A metro station ventilation control method based on adaptive rolling optimization

Abstract

The application provides a subway station ventilation control method based on adaptive rolling optimization, comprising the following steps: S1, collecting operation data of the subway station by a ventilation control system; S2, constructing a first-order thermal balance simplified physical model of a platform of the subway station; performing online parameter identification to obtain physical model parameters; S3, comparing a model predicted temperature value with a platform temperature to obtain an original residual error and a bias correction amount; S4, constructing an enhanced prediction model; S5, performing rolling optimization to obtain a wind volume control sequence; and S6, finally extracting a control amount in the wind volume control sequence and issuing the control amount to a fan frequency converter for ventilation control, and re-executing S1 to S6 at the next sampling time. Under the premise of strictly guaranteeing passenger thermal comfort (PMV), optimal energy-saving wind volume strategies are calculated and executed in real time, model mismatch and early morning peak cold stimulation effects can be effectively solved, and efficient, comfortable and stable operation of the station environment is realized.

Description

Technical Field

[0001] This invention relates to the field of subway ventilation control, and more particularly to a subway station ventilation control method based on adaptive rolling optimization. Background Technology

[0002] The ventilation and air conditioning system (the environmental control subsystem in the BAS) is a critical infrastructure for ensuring air quality, thermal comfort, and normal equipment operation in subway stations. However, it is also a major energy consumer in subway operations, typically accounting for 40% of the station's total electricity consumption. How to minimize the energy consumption of the ventilation and air conditioning system while ensuring passenger comfort has become a critical technical problem that the rail transit industry urgently needs to solve.

[0003] As the backbone of modern urban public transportation, subway stations are typical large-scale underground public building spaces. Compared with above-ground buildings, subway stations have the following significant characteristics: (1) High-density crowds, especially during morning and evening peak hours, result in a huge instantaneous passenger flow, which leads to a sharp increase and drastic fluctuation in the heat dissipation and moisture dissipation load of personnel.

[0004] (2) The station is highly enclosed. It is usually buried deep underground and has poor natural ventilation. It mainly relies on mechanical ventilation and air conditioning systems to maintain the station environment.

[0005] (3) The large thermal inertia of the concrete enclosure structure and soil has a huge heat storage capacity, which makes the response of the ambient temperature to external changes significantly lagging.

[0006] (4) Multiple disturbance sources are coupled and affected by multiple factors such as outdoor meteorological conditions, train operation (piston wind, brake heat), platform screen door opening and closing, equipment heating and random passenger flow. The environmental parameters are highly nonlinear, time-varying and uncertain.

[0007] Traditional control methods (such as PID control) or predictive control based on fixed models fail to consider the slow time-varying characteristics of station environmental parameters (such as the heat transfer coefficient of the building envelope and the efficiency decay caused by equipment aging) with the seasons and time, resulting in a serious mismatch between the control model and the actual object and a decrease in control accuracy.

[0008] Existing methods are ill-equipped to effectively handle unpredictable random disturbances (such as heat exchange caused by frequent opening and closing of platform screen doors, sudden surges in passenger flow, etc.), and lack effective online compensation mechanisms, resulting in large temperature fluctuations within the station and impacting passenger comfort. Furthermore, they lack the ability to predict future load changes, often relying on "post-event adjustments," which exhibit significant control lag. Summary of the Invention

[0009] The purpose of this invention is to provide a subway station ventilation control method based on adaptive rolling optimization, which can construct a dynamic model that can accurately track environmental changes by using adaptive model predictive control (MPC) technology and online parameter identification and feedback deviation correction mechanism of RLS.

[0010] This invention is achieved through the following technical solution: A subway station ventilation control method based on adaptive rolling optimization includes: S1. The ventilation control system collects operational data from the subway station, including platform temperature, carbon dioxide concentration, outdoor temperature, passenger flow, supply air temperature, and fan operating frequency, and obtains historical operational data of the subway station. S2. Construct a simplified first-order thermal equilibrium physical model of the subway station platform based on the collected operational data; then, based on the current operational data and historical operational data, use a recursive least squares algorithm with a forgetting factor to identify the parameters of the simplified first-order thermal equilibrium physical model online, and obtain the physical model parameters. S3. Calculate the model predicted temperature value of the platform based on the physical model parameters, compare the model predicted temperature value with the platform temperature to obtain the original residual, and perform low-pass filtering on the original residual to obtain the deviation correction amount. S4. Integrate the physical model parameters with the deviation correction amount to construct an enhanced prediction model; S5. Based on the enhanced prediction model, predict the platform temperature within the set time domain, construct a carbon dioxide concentration prediction model and predict the carbon dioxide concentration within the set time domain; then construct a rolling optimization objective function with ventilation energy consumption, passenger thermal comfort and air volume change rate as optimization objectives, and perform rolling optimization in combination with carbon dioxide concentration prediction results to obtain the air volume control sequence. S6. Finally, extract the current control quantity from the air volume control sequence and send it to the fan frequency converter for ventilation control. S1 to S6 are executed again at the next sampling time.

