data to obtain optimized setting parameters for the compressor and the main expansion valve, and to perform PID adaptive adjustment of the optimized setting parameters to monitor the operating status of the compressor and constrain the safety boundary.

[0091] Table 1 shows the high energy efficiency data of the magnetic levitation unit under partial load. Through the rolling optimization strategy of the MPC algorithm, the operating load of the magnetic levitation compressor is adjusted to the high-efficiency range of 25%-60%. The large-span intermittent operation mode in the traditional control is replaced by small-amplitude continuous adjustment, thereby avoiding the low-efficiency zone of full-load operation and giving full play to the high-efficiency and energy-saving advantages of the magnetic levitation compressor under partial load conditions, so as to achieve high-efficiency and stable operation of the unit.

[0092] Table 1

[0093]

[0094] like Figure 2 This application also provides a control optimization system for a magnetic levitation direct expansion air conditioning unit based on the MPC algorithm, which may include:

[0095] The data acquisition module is used to collect unit operating status parameters, external environmental parameters, and target control parameters in real time.

[0096] The module is used to build a dynamic prediction model for the unit based on the fusion of mechanism modeling and data-driven approach, and to establish the mapping relationship between control variables, disturbance variables and output variables;

[0097] The configuration module is used to set the multi-objective rolling optimization function and the constraints for safe operation of the equipment; the constraints include control variable constraints and output variable constraints.

[0098] The prediction module is used to predict future operating conditions using the prediction model and solve for the optimal control sequence under the constraints.

[0099] The output correction module is used to execute the first control command of the optimal control sequence and collect the actual output feedback in the next cycle. Based on the deviation between the actual and the prediction, the prediction model is corrected online to realize closed-loop rolling optimization control.

[0100] Specific limitations regarding the control optimization system for magnetic levitation direct expansion air conditioning units based on the MPC algorithm can be found in the limitations of the control optimization method for magnetic levitation direct expansion air conditioning units based on the MPC algorithm mentioned above, and will not be repeated here. Each module in the aforementioned control optimization system for magnetic levitation direct expansion air conditioning units based on the MPC algorithm can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0101] In one embodiment, an electronic device is provided, which may be a computer, and its internal structure diagram may be as follows: Figure 3As shown, the electronic device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for control optimization data of the magnetic levitation direct expansion air conditioning unit. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a control optimization method for a magnetic levitation direct expansion air conditioning unit.

[0102] Those skilled in the art will understand that, Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0103] In one embodiment of this application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the above-described control optimization method for magnetic levitation direct expansion air conditioning units based on the MPC algorithm.

[0104] In one embodiment of this application, a computer program product is provided, including a computer program / instructions, which, when executed by a processor, implements the steps of the above-described control optimization method for magnetic levitation direct expansion air conditioning units based on the MPC algorithm.

[0105] The computer-readable storage medium and computer program product provided in this embodiment are similar in implementation principle and technical effect to the above method embodiments, and will not be described again here.

[0106] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0107] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0108] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A control optimization method for a magnetic levitation direct expansion air conditioning unit based on the MPC algorithm, characterized in that, The method includes: Real-time acquisition of unit operating status parameters, external environmental parameters, and target control parameters; A dynamic prediction model for the unit is constructed based on a fusion of mechanism modeling and data-driven approach, and a mapping relationship between control variables, disturbance variables and output variables is established. Define a multi-objective rolling optimization function and constraints for safe equipment operation; the constraints include control variable constraints and output variable constraints. The prediction model is used to predict future operating conditions, and the optimal control sequence is solved under the constraints. The first control command of the optimal control sequence is executed, and the actual output feedback is collected in the next cycle. The prediction model is corrected online based on the deviation between the actual and the prediction to achieve closed-loop rolling optimization control.

2. The method according to claim 1, characterized in that, The real-time acquisition of unit operating status parameters, external environmental parameters, and target control parameters includes: Temperature and humidity sensors are installed on the outdoor side of the unit to collect outdoor ambient temperature and humidity. Temperature and humidity sensors are installed in the return air duct to collect indoor return air temperature and humidity. Air volume sensors are installed in the fresh air duct to collect fresh air volume. Load sensors are installed indoors to collect indoor cooling load. Temperature sensors and pressure sensors are installed at the inlet and outlet of the evaporator and condenser, respectively. The unit's electrical control system collects data on the magnetic levitation compressor speed, variable frequency fan operating frequency, electronic expansion valve opening, and real-time energy consumption, and sets target values ​​for supply air temperature and humidity.

