Energy-saving optimization control method of air conditioner rotating wheel dehumidification system based on mechanism model

Abstract

The invention discloses an energy-saving optimization control method for an air conditioner rotating wheel dehumidification system based on a mechanism model, and belongs to the technical field of air conditioner energy-saving control. The method specifically comprises the steps that the mechanism model is established and verified by collecting historical operation data; based on the verified model, predicting a system state by using a model prediction control method, and adaptively adjusting the model and the prediction control method by using an adaptive learning algorithm and a deep reinforcement learning algorithm; calculating an optimal control input through an optimization algorithm by using the adjusted model and a predictive control method in combination with a current state and load prediction; monitoring system operation data in real time, and correcting and optimizing the model by adopting an intelligent decision algorithm and a deep reinforcement learning algorithm; the optimal control input is applied to the system, the control strategy is implemented, the effect is evaluated, the control strategy is optimized, and efficient and intelligent dehumidification control is achieved.

Description

Technical Field

[0001] The invention belongs to the technical field of air-conditioning energy-saving control, and in particular to an energy-saving optimization control method for an air-conditioning rotary dehumidification system based on a mechanism model. Background Art

[0002] In industrial production and living environments, the control of air humidity is crucial. The air conditioning rotary dehumidification system is a common dehumidification equipment, widely used in industrial, commercial and civil buildings to adjust the air humidity. However, traditional control methods mainly rely on empirical rules or simple feedback control, which makes it difficult to achieve the optimal operation of the system, resulting in high energy consumption. The existing control methods mainly rely on empirical rules or simple feedback control, which makes it difficult to achieve the optimal operation of the system. Most of them are based on experience or simple set value control, and do not fully consider the complex relationship between the various parameters in the system and the changes in actual operating conditions, resulting in serious energy waste and high operating costs. With the increasing requirements for energy conservation and emission reduction, how to reduce system energy consumption by optimizing control methods has become an urgent problem to be solved.

[0003] For example, the Chinese patent with the authorization announcement number CN115325665B discloses a central air-conditioning energy-saving optimization control method, device and system, including: responding to a startup instruction to output a test instruction for selecting a system operation mode; responding to the test instruction to select an operation parameter combination in a preset database, and issuing a current round of test requests to the central air-conditioning system according to the operation parameter combination; receiving the energy efficiency test result obtained by the central air-conditioning system in response to the test request, and outputting the next round of test instructions based on the mapping relationship table between the energy efficiency test result and the preset detection mechanism; repeating multiple rounds of testing, when the number of the operation parameter combinations that have been run reaches a set ratio of the total number of all the operation parameter combinations in the database, outputting the operation parameter combination with the highest energy efficiency as the final test result and configuring the final test result to the central air-conditioning system. This technical solution has the effect of facilitating the central air-conditioning system to maintain efficient operation.

[0004] For example, a Chinese patent with authorization announcement number CN104633857B discloses an air conditioning energy-saving optimization control method and device, including: collecting and storing the current operating parameters of the air conditioning system's chiller unit, chilled water pump, and cooling water pump according to a preset cycle; determining the energy consumption increment of the air conditioning system under the operating state of the operating parameters to be adjusted relative to the current operating state according to the current operating parameters and the operating parameters to be adjusted input by the user; when the energy consumption increment is less than zero, adjusting the operating state of the air conditioning system according to the operating parameters to be adjusted. The air conditioning control method and device provided by this technical solution adjusts the air conditioning system according to the operating parameters to be adjusted when the energy consumption increment is less than zero, thereby achieving energy saving, which is easy to implement.

[0005] The above existing technologies have the following problems: lack of utilization of the inherent physical laws of the system and lack of consideration of the comprehensive physical characteristics of the system; reliance on the operating parameter combination in the preset database for testing and lack of real-time adaptability; lack of dynamic adjustment capability and long-term optimization capability, and will cause energy consumption and waste of resources. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention proposes an energy-saving optimization control method for an air-conditioning rotary dehumidification system based on a mechanism model. By collecting historical operation data, the mechanism model is established and verified; based on the verified model, the system state is predicted by the model predictive control method, and the model and the predictive control method are adaptively adjusted using an adaptive learning algorithm and a deep reinforcement learning algorithm; the adjusted model and predictive control method are used, combined with the current state and load forecast, to calculate the optimal control input through an optimization algorithm; the system operation data is monitored in real time, and the model is corrected and optimized using an intelligent decision-making algorithm and a deep reinforcement learning algorithm; the optimal control input is applied to the system, the control strategy is implemented, the effect is evaluated and the control strategy is optimized, so as to achieve efficient and intelligent dehumidification control.

[0007] To achieve the above object, the present invention provides the following technical solutions: The energy-saving optimization control method of the air-conditioning rotary dehumidification system based on the mechanism model includes: Step S1: collecting historical operation data of the air-conditioning rotary dehumidification system, establishing a mechanism model of the air-conditioning rotary dehumidification system, and verifying the mechanism model; Step S2: Based on the verified mechanism model, a model predictive control method is used to predict the operating state of the air-conditioning rotary dehumidification system, and according to the prediction results and the real-time collected operating data, an adaptive learning algorithm and a deep reinforcement learning algorithm are used to adaptively adjust the mechanism model and the model predictive control method; Step S3: Based on the adaptively adjusted mechanism model, the current operating state and load of the air-conditioning rotary dehumidification system are predicted, and the optimal control input is calculated through the optimization algorithm; Step S4: using intelligent decision-making algorithms and deep reinforcement learning algorithms to correct and iterate the adaptively adjusted mechanism model; Step S5: Apply the calculated optimal control input to the air-conditioning rotary dehumidification system, implement the control strategy, monitor the operating status and energy consumption of the system in real time, evaluate the effect of the control strategy, and adjust and optimize the control strategy based on the evaluation results.