[0011] Compared with previous technologies, the beneficial effects of the present invention are as follows: 1. Under the premise of strictly ensuring passenger thermal comfort (PMV), this invention calculates and executes the optimal energy-saving air volume strategy in real time, which can effectively solve the model mismatch and the morning peak cooling effect, and achieve efficient, comfortable and stable operation of the station environment.

[0012] 2. By introducing recursive least squares (RLS) with a forgetting factor, the time-varying physical parameters of the station's thermodynamic model are identified in real time online. This enables the control model to automatically track characteristic changes caused by seasonal changes and equipment aging, fundamentally eliminating the impact of model mismatch on control accuracy. 3. A rolling time-domain optimization mechanism is used to achieve forward-looking adjustment of future loads and overcome the large lag characteristic. At the same time, combined with the feedback deviation correction link based on residual analysis, real-time dynamic compensation is performed for unmodeled dynamics and sudden random disturbances to ensure that the platform temperature remains stable even under drastic fluctuations in passenger flow.

[0013] 4. Abandoning the control mode of a single temperature setpoint, the system directly uses passenger thermal comfort (PMV), fan operating energy consumption, and CO2 concentration as comprehensive optimization objectives. Under the premise of meeting comfort constraints, the system solves the optimal airflow control sequence in real time within a finite time domain to find the "most economical operating point" of the system, significantly reducing operating energy consumption. Attached Figure Description

[0014] Figure 1 This is a flowchart of the main process of the present invention; Figure 2 This is a block diagram of the ventilation control system. Figure 3 A comparison chart of PMV indicators; Figure 4 This is a graph showing the air volume adjustment curve. Figure 5 Identify the parameter tracking graph for RLS; Figure 6 The adaptive control curves for temperature and air volume are shown. Figure 7 An adaptive control curve for passenger flow and carbon dioxide concentration under simulated high passenger flow conditions. Detailed Implementation