3. The method according to claim 1, characterized in that, The method of constructing a dynamic prediction model for the unit based on the fusion of mechanism modeling and data-driven approaches includes: A unit mechanism model is established based on thermodynamic and fluid mechanics principles. The parameters of the mechanism model are corrected and optimized using the least squares method, support vector machine, or neural network with historical operating data to obtain the prediction model y(k+i|k) = f[u(k+i-1|k), d(k+i|k), y(k|k)]; Where i = 1, 2, ..., N_p, N_p is the prediction time domain length; y(k+i|k) represents the output variable value at time k+i predicted based on the measurement value at time k; u(k+i-1|k) represents the control variable value at time k+i-1 for the decision at time k; d(k+i|k) represents the disturbance variable value at time k+i predicted at time k.

4. The method according to claim 3, characterized in that, Define the multi-objective rolling optimization function and the constraints for safe operation of the equipment, including constructing the optimization objective function: J=ω1·Σ(T_supply(k+i|k)-T_supply_ref)²+ω2·Σ(RH_supply(k+i|k)-RH_supply_ref)²+ω3·ΣP_energy(k+i|k) Where i = 1, 2, ..., N_p; ω1, ω2, ω3 are the weighting coefficients for supply air temperature deviation, supply air humidity deviation, and operating energy consumption, respectively; Σ represents the summation of various deviations in the prediction time domain; T_supply(k+i|k) represents the supply air temperature predicted at time k+i based on the measurement value at time k; T_supply_ref represents the preset target value for supply air temperature; RH_supply(k+i|k) represents the supply air humidity predicted at time k+i based on the measurement value at time k; RH_supply_ref represents the preset target value for supply air humidity; P_energy(k+i|k) represents the unit operating energy consumption predicted at time k+i based on the measurement value at time k.

5. The method according to claim 1, characterized in that, The method of using a predictive model to predict future operating conditions and solving for the optimal control sequence under satisfied constraints includes: Based on the actual output and predicted disturbance values ​​at the current moment, the output variables at the next N_p moments are predicted using a prediction model. Sequence quadratic programming or particle swarm optimization algorithms are then used to solve for the optimal control sequence at the next N_c moments, provided that the constraints are met. The control time domain length N_c does not exceed the prediction time domain length N_p.

6. The method according to claim 1, characterized in that, The first control command of the optimal control sequence is executed, and the actual output feedback is collected in the next cycle. The prediction model is corrected online based on the deviation between the actual and predicted outputs, including: The first control quantity of the optimal control sequence is output to the actuators of the magnetic levitation compressor, variable frequency fan and electronic expansion valve. In the next control cycle, the actual operating parameters are collected again, the deviation between the actual output and the predicted output is calculated, and the deviation is used to perform online parameter correction of the prediction model.

7. The method according to claim 1, characterized in that, The control variable constraints include magnetic levitation compressor speed constraints, variable frequency fan frequency constraints, and electronic expansion valve opening constraints; the output variable constraints include allowable fluctuation range constraints for supply air temperature and allowable fluctuation range constraints for supply air humidity.

8. The method according to claim 1, characterized in that, The method further includes a feedforward control step: Based on the historical data and trends of the external environmental parameters and cooling load, the trajectory of the change of disturbance variables in the future prediction time domain is predicted. The predicted change trajectory of the disturbance variable is input into the unit dynamic prediction model, so that the prediction model can respond in advance to the changes in the external environmental parameters and cooling load when solving the optimal control sequence, thereby realizing feedforward compensation control.

9. A control optimization system for a magnetic levitation direct expansion air conditioning unit based on the MPC algorithm, characterized in that, The system includes: The data acquisition module is used to collect unit operating status parameters, external environmental parameters, and target control parameters in real time. The module is used to build a dynamic prediction model for the unit based on the fusion of mechanism modeling and data-driven approach, and to establish the mapping relationship between control variables, disturbance variables and output variables; The configuration module is used to set the multi-objective rolling optimization function and the constraints for safe operation of the equipment; the constraints include control variable constraints and output variable constraints. The prediction module is used to predict future operating conditions using the prediction model and solve for the optimal control sequence under the constraints. The output correction module is used to execute the first control command of the optimal control sequence and collect the actual output feedback in the next cycle. Based on the deviation between the actual and the prediction, the prediction model is corrected online to realize closed-loop rolling optimization control.

10. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program that, when executed by the processor, implements the method as described in any one of claims 1 to 8.

Images