[0008] Specifically, the specific steps of establishing the mechanism model of the air-conditioning rotary dehumidification system in step S1 include: S1.1: Collect historical operation data of the air-conditioning rotary dehumidification system and perform preprocessing; S1.2: Determine the scope of the air conditioning rotary dehumidification system and analyze the functions of each component; the components include rotary dehumidifier, regenerative heating device, fan, and air handling channel; S1.3: Based on the analysis results of S1.2, considering the adsorption characteristics of the desiccant, mass transfer and heat transfer process, establish the wheel dehumidification and absorption model The wheel moisture absorption and dehumidification model includes an adsorption isotherm model and a mass transfer and heat transfer model, wherein q represents the adsorption amount of the moisture absorbent, represents the maximum adsorption capacity, K represents the adsorption equilibrium constant, is the water vapor partial pressure, is the water vapor density, represents the volume of the control volume, and denote the water vapor mass flow rates entering and leaving the control volume, respectively, represents the adsorption mass flow rate of the desiccant, t represents the time, represents the air density, represents the constant pressure specific heat capacity of air, T represents the air temperature, and denote the heat fluxes entering and leaving the control volume, respectively, is the adsorption heat flow, represents partial derivative; S1.4: Similarly, based on the analysis results of S1.2 and the thermodynamic properties of air, an air heat and moisture exchange model is established. ,in, represents the air mass flow rate, and Respectively represent the moisture content of the air when it enters and leaves the rotor, represents the change in the mass flow rate of water vapor in the air, and are the enthalpy values ​​of air entering and leaving the rotor, represents the sensible heat exchange capacity, represents the amount of latent heat exchange; S1.5: Establish component characteristic model based on the fan and heating device ,in, represents the wind pressure of the fan, Q represents the wind volume of the fan, , , represents the fitting coefficient, represents heating power, V represents voltage, and I represents current.

[0009] Specifically, the specific steps of establishing the mechanism model of the air-conditioning rotary dehumidification system in step S1 also include: S1.6: Integrate the wheel moisture absorption and dehumidification model, air heat and moisture exchange model and component characteristic model to form a mechanism model of the air conditioning wheel dehumidification system; S1.7: Use the finite difference method to discretize the mechanism model equation of the air-conditioning rotary dehumidification system into a difference equation and solve it to obtain the operating state parameters of the air-conditioning rotary dehumidification system under different working conditions; S1.8: Verify the output result of the mechanism model of the air-conditioning rotary dehumidification system according to the preprocessed historical operation data, and calculate the mean square error index by comparing the prediction result of the mechanism model with the actual operation data; If the mean square error index is less than or equal to the preset threshold, the mechanism model is considered valid; If the mean square error index is greater than the preset threshold, the mechanism model is adjusted.

[0010] Specifically, the specific steps of S1.7 include: S1.71: Obtain the mechanism model of the air-conditioning rotary dehumidification system, and discretize it in the space domain and time domain to generate the mechanism model equation; the space domain is to divide the solution area of ​​the rotary dehumidification system into a series of uniform grid points; the time domain is to divide it into discrete time steps; S1.72: For each grid cell, transform the mechanism model equation into a difference equation using the forward difference format; S1.73: The discretized difference equations are organized into a system of linear algebraic equations, and the Jacobi iteration method is used to solve the system of linear algebraic equations to obtain the variable values ​​of each grid node at different time steps, that is, the operating state parameters of the system.

[0011] Specifically, the specific steps of step S2 include: S2.1: Setting the prediction horizon and control time domain , and according to the operation target of the air-conditioning rotary dehumidification system, the objective function J is established, and the formula is: ; in, represents the system state at time k+1 predicted based on the information at time k, represents the state variable of the air-conditioning rotary dehumidification system at time k, E represents the state error, represents the control input at time k, and R represents the weighted matrix of the control input; S2.2: Using the verified mechanism model, combined with the current operating status of the air-conditioning rotary dehumidification system and the interference factor results, predict the operating status of the air-conditioning rotary dehumidification system within the set prediction time domain; S2.3: Use sensors to collect the operating data of the air-conditioning rotary dehumidification system in real time, compare the operating status in the time domain in S2.2 with the real-time collected data, and calculate the error between the two; S2.4: According to the error results, the parameters of the mechanism model are adjusted using an adaptive learning algorithm based on the least squares method.

[0012] Specifically, the specific steps of step S2 also include: S2.5: The operating state of the air-conditioning rotary dehumidification system is used as the state space, the control input is used as the action space, and the reward function is defined according to the operating goal of the system. ; The control input includes the wheel speed and heating power; the operation target includes energy saving and humidity control accuracy, where z represents the air-conditioning wheel dehumidification system, g represents the control action, and represents the weighting coefficient, G represents the system energy consumption, hum represents the operating status of the air-conditioning rotary dehumidification system, Indicates the operating status reference value of the air-conditioning rotary dehumidification system; S2.6: Use deep reinforcement learning algorithms to learn the optimal control strategy by interacting with the air conditioning rotary dehumidification system; S2.7: Update the control strategy in the model predictive control method based on the optimal strategy learned by the deep reinforcement learning model.