[0015] The present invention will now be described in detail with reference to the accompanying drawings, but the scope of protection of the present invention is not limited to the following description: like Figures 1 to 7 As shown, a subway station ventilation control method based on adaptive rolling optimization includes: S1. The ventilation control system collects operational data from the subway station, including platform temperature, carbon dioxide concentration, outdoor temperature, passenger flow, supply air temperature, and fan operating frequency, and obtains historical operational data and control parameters of the subway station. Furthermore, the operational data also needs to undergo amplitude limiting filtering, normalization, and abnormal bad pixel removal processing; The platform temperature and carbon dioxide concentration are collected by temperature and humidity sensors and carbon dioxide concentration sensors installed in the public area of ​​the platform, and the outdoor temperature is collected by outdoor meteorological sensors installed at the station's fresh air inlet. The passenger flow, air supply temperature and fan operating frequency are all obtained through the station's integrated monitoring system, and the station's integrated monitoring system communicates with the data acquisition system via OPC or Modbus protocols. It should be noted that, is the total air volume delivered by the air conditioning fan at time k, in cubic meters per hour, and is the control input variable of the ventilation control system, representing the total mixed air volume (the air volume after the fresh air and return air are mixed) delivered by the air conditioning unit to the public area of ​​the platform. Convert the fan frequency to air volume The formula is expressed as: ; in For the rated air supply volume, The rated frequency for the operation of the fan. This refers to the operating frequency of the fan; and the rated operating frequency of a conventional fan is 50Hz. S2. Construct a simplified first-order thermal equilibrium physical model of the subway station platform based on the collected operational data; then, based on the current operational data and historical operational data, use a recursive least squares algorithm with a forgetting factor to identify the parameters of the simplified first-order thermal equilibrium physical model online, and obtain the physical model parameters. Furthermore, the formula for the simplified physical model of first-order thermal equilibrium described in S2 is expressed as follows: ; in, Let K be the platform temperature at time k. Let K be the outdoor temperature at time k. Let K be the supply air temperature at time k. Let K be the air volume at time k. Let be the passenger flow at time k, and let a, b, and d be the physical parameters to be identified. ξ(k) The model perturbation term; the model perturbation term ξ(k) The initial value is 0, and during system operation, it will adapt to the working conditions of the simulation model by changing the model parameters step by step. Furthermore, in the physical model parameters, Used to characterize the effect of heat transfer from the building envelope on the platform temperature. Used to characterize the impact of ventilation, air conditioning and cooling on platform temperature. Used to characterize the effect of personnel heat dissipation on platform temperature; The recursive least squares algorithm is used to construct the regression vector. Its formula is expressed as: ; The gain matrix is ​​then updated according to the following recursive formula. Physical model parameters and covariance matrix : ; ; ; ; in, The increment is the temperature change, and λ is the forgetting factor, ranging from 0.95 to 0.99. S3. Calculate the model predicted temperature value of the platform based on the physical model parameters, compare the model predicted temperature value with the platform temperature to obtain the original residual, and perform low-pass filtering on the original residual to obtain the deviation correction amount. Furthermore, in S3, the model-predicted temperature values ​​are compared with the platform temperature to obtain the original residuals. Its formula is expressed as: ; Then, a first-order low-pass filter is used to filter the original residual. Low-pass filtering is performed to obtain the deviation correction at time k. Its formula is expressed as: ; in, Let K be the platform temperature at time k-1. The model predicts the temperature value at time k-1. These are the physical model parameters at time k-1. Let α be the deviation correction amount at time k-1, and α be the filter coefficient and α∈(0,1); S4. Integrate the physical model parameters with the deviation correction amount to construct an enhanced prediction model; Furthermore, the formula for the enhanced prediction model described in S4 is expressed as follows: ; Among them, the This is the transpose of the regression vector constructed in step S2. Let be the deviation correction amount at time k. These are the parameters of the physical model; S5. Based on the enhanced prediction model, predict the platform temperature within the set time domain, construct a carbon dioxide concentration prediction model and predict the carbon dioxide concentration within the set time domain; then construct a rolling optimization objective function with ventilation energy consumption, passenger thermal comfort and air volume change rate as optimization objectives, and perform rolling optimization in combination with carbon dioxide concentration prediction results to obtain the air volume control sequence. Furthermore, in S5, a carbon dioxide concentration prediction model is constructed based on the mass conservation relationship. The formula for the carbon dioxide concentration prediction model is expressed as follows: ; in, Let be the carbon dioxide concentration at time k. Let be the carbon dioxide concentration at time k+1, Δt be the control step size, V be the effective volume of the platform, and η be the fresh air ratio coefficient. Outdoor carbon dioxide concentration, Per capita carbon dioxide production rate; The set time domain is In each control step, the formula for the rolling optimization objective function is expressed as: ; in, Let i be the energy consumption of the ventilation control system in the i-th control step. For the passenger thermal comfort at the i-th control step, Let be the rate of change of air volume at adjacent control times before and after the i-th control step. , and These are the corresponding penalty factors; Normally set The system consists of 6 control steps, each lasting 5 minutes, corresponding to a 30-minute prediction time domain. After rolling optimization, the first airflow control quantity Q(k) is extracted from the obtained airflow control sequence and converted into a frequency control command for the fan inverter. The energy consumption of the ventilation control system is determined according to the fan similarity law, which satisfies the condition P∝Q. 3 The passenger thermal comfort value (PMV) is calculated using the Fanger thermal comfort model, with the human metabolic rate set at 1.2 Mt, clothing thermal resistance set at 0.5 clo, and relative humidity set at 60%. The square of the PMV is used as the thermal comfort penalty. Furthermore, the Passenger Thermal Comfort Value (PMV) index is calculated based on the Fanger thermal comfort equation in the international standard ISO7730 or the Chinese national standard GB / T18049. It is a comprehensive function of air temperature, mean radiant temperature, wind speed, relative humidity, human metabolic rate, and clothing thermal resistance. In the simplified engineering calculation of this invention, relative humidity (e.g., 60%), wind speed (e.g., 0.3 m / s), and clothing thermal resistance (0.5 clo in summer) are set as constants. The platform temperature is mainly optimized as the core variable affecting PMV, thereby ensuring both calculation efficiency and accuracy of thermal comfort evaluation. S6. Finally, extract the current control quantity from the air volume control sequence and send it to the fan frequency converter for ventilation control. S1 to S6 are re-executed at the next sampling time. The air volume control sequence The set of optimal air supply volumes within a given time domain obtained by solving the rolling optimization objective function can be expressed as follows: ; The optimized air supply volume command for the future k+i time step is obtained. Based on the rolling optimization principle, only the first term of the sequence is applied to the control quantity of the ventilation and air conditioning system, and the remaining elements of the sequence are not directly executed. When the sampling time is updated, a new control sequence is resolved by combining the actual sensor measurement status and the updated identification model. Each air supply volume term in the sequence is constrained by the fan amplitude and the rate of change of air volume, while also satisfying the hard constraints of station temperature and carbon dioxide concentration. The fan energy consumption, thermal comfort PMV index, and actuator smoothness are comprehensively balanced.