[0013] Specifically, the specific steps of step S3 include: S3.1: Obtain the adaptively adjusted mechanism model and current operating data of the air-conditioning rotary dehumidification system; S3.2: Based on the historical operation data, use a time series analysis method to predict the load of the air-conditioning rotary dehumidification system; the load includes dehumidification load, heating load, and ventilation load; S3.3: Determine the constraints based on the operating requirements of the air conditioning rotary dehumidification system ,in, and Respectively represent the lower and upper limits of the wheel speed, represents the wheel speed at time k, and Respectively represent the lower and upper limits of the heating power, represents the heating power at time k, and Respectively represent the lower and upper limits of air humidity. represents the air humidity at time k.

[0014] Specifically, the specific steps of step S3 also include: S3.4: Substitute the current operating state, load forecast results, objective function and constraint conditions into the genetic algorithm, and the genetic algorithm searches for a control input combination that makes the objective function reach the optimal value through iterative calculation; the control input includes the rotation speed of the rotor, the power of the heating device, and the air volume of the fan; S3.5: Verify the calculated optimal control input to check whether it meets the physical feasibility and operation requirements of the system; If the optimal control input does not satisfy the constraints or causes unstable system operation, the parameters of the genetic algorithm are readjusted and the calculation is performed again.

[0015] Specifically, the model predictive control algorithm predicts the operating results of the air-conditioning rotary dehumidification system under different control strategies based on the current system operating status and load prediction, and obtains the optimal control parameters through optimization calculation, including the rotary wheel speed, heating power, and fan frequency.

[0016] Specifically, the mechanism model includes the adsorption and desorption process of the wheel and the heat exchange process of the air flow; the historical operation data of the air-conditioning rotary dehumidification system includes indoor and outdoor air humidity, temperature, air volume and energy consumption; the optimal control input includes the optimal speed of the wheel, the optimal power of the heating device and the optimal frequency of the fan.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention proposes an energy-saving optimization control method for an air-conditioning rotary dehumidification system based on a mechanism model. The mechanism model is established by collecting historical operation data, and the system state is predicted by a model predictive control method. The model and the predictive control method are adaptively adjusted in combination with an adaptive learning algorithm and a deep reinforcement learning algorithm, thereby achieving efficient and precise control of the air-conditioning rotary dehumidification system. This can not only improve the operating efficiency of the system, but also effectively reduce energy consumption and achieve the goal of energy conservation and emission reduction.

[0018] The present invention proposes an energy-saving optimization control method for an air-conditioning rotary dehumidification system based on a mechanism model. It also monitors the system operation data in real time, and uses an intelligent decision-making algorithm and a deep reinforcement learning algorithm to correct and iterate the mechanism model to ensure that the model always reflects the latest status of the system. This real-time feedback and iterative optimization mechanism enables the control strategy to continuously adapt to changes in the system environment, further improving the stability and reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of the energy-saving optimization control method of the air-conditioning rotary dehumidification system based on the mechanism model of the present invention; Figure 2 It is a principle flow chart of the energy-saving optimization control method of the air-conditioning rotary dehumidification system based on the mechanism model of the present invention; Figure 3 The present invention is a flow chart of the mechanism model construction of the energy-saving optimization control method of the air-conditioning rotary dehumidification system based on the mechanism model. DETAILED DESCRIPTION

[0020] Example 1 See also Figure 1-Figure 2 The present invention provides an embodiment: an energy-saving optimization control method for an air-conditioning rotary dehumidification system based on a mechanism model, comprising the following steps: Step S1: collecting historical operation data of the air-conditioning rotary dehumidification system, establishing a mechanism model of the air-conditioning rotary dehumidification system, and verifying the mechanism model; Furthermore, the process of collecting historical operation data includes: (1) Data collection: Determine the type and scope of data that needs to be collected, such as temperature, humidity, pressure, flow, and key energy consumption parameters; Select appropriate data collection equipment and sensors to ensure data accuracy and reliability; Set the frequency and time range of data collection to obtain sufficient historical operation data.

[0021] (2) Data preprocessing: Clean the collected data to remove outliers and noise; Normalize the data to make different parameters comparable; Sort the data into time series to facilitate subsequent modeling and analysis.

[0022] Step S2: Based on the verified mechanism model, a model predictive control method is used to predict the operating state of the air-conditioning rotary dehumidification system, and according to the prediction results and the real-time collected operating data, an adaptive learning algorithm and a deep reinforcement learning algorithm are used to dynamically adjust the mechanism model and the model predictive control method; Step S3: Based on the adaptively adjusted mechanism model, the current operating state and load of the air-conditioning rotary dehumidification system are predicted, and the optimal control input is calculated through the optimization algorithm; Step S4: using intelligent decision-making algorithms and deep reinforcement learning algorithms to correct and iterate the adaptively adjusted mechanism model; Furthermore, the specific steps of step S4 include: (1) Obtaining a mechanism model after adaptive adjustment; (2) Collect and preprocess system operation data; (3) Using a fuzzy logic algorithm to make a preliminary correction to the adaptively adjusted mechanism model, and at the same time, using a deep reinforcement learning algorithm to further optimize the mechanism model and update the parameters. The fuzzy logic algorithm is a prior art in this field and is not an inventive solution of the present application, and will not be described in detail here. (4) Evaluate the performance of the modified and optimized mechanism model and calculate the error index. In the present invention, the error index adopts the mean square error, which is the prior art content in this field and is not the inventive solution of the present application, and will not be described in detail here; (5) Determine whether the termination condition is met. If not, return to (3) and continue iterating; (6) Output the corrected and optimized mechanism model.