[0016] In summary, to verify the effectiveness of the control strategy proposed in this invention, a continuous simulation scenario was constructed for the typical summer weekday operating hours (6:00-23:00) in Fuzhou, covering the pre-cooling period, morning and evening peak high-load periods, and nighttime low-load periods. The simulation input parameters were set as follows: ① Meteorological parameters: Data from typical summer days in Fuzhou were used, showing daily fluctuations in outdoor temperature with a peak of approximately 36℃; ② Passenger flow data: Based on the actual statistical patterns of subway stations, it shows typical load characteristics of morning and evening peaks, off-peak and low peaks during the day, and participates in prediction and optimization calculations as the main disturbance of the system; ③ Supply air temperature: set according to the operating period: 16℃ during peak period, 20℃ during off-peak period, and 26℃ during low load period; The table below provides detailed simulation settings:

[0017] in, Figure 3 The diagram depicts the intraday variation curves of PMV (Personal Ventilation Volume) in the public area of ​​a subway station under four control strategies. The light green area represents the thermal comfort range specified by the ISO 7730 standard (PMV ∈ [-0.5, 0.5]). The subway station ventilation control method based on adaptive rolling optimization proposed in this invention is represented by the "SLSQP adaptive" curve in the diagram. It exhibits the best control effect, with the PMV consistently maintained within the comfort range throughout the day. The curve shows gentle fluctuations and no significant overshoot or oscillations even during periods of drastic passenger flow fluctuations in the morning and evening rush hours, demonstrating excellent load tracking capability and steady-state control accuracy. Traditional timed segmented control strategies are open-loop controls, unable to adapt to dynamic load changes, resulting in significant overcooling during peak load switching phases, with PMV repeatedly exceeding the lower limit of the comfort range. While PID control possesses closed-loop regulation capabilities, it suffers from regulation lag, poor adaptability to operating conditions, and significant overshoot. The PSO adaptive strategy produces a PMV curve closer to the neutral center of the comfort range, exhibiting excellent steady-state optimization accuracy; however, its single-step optimization calculation time is significantly higher than the "SLSQP adaptive" curve of this invention, making it difficult to match the rolling control cycle of subway ventilation and air conditioning systems. Considering both regulation accuracy and real-time performance, the method proposed in this invention is better suited to the requirements of online optimization operation.

[0018] Figure 4 The figure describes the variation curves of the total daily air supply volume in the public area of ​​a subway station under four strategies. The subway station ventilation control method based on adaptive rolling optimization proposed in this invention is represented by the "SLSQP adaptive" curve in the figure. This curve shows that the air supply volume is continuously and dynamically adjusted according to the load changes within the station. During off-peak and low-load periods, the air supply output is rapidly reduced to accurately match real-time demand and effectively avoid excessive air supply. Traditional strategies present a stepped fixed air volume, maintaining a relatively high basic air supply volume during off-peak periods, resulting in significant energy redundancy. PID control is relatively slow in adjustment, but its overall air supply level is higher than that of the adaptive strategy. The air volume variation trend of PSO adaptive is similar to that of SLSQP. Based on the total daily energy consumption statistics, compared with traditional timed segmented control, the SLSQP adaptive control strategy of this invention can achieve an energy saving effect of 7.9%, verifying the significant role of the dual adaptive rolling optimization framework in improving energy efficiency.