[0023] Step S5: Apply the calculated optimal control input to the air-conditioning rotary dehumidification system, implement the control strategy, monitor the operating status and energy consumption of the system in real time, evaluate the effect of the control strategy, and adjust and optimize the control strategy based on the evaluation results.

[0024] Furthermore, the specific steps of step S5 include: (1) Converting the calculated optimal control input, such as the rotor speed, heating power, and fan air volume, into a control signal for an actuator of the air-conditioning rotor dehumidification system, for example, converting a digital signal into an analog voltage or current signal, wherein the digital-to-analog conversion is a prior art in the art and is not an inventive solution of the present application and is not described in detail herein; (2) transmitting control signals to actuators of the air-conditioning rotary dehumidification system, such as a frequency converter and a heater controller, through a control system so that the system operates according to the optimal control input; the frequency converter is used to adjust the rotary speed and the fan air volume; the heater controller is used to adjust the heating power; (3) Acquire real-time operation data, transmit the collected real-time operation data to the monitoring center through the wireless communication network, and store it in real time for subsequent analysis and processing; (4) According to the system's operating objectives, set indicators for evaluating the effectiveness of the control strategy, including energy consumption reduction rate, humidity control accuracy, and temperature stability; (5) Calculate the values ​​of various evaluation indicators based on real-time monitoring data. For example, the energy consumption reduction rate can be calculated by comparing the current energy consumption with the historical energy consumption or the benchmark energy consumption; the humidity control accuracy can be evaluated by calculating the deviation between the actual humidity and the set humidity; (6) Compare the calculated evaluation index value with the pre-set expected target to determine whether the control strategy has achieved the expected effect; If there is a deviation between the evaluation index value and the expected target, analyze the reasons for the deviation, which may include inaccurate model, changes in external interference factors, and actuator failure; (7) Adjust the control strategy according to the cause of the deviation. For example, if the energy consumption reduction rate does not reach the expected target, the heating power can be appropriately reduced or the wheel speed can be adjusted; if the humidity control accuracy is not high, the dehumidification time can be increased or the ventilation volume can be adjusted; (8) Using the adjusted control strategy, combined with the current system operating status and load forecast, recalculate the optimal control input to achieve optimization of the control strategy.

[0025] For example, the drug production workshop of a large pharmaceutical factory has extremely strict requirements on air humidity. The humidity needs to be controlled between 30%-40%RH to ensure the stability of drug quality. The original air-conditioning rotary dehumidification system adopts traditional control methods, which has high energy consumption and insufficient humidity control accuracy. Humidity fluctuations often exceed the allowable range. To solve these problems, the energy-saving optimization control method based on the mechanism model is adopted to upgrade the system, including: Over a one-month period, high-precision sensors were used to collect real-time historical operating data of the air-conditioning rotary dehumidification system, including indoor and outdoor air temperature, humidity, rotary speed, regeneration air temperature and flow, fan operating frequency, and system energy consumption. The collection frequency was set to once per minute to ensure data integrity and accuracy. The system scope was determined, including components such as rotary dehumidifiers, regeneration heating devices, fans, and air handling channels. The adsorption characteristics of the desiccant and the mass and heat transfer processes were considered, and the Langmuir adsorption isotherm model and mass and energy conservation equations were used to establish a rotary dehumidification model. According to the thermodynamic properties of air, An air heat and moisture exchange model is established to describe the temperature and humidity changes of the air in the system. Performance characteristic models are established for the fan and heating device, such as the air volume-air pressure-power relationship model of the fan and the relationship model between the heating power and temperature and energy consumption of the heating device. The above sub-models are integrated to form a complete mechanism model of the air-conditioning rotary dehumidification system. The mechanism model is verified using the historical operation data collected in the second half of the month. By comparing the operating status parameters predicted by the model, such as air humidity and temperature, with the actual measured values, the mean square error and mean absolute error are calculated. After multiple parameter adjustments and optimizations, the error is controlled within 5%, indicating that the mechanism model The model has high accuracy and reliability; based on the verified mechanism model, the model predictive control method is adopted, with the next 24 hours as the prediction time domain, to predict the operating status of the air conditioning rotary dehumidification system, taking into account the production plan and weather changes of the pharmaceutical production workshop and other factors, to predict parameters such as air humidity, temperature and energy consumption at different times; the system operation data is collected in real time, the prediction results are compared with the real-time data, and the adaptive learning algorithm is used to dynamically adjust the parameters of the mechanism model according to the error between the two to improve the adaptability of the model. At the same time, the deep reinforcement learning algorithm is introduced to continuously optimize the control strategy of the model predictive control method through the reward mechanism, so that the system The system can better cope with various complex working conditions; combined with the production plan of the pharmaceutical production workshop, the flow of personnel and outdoor weather changes and other factors, the load of the air-conditioning rotary dehumidification system is predicted. For example, during the peak production period, the heat and humidity generated by the personnel and equipment in the workshop increase, and the system load increases accordingly. During the rest period at night, the load is relatively low; based on the adaptively adjusted mechanism model and model predictive control method, the genetic algorithm is used as the optimization algorithm, with the minimum system energy consumption as the objective function. According to the current operating status and load forecast, the optimal control input is calculated, including the optimal speed of the rotary wheel, the optimal power of the heating device, the optimal frequency of the fan, etc.; Install multiple high-precision sensors at key locations of the air-conditioning rotary dehumidification system to monitor the system's operating status in real time, including air temperature, humidity, flow, rotary speed, temperature and power of the heating device and other parameters. The sensor data is updated every 10 seconds to ensure that the real-time changes of the system can be reflected in a timely manner. The real-time monitored operating data is input into the mechanism model, and an intelligent decision-making algorithm is used to determine whether the mechanism model needs to be corrected based on the data changes. At the same time, the deep reinforcement learning algorithm is used to iteratively update the mechanism model through continuous learning and optimization, so that the mechanism model can more accurately reflect the actual operation of the system. The calculated optimal control input is sent to the execution of the air-conditioning rotary dehumidification system. Mechanisms, such as adjusting the speed of the rotary motor, the power of the heating device, the frequency of the fan, etc., are used to implement the control strategy; the operating status and energy consumption of the system are monitored in real time, the air humidity control accuracy and system energy consumption before and after the implementation of the control strategy are compared, and evaluation indicators, such as humidity control error rate and energy consumption reduction rate, are set. If the humidity control error rate exceeds 3% or the energy consumption reduction rate does not reach the expected target, then according to the evaluation results, the parameters in the control strategy are readjusted, such as adjusting the prediction time domain of the model predictive control method, the parameters of the optimization algorithm, etc., for further optimization; after running for a predetermined time, the energy-saving optimization control method of the air-conditioning rotary dehumidification system has achieved good results, the air humidity control accuracy has been significantly improved, and the humidity fluctuation range is controlled within ±1% RH, which meets the strict requirements of the pharmaceutical production workshop. At the same time, the system energy consumption is reduced by more than 25%, reducing production costs.