[0019] Furthermore, to verify the robustness of this invention under time-varying or sudden disturbances in system characteristics, a step disturbance experiment was designed: at 12:00 during the simulation, the overall heat transfer coefficient α of the system was artificially increased from the initial value of 0.10 to 0.20 to simulate typical faults such as a sudden decrease in the thermal insulation performance of the building envelope or sensor drift, providing a model for... Figure 5 The RLS identification parameter tracking diagram shown and as follows Figure 6 The diagram shows the adaptive control curves for temperature and air volume; where... Figure 5 This indicates that the identification parameters of the RLS algorithm quickly detected dynamic changes after parameter mutations. After a brief adjustment, the identification parameters reconverged to the new true values ​​within about 13 minutes. The convergence was smooth and the overshoot was small, proving that the RLS identification algorithm has good real-time tracking ability and strong robustness. Figure 6 This indicates that the platform temperature deviated briefly, but the controller responded quickly based on the updated parameter estimates and dynamically increased the air supply to compensate. The temperature stabilized to the set value of 26°C within about 15 minutes. No significant overshoot or oscillation occurred during the recovery process, which verifies that the subway station ventilation control method based on adaptive rolling optimization proposed in this invention can still maintain the closed-loop stability of the system and the ability to track the set value even when the model is mismatched.

[0020] Figure 7This paper describes the coordination capability of the present invention under multi-objective constraint conflicts, and tests were conducted on a sudden surge in passenger flow during the morning rush hour: The scenario was set between 12:00 and 13:00, with the number of passengers entering the station suddenly increasing to 15,000. Following the surge in passenger flow, the CO2 concentration inside the station rapidly increased. When the predictive model determined that it was about to exceed the national limit of 1000 ppm, the model's predictive control framework automatically triggered a constraint priority mechanism, prioritizing air quality over thermal comfort and forcing the fans to operate at full capacity. The CO2 concentration plateaued during the peak passenger flow period and then quickly fell back below the limit, demonstrating that the multi-objective optimization framework of the present invention can effectively coordinate the competitive relationship between thermal comfort and air quality, ensuring environmental safety and reliable system operation even under extreme loads.

[0021] Furthermore, the ventilation control system includes: The data acquisition module is used to collect operational data of the subway station, including platform temperature, carbon dioxide concentration, outdoor temperature, passenger flow, air supply temperature, and fan operating frequency, and to obtain historical operational data and control parameters of the subway station. An adaptive prediction module, connected to the data acquisition module, is used to construct a first-order simplified physical model of thermal equilibrium based on the operating data, and to perform online parameter identification on the first-order simplified physical model of thermal equilibrium to obtain physical model parameters; and to generate a deviation correction amount based on the original residual between the model predicted temperature value and the station temperature, in order to construct an enhanced prediction model. A rolling optimization control module, connected to the adaptive prediction module, is used to perform rolling optimization based on the enhanced prediction model and the carbon dioxide concentration prediction model, with ventilation energy consumption, passenger thermal comfort, and air volume change rate as optimization objectives, and to solve for the air volume control sequence; and An execution module, connected to the rolling optimization control module, is used to extract the current control quantity in the air volume control sequence and convert it into a frequency control command for the fan inverter to drive the return air fan and / or exhaust fan of the subway station to operate. In addition, the adaptive prediction module includes an RLS parameter estimation unit and a model reconstruction and bias correction unit; the RLS parameter estimation unit is used to calculate the time-varying parameters of the first-order thermal equilibrium simplified physical model in real time using a recursive least squares algorithm with a forgetting factor; the model reconstruction and bias correction unit is used to generate a bias correction term based on the residual between the actual measured temperature and the theoretical value of the model, and output the enhanced prediction model by combining the physical model parameters with the bias correction term. The rolling optimization control module adopts a model predictive control algorithm and is equipped with a nonlinear optimization solver SLSQP, which is used to calculate the optimal air volume control sequence in the future finite time domain based on passenger comfort, energy consumption and equipment movement amplitude in each control cycle. The execution module is used to convert the current optimal control quantity output by the rolling optimization control module into the frequency command of the frequency converter VFD, and drive the return air fan and / or exhaust fan of the subway station through the frequency converter.