[0026] Example 2 See also Figure 3 In this embodiment, the specific steps of establishing the mechanism model of the air-conditioning rotary dehumidification system in step S1 include: S1.1: Collect historical operation data of the air-conditioning rotary dehumidification system and perform preprocessing; S1.2: Determine the scope of the air conditioning rotary dehumidification system and analyze the functions of each component; the components include rotary dehumidifier, regenerative heating device, fan, and air handling channel; Furthermore, analyzing the functions of each component means understanding the role of each component in the dehumidification process. For example, the rotary dehumidifier achieves dehumidification by absorbing moisture through the desiccant, and the regeneration heating device provides the heat required for the regeneration of the rotary wheel.

[0027] S1.3: Based on the analysis results of S1.2, considering the adsorption characteristics of the desiccant, mass transfer and heat transfer process, establish the wheel dehumidification and absorption model The wheel moisture absorption and dehumidification model includes an adsorption isotherm model and a mass transfer and heat transfer model, wherein q represents the adsorption amount of the moisture absorbent, represents the maximum adsorption capacity, K represents the adsorption equilibrium constant, is the water vapor partial pressure, is the water vapor density, represents the volume of the control volume, and denote the water vapor mass flow rates entering and leaving the control volume, respectively, represents the adsorption mass flow rate of the desiccant, t represents the time, represents the air density, represents the constant pressure specific heat capacity of air, T represents the air temperature, and denote the heat fluxes entering and leaving the control volume, respectively, is the adsorption heat flow, represents partial derivative; It should be noted that It represents the rate of change of the energy of the air in the control body over time, that is, it reflects the change of the energy in the control body over time due to the combined factors of air density, constant pressure specific heat capacity, temperature and control body volume at different times.

[0028] S1.4: Similarly, based on the analysis results of S1.2 and the thermodynamic properties of air, an air heat and moisture exchange model is established. ,in, represents the air mass flow rate, and Respectively represent the moisture content of the air when it enters and leaves the rotor, represents the change in the mass flow rate of water vapor in the air, and are the enthalpy values ​​of air entering and leaving the rotor, represents the sensible heat exchange capacity, represents the amount of latent heat exchange; S1.5: Establish component characteristic model based on the fan and heating device ,in, represents the wind pressure of the fan, Q represents the wind volume of the fan, , , represents the fitting coefficient, Indicates heating power, V indicates voltage, and I indicates current; S1.6: Integrate the wheel moisture absorption and dehumidification model, air heat and moisture exchange model and component characteristic model to form a mechanism model of the air conditioning wheel dehumidification system; S1.7: Use the finite difference method to discretize the mechanism model equation of the air-conditioning rotary dehumidification system into a difference equation and solve it to obtain the operating state parameters of the air-conditioning rotary dehumidification system under different working conditions; S1.8: Verify the output result of the mechanism model of the air-conditioning rotary dehumidification system according to the pre-processed historical operation data, and calculate the mean square error index by comparing the prediction result of the mechanism model with the actual operation data, wherein the mean square error calculation formula is the prior art content in this field, is not the inventive solution of the present application, and is not described in detail here; If the mean square error index is less than or equal to the preset threshold, the mechanism model is considered valid; If the mean square error index is greater than the preset threshold, the mechanism model is adjusted.

[0029] The specific steps of S1.7 include: S1.71: Obtain the mechanism model of the air-conditioning rotary dehumidification system, and discretize it in the space domain and time domain to generate the mechanism model equation; the space domain is to divide the solution area of ​​the rotary dehumidification system into a series of uniform grid points; the time domain is to divide it into discrete time steps; Furthermore, the specific steps of S1.71 include: (1) Sorting out the working principles and interaction relationships of various components in the air-conditioning rotary dehumidification system, and summarizing the application equations of mass conservation, energy conservation, mass transfer and heat transfer to form a complete mechanism model; (2) According to the geometry of the system and the computational requirements, the spatial area of ​​the entire system is divided into multiple small control volumes or grid units. For example, for a rotary dehumidifier, it can be divided according to the axial direction, radial direction, etc.; for an air treatment channel, it can be divided according to its length, cross section, etc.; (3) Determine the nodes in each control volume or grid. These nodes will be used to describe physical quantities, such as temperature, humidity, and pressure. At the same time, clarify the boundaries of the system, such as import and export boundaries, interfaces with other components, etc., and determine the values ​​of physical quantities on the boundaries or the types of boundary conditions; (4) Determine the time step according to the dynamic characteristics and calculation accuracy requirements of the air-conditioning rotary dehumidification system , where a time step that is too small will increase the amount of calculation, while a time step that is too large may affect the calculation accuracy and stability; (5) Divide the entire time process into n discrete time nodes ; (6) For each discrete control volume and time node, the algebraic equations of mass conservation, energy conservation, and mass and heat transfer are organized to form a set of equations containing all unknown quantities, that is, a set of equations for the physical quantities of each node, also called the discretized mechanism model equations.

[0030] S1.72: For each grid cell, transform the mechanism model equation into a difference equation using the forward difference format; Furthermore, the specific steps of S1.72 include: (1) Obtaining the discretized mechanism model equation; (2) Divide the entire computational area into grids, determine the number and position of each grid cell, and at the same time, clearly identify the variables that need to be solved in each grid cell, such as temperature, humidity, etc., and mark the values ​​of the variables at the grid nodes; (3) For the time derivative term in the mechanism model equation , discretized using the forward difference format, where the forward difference formula is: ,in, Indicates the variables to be determined, such as temperature, humidity, etc. and Respectively represent the values ​​of the variable at the current time step n and the next time step n+1, and Respectively represent variables The values ​​at two adjacent grid nodes i and i+1, represents the spatial grid spacing, and x represents the number of spatial grids; (4) The time derivative and space derivative in the mechanism model equation are replaced by the forward difference format to obtain the difference equation about the variable value at the grid node.

[0031] S1.73: The discretized differential equations are organized into the form of a system of linear algebraic equations, and the Jacobi iteration method is used to solve the system of linear algebraic equations to obtain the variable values ​​of each grid node at different time steps, that is, the operating state parameters of the system. The process of solving the system of linear algebraic equations by the Jacobi iteration method is the prior art content in this field, and is not the inventive solution of the present application, and will not be elaborated here.

[0032] The specific steps of step S2 include: S2.1: Setting the prediction horizon and control time domain , and according to the operation target of the air-conditioning rotary dehumidification system, the objective function J is established, and the formula is: ; in, represents the system state at time k+1 predicted based on the information at time k, represents the state variable of the air-conditioning rotary dehumidification system at time k, E represents the state error, represents the control input at time k, and R represents the weighted matrix of the control input; S2.2: Using the verified mechanism model, combined with the current operating status of the air-conditioning rotary dehumidification system and the interference factor results, predict the operating status of the air-conditioning rotary dehumidification system within the set prediction time domain; Furthermore, the specific steps of S2.2 include: (1) Obtain a verified mechanism model; (2) Use various sensors to collect the current operating data of the air-conditioning rotary dehumidification system in real time; (3) Identify interference factors that affect the operation of the air-conditioning rotary dehumidification system and obtain the predicted values ​​of outdoor environmental parameters through weather forecasts; the interference factors include changes in outdoor temperature, humidity, wind speed, and indoor human activities and equipment heat dissipation; (4) Obtaining the predicted time domain ; (5) Use the current operating data and interference factor results as input and substitute them into the verified mechanism model; (6) The calculation is gradually carried out in the prediction time domain according to the time step, and the system state variable values ​​of each time step are calculated in sequence using the equations and parameters of the mechanism model; (7) Post-process the predicted operating data, such as visualizing the data, analyzing the changing trends and characteristics of the data, and outputting the prediction results.

[0033] S2.3: Use sensors to collect the operating data of the air-conditioning rotary dehumidification system in real time, compare the operating status in the time domain in S2.2 with the real-time collected data, and calculate the error between the two; S2.4: According to the error result, the parameters of the mechanism model are adjusted using an adaptive learning algorithm based on the least squares method, wherein the least squares method is a prior art content in this field and is not an inventive solution of the present application, and is not described in detail here; S2.5: The operating state of the air-conditioning rotary dehumidification system is used as the state space, the control input is used as the action space, and the reward function is defined according to the operating goal of the system. ; The control input includes the wheel speed and heating power; the operation target includes energy saving and humidity control accuracy, where z represents the state of the air-conditioning wheel dehumidification system, g represents the control action, and represents the weighting coefficient, G represents the system energy consumption, hum represents the operating status of the air-conditioning rotary dehumidification system, Indicates the operating status reference value of the air-conditioning rotary dehumidification system; S2.6: Use deep reinforcement learning algorithms to learn the optimal control strategy by interacting with the air conditioning rotary dehumidification system; Furthermore, the specific steps of S2.6 include: (1) Obtain state space, action space, and reward function; (2) The air-conditioning rotary dehumidification system is regarded as an environment that generates corresponding feedback according to the actions of the intelligent agent; the intelligent agent is the controller; (3) Construct the state space based on the operating state parameters of the air-conditioning rotary dehumidification system; (4) Determine the control actions that the agent can take. In the present invention, for the air-conditioning rotary dehumidification system, the actions usually include adjusting the rotary speed of the rotary wheel, changing the power of the heating device, and adjusting the air volume of the fan. The action space is discrete, for example, the rotary speed of the rotary wheel is divided into several fixed gears. (5) Based on the characteristics and requirements of the problem, a deep reinforcement learning algorithm based on a deep Q network is used to allow the agent to interact and learn with the environment: The agent is based on the current state , using a greedy strategy to select an action ; Execute an action After that, the environment generates the next state according to the physical laws of the system. and rewards ; The sample Store into the experience replay buffer; According to the selected deep reinforcement learning algorithm based on deep Q network, the parameters of the neural network are updated using the stored samples to optimize the strategy; (6) When the termination condition is met, the training is stopped to obtain the optimal control strategy learned; the termination condition is a preset number of training steps.