[0022] Although the present invention has been illustrated and described through specific embodiments and alternative methods, it should be understood that various changes and modifications may be made without departing from the spirit and scope of the invention. Therefore, it should be understood that the present invention is not limited in any sense except by the appended claims and their equivalents.

Claims

1. A subway station ventilation control method based on adaptive rolling optimization, characterized in that, include: S1. The ventilation control system collects operational data from the subway station, including platform temperature, carbon dioxide concentration, outdoor temperature, passenger flow, supply air temperature, and fan operating frequency, and obtains historical operational data of the subway station. S2. Construct a simplified first-order thermal equilibrium physical model of the subway station platform based on the collected operational data; then, based on the current operational data and historical operational data, use a recursive least squares algorithm with a forgetting factor to identify the parameters of the simplified first-order thermal equilibrium physical model online, and obtain the physical model parameters. S3. Calculate the model predicted temperature value of the platform based on the physical model parameters, compare the model predicted temperature value with the platform temperature to obtain the original residual, and perform low-pass filtering on the original residual to obtain the deviation correction amount. S4. Integrate the physical model parameters with the deviation correction amount to construct an enhanced prediction model; S5. Based on the enhanced prediction model, predict the platform temperature within the set time domain, construct a carbon dioxide concentration prediction model and predict the carbon dioxide concentration within the set time domain; then construct a rolling optimization objective function with ventilation energy consumption, passenger thermal comfort and air volume change rate as optimization objectives, and perform rolling optimization in combination with carbon dioxide concentration prediction results to obtain the air volume control sequence. S6. Finally, extract the current control quantity from the air volume control sequence and send it to the fan frequency converter for ventilation control. S1 to S6 are executed again at the next sampling time.

2. The subway station ventilation control method based on adaptive rolling optimization according to claim 1, characterized in that, The formula for the simplified physical model of first-order thermal equilibrium described in S2 is expressed as follows: ； in, Let K be the platform temperature at time k. Let K be the outdoor temperature at time k. Let K be the supply air temperature at time k. Let K be the air volume at time k. Let ξ(k) be the passenger flow at time k, a, b, and d be the physical parameters to be identified, and ξ(k) be the model disturbance term.

3. The subway station ventilation control method based on adaptive rolling optimization according to claim 2, characterized in that, The recursive least squares algorithm described in S2 is used to construct the regression vector. Its formula is expressed as: ； The gain matrix is ​​then updated according to the following recursive formula. Physical model parameters and covariance matrix : ； ； ； ； in, Let λ be the temperature change increment, and λ be the forgetting factor, with a value ranging from 0.95 to 0.

99.

4. The subway station ventilation control method based on adaptive rolling optimization according to claim 3, characterized in that, In S3, the model-predicted temperature values ​​are compared with the platform temperature to obtain the original residuals. Its formula is expressed as: ； Then, a first-order low-pass filter is used to filter the original residual. Low-pass filtering is performed to obtain the deviation correction at time k. Its formula is expressed as: ； in, Let K be the platform temperature at time k-1. The model predicts the temperature value at time k-1. These are the physical model parameters at time k-1. Let α be the deviation correction amount at time k-1, and α be the filter coefficient and α∈(0,1).

5. A subway station ventilation control method based on adaptive rolling optimization according to claim 4, characterized in that, The formula for the enhanced prediction model described in S4 is expressed as follows: ； Among them, the This is the transpose of the regression vector constructed in step S2. Let be the deviation correction amount at time k. These are the parameters of the physical model.

6. A subway station ventilation control method based on adaptive rolling optimization according to claim 5, characterized in that, In S5, a carbon dioxide concentration prediction model is constructed based on the mass conservation relationship. The formula for the carbon dioxide concentration prediction model is expressed as follows: ； in, Let be the carbon dioxide concentration at time k. Let be the carbon dioxide concentration at time k+1, Δt be the control step size, V be the effective volume of the platform, and η be the fresh air ratio coefficient. Outdoor carbon dioxide concentration, This refers to the per capita carbon dioxide production rate.

7. A subway station ventilation control method based on adaptive rolling optimization according to claim 1, characterized in that, The time domain set in S5 is In each control step, the formula for the rolling optimization objective function is expressed as: ； in, Let i be the energy consumption of the ventilation control system in the i-th control step. For the passenger thermal comfort at the i-th control step, Let be the rate of change of air volume at adjacent control times before and after the i-th control step. , and These are the corresponding penalty factors.

Images