[0034] S2.7: Update the control strategy in the model predictive control method based on the optimal strategy learned by the deep reinforcement learning model.

[0035] The specific steps of step S3 include: S3.1: Obtain the adaptively adjusted mechanism model and current operating data of the air-conditioning rotary dehumidification system; S3.2: Based on the historical operation data, a time series analysis method is used to predict the load of the air-conditioning rotary dehumidification system; the load includes a dehumidification load, a heating load, and a ventilation load. The time series analysis method is a prior art in the art and is not an inventive solution of the present application, and is not described in detail here; S3.3: Determine the constraints based on the operating requirements of the air conditioning rotary dehumidification system ,in, and Respectively represent the lower and upper limits of the wheel speed, represents the wheel speed at time k, and Respectively represent the lower and upper limits of the heating power, represents the heating power at time k, and Respectively represent the lower and upper limits of air humidity. represents the air humidity at time k; S3.4: Substitute the current operating state, load forecast results, objective function and constraints into the genetic algorithm, and the genetic algorithm searches for a control input combination that makes the objective function reach the optimal value through iterative calculation; the control input includes the rotation speed of the rotor, the power of the heating device, and the air volume of the fan, wherein the genetic algorithm is the prior art content in this field, and is not the inventive solution of the present application, and is not described in detail here; S3.5: Verify the calculated optimal control input to check whether it meets the physical feasibility and operation requirements of the system; If the optimal control input does not satisfy the constraints or causes unstable system operation, the parameters of the genetic algorithm are readjusted and the calculation is performed again.

[0036] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the purpose and scope of protection of the present invention, and all of these are within the protection of the present invention.

Claims

1. An energy-saving optimization control method for an air-conditioning rotary dehumidification system based on a mechanism model, characterized in that: include: Step S1: collecting historical operation data of the air-conditioning rotary dehumidification system, establishing a mechanism model of the air-conditioning rotary dehumidification system, and verifying the mechanism model; Step S2: Based on the verified mechanism model, a model predictive control method is used to predict the operating state of the air-conditioning rotary dehumidification system, and according to the prediction results and the real-time collected operating data, an adaptive learning algorithm and a deep reinforcement learning algorithm are used to adaptively adjust the mechanism model and the model predictive control method; Step S3: Based on the adaptively adjusted mechanism model, the current operating state and load of the air-conditioning rotary dehumidification system are predicted, and the optimal control input is calculated through the optimization algorithm; Step S4: using intelligent decision-making algorithms and deep reinforcement learning algorithms to correct and iterate the adaptively adjusted mechanism model; Step S5: Apply the calculated optimal control input to the air-conditioning rotary dehumidification system, implement the control strategy, monitor the operating status and energy consumption of the system in real time, evaluate the effect of the control strategy, and adjust and optimize the control strategy based on the evaluation results.

2. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 1 is characterized in that: The specific steps of establishing the mechanism model of the air-conditioning rotary dehumidification system in step S1 include: S1.1: Collect historical operation data of the air-conditioning rotary dehumidification system and perform preprocessing; S1.2: Determine the scope of the air conditioning rotary dehumidification system and analyze the functions of each component; the components include rotary dehumidifier, regenerative heating device, fan, and air handling channel; S1.3: Based on the analysis results of S1.2, considering the adsorption characteristics of the desiccant, mass transfer and heat transfer process, establish the wheel dehumidification and absorption model The wheel moisture absorption and dehumidification model includes an adsorption isotherm model and a mass transfer and heat transfer model, wherein q represents the adsorption amount of the moisture absorbent, represents the maximum adsorption capacity, K represents the adsorption equilibrium constant, is the water vapor partial pressure, is the water vapor density, represents the volume of the control volume, and denote the water vapor mass flow rates entering and leaving the control volume, respectively, represents the adsorption mass flow rate of the desiccant, t represents the time, represents the air density, represents the constant pressure specific heat capacity of air, T represents the air temperature, and denote the heat fluxes entering and leaving the control volume, respectively, is the adsorption heat flow, represents partial derivative; S1.4: Similarly, based on the analysis results of S1.2 and the thermodynamic properties of air, an air heat and moisture exchange model is established. ,in, represents the air mass flow rate, and Respectively represent the moisture content of the air when it enters and leaves the rotor, represents the change in the mass flow rate of water vapor in the air, and are the enthalpy values ​​of air entering and leaving the rotor, represents the sensible heat exchange capacity, represents the amount of latent heat exchange; S1.5: Establish component characteristic model based on the fan and heating device ,in, represents the wind pressure of the fan, Q represents the wind volume of the fan, , , represents the fitting coefficient, represents heating power, V represents voltage, and I represents current.

3. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 2 is characterized in that: The specific steps of establishing the mechanism model of the air-conditioning rotary dehumidification system in step S1 also include: S1.6: Integrate the wheel moisture absorption and dehumidification model, air heat and moisture exchange model and component characteristic model to form a mechanism model of the air conditioning wheel dehumidification system; S1.7: Use the finite difference method to discretize the mechanism model equation of the air-conditioning rotary dehumidification system into a difference equation and solve it to obtain the operating state parameters of the air-conditioning rotary dehumidification system under different working conditions; S1.8: Verify the output result of the mechanism model of the air-conditioning rotary dehumidification system according to the preprocessed historical operation data, and calculate the mean square error index by comparing the prediction result of the mechanism model with the actual operation data; If the mean square error index is less than or equal to the preset threshold, the mechanism model is considered valid; If the mean square error index is greater than the preset threshold, the mechanism model is adjusted.

4. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 3 is characterized in that: The specific steps of S1.7 include: S1.71: Obtain the mechanism model of the air-conditioning rotary dehumidification system, and discretize it in the space domain and time domain to generate the mechanism model equation; the space domain is to divide the solution area of ​​the rotary dehumidification system into a series of uniform grid points; the time domain is to divide it into discrete time steps; S1.72: For each grid cell, transform the mechanism model equation into a difference equation using the forward difference format; S1.73: The discretized difference equations are organized into a system of linear algebraic equations, and the Jacobi iteration method is used to solve the system of linear algebraic equations to obtain the variable values ​​of each grid node at different time steps, that is, the operating state parameters of the system.

5. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 4 is characterized in that: The specific steps of step S2 include: S2.1: Setting the prediction horizon and control time domain , and according to the operation target of the air-conditioning rotary dehumidification system, the objective function J is established, and the formula is: ； in, represents the system state at time k+1 predicted based on the information at time k, represents the state variable of the air-conditioning rotary dehumidification system at time k, E represents the state error, represents the control input at time k, and R represents the weighted matrix of the control input; S2.2: Using the verified mechanism model, combined with the current operating status of the air-conditioning rotary dehumidification system and the interference factor results, predict the operating status of the air-conditioning rotary dehumidification system within the set prediction time domain; S2.3: Use sensors to collect the operating data of the air-conditioning rotary dehumidification system in real time, compare the operating status in the time domain in S2.2 with the operating data collected in real time, and calculate the error between the two; S2.4: According to the error results, the parameters of the mechanism model are adjusted using an adaptive learning algorithm based on the least squares method.

6. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 5 is characterized in that: The specific steps of step S2 also include: S2.5: The operating state of the air-conditioning rotary dehumidification system is used as the state space, the control input is used as the action space, and the reward function is defined according to the operating goal of the system. ; The control input includes the wheel speed and heating power; the operation target includes energy saving and humidity control accuracy, where z represents the air-conditioning wheel dehumidification system, g represents the control action, and represents the weighting coefficient, G represents the system energy consumption, hum represents the operating status of the air-conditioning rotary dehumidification system, Indicates the operating status reference value of the air-conditioning rotary dehumidification system; S2.6: Use deep reinforcement learning algorithms to learn the optimal control strategy by interacting with the air conditioning rotary dehumidification system; S2.7: Update the control strategy in the model predictive control method based on the optimal strategy learned by the deep reinforcement learning model.

7. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 6 is characterized in that: The specific steps of step S3 include: S3.1: Obtain the adaptively adjusted mechanism model and current operating data of the air-conditioning rotary dehumidification system; S3.2: Based on the historical operation data, use a time series analysis method to predict the load of the air-conditioning rotary dehumidification system; the load includes dehumidification load, heating load, and ventilation load; S3.3: Determine the constraints based on the operating requirements of the air conditioning rotary dehumidification system ,in, and Respectively represent the lower and upper limits of the wheel speed, represents the wheel speed at time k, and Respectively represent the lower and upper limits of the heating power, represents the heating power at time k, and Respectively represent the lower and upper limits of air humidity. represents the air humidity at time k.

8. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 7 is characterized in that: The specific steps of step S3 also include: S3.4: Substitute the current operating state, load forecast results, objective function and constraint conditions into the genetic algorithm, and the genetic algorithm searches for a control input combination that makes the objective function reach the optimal value through iterative calculation; the control input includes the rotation speed of the rotor, the power of the heating device, and the air volume of the fan; S3.5: Verify the calculated optimal control input to check whether it meets the physical feasibility and operation requirements of the system; If the optimal control input does not satisfy the constraints or causes unstable system operation, the parameters of the genetic algorithm are readjusted and the calculation is performed again.

9. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 8, characterized in that: The model predictive control algorithm predicts the operating results of the air-conditioning rotary dehumidification system under different control strategies based on the current system operating status and load prediction, and obtains the optimal control parameters, including the rotary wheel speed, heating power, and fan frequency, through optimization calculation.

10. The energy-saving optimization control method for the air-conditioning rotary dehumidification system based on the mechanism model according to claim 9, characterized in that: The mechanism model includes the adsorption and desorption process of the wheel and the heat exchange process of the air flow; the historical operation data of the air-conditioning rotary dehumidification system includes indoor and outdoor air humidity, temperature, air volume and energy consumption; the optimal control input includes the optimal speed of the wheel, the optimal power of the heating device and the optimal frequency of the fan.

Images