Scheduling optimization method based on multi-target rail transit station air conditioning system

Abstract

The invention provides a scheduling optimization method based on a multi-target rail transit station air conditioning system, and the method comprises the following steps: P1, collecting the building parameter information of a rail transit station, and constructing a rail transit station 3D model recognized by EnergyPlus through SketchUp software; p2, modifying key parameters of the air conditioning system; p3, building an intelligent optimization model of the air conditioning system, calculating an error evaluation index by using verification set data for evaluation, performing real-time simulation by using the intelligent optimization model after the model can be used, and adjusting the air conditioning system until a scheduling scheme of the air conditioning system is determined; and P4, acquiring real-time monitoring data through indoor and outdoor temperature sensor monitoring equipment, and inputting the real-time monitoring data into the intelligent optimization model. Intelligent dispatching and efficient operation of the air conditioning system are achieved, the relation between the comfort degree and energy consumption is balanced, parameters can be set and adjusted autonomously according to actual requirements, and high flexibility and customizability are achieved.

Description

Technical Field

[0001] The invention relates to the technical field of intelligent control of rail transit buildings, and in particular to a scheduling optimization method based on a multi-objective rail transit station air conditioning system. Background Art

[0002] With the development of society and people's growing demand for a comfortable environment, air-conditioning systems are increasingly used in buildings and have become an important means to improve the comfort of living, working and commercial spaces. However, existing air-conditioning scheduling methods usually only consider the optimization of building energy consumption and only use energy saving as the only criterion for building energy management and intelligent control, while ignoring multiple factors such as human comfort and environmental protection. This single-dimensional optimization method often brings huge energy costs while improving comfort. Especially at a time when global energy issues are becoming increasingly prominent and environmental protection is receiving increasing attention, there is an urgent need for a more comprehensive and intelligent air-conditioning system scheduling solution to balance the relationship between comfort, energy efficiency and environmental protection.

[0003] In the determination of the previous air-conditioning system scheduling scheme, most of the related research is devoted to building energy conservation. For example, patent number CN108804802A discloses an energy-saving design method for office building space with the goal of controlling summer air-conditioning refrigeration energy consumption. The basic body model of the office building is determined according to the physical environment and the body design factors affecting the energy consumption of the building are determined; the basic body model established in the DesignBuilder software is used to simulate and analyze the summer air-conditioning refrigeration energy consumption with different plane forms, plane aspect ratios, and floor heights as single variables; the office building body model with relatively low summer air-conditioning refrigeration energy consumption is obtained, and on this basis, the summer air-conditioning refrigeration energy consumption simulation analysis is performed again with different building orientations and space combination forms as single variables; the passive energy-saving design potential of the building body is further explored to achieve the energy-saving goal; some results are also committed to balancing comfort and energy saving; for example, CN116358114A discloses an air-conditioning temperature control method based on deep reinforcement learning, Q-value neural network training is performed based on the TD3 algorithm, and the trained Q-value neural network is deployed in the air-conditioning control system to adjust the air-conditioning temperature in real time.

[0004] The above-mentioned design method for air-conditioning system scheduling still has the following shortcomings: only considering the single factor of building energy conservation and scheduling the air-conditioning system only based on monitoring data often leads to data missing and lack of data sets; therefore, an air-conditioning system scheduling scheme that can comprehensively consider multiple factors is needed to minimize energy consumption while ensuring human comfort. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a scheduling optimization method for the air-conditioning system of a rail transit station based on a multi-objective method. By adopting the number of human discomfort hours and building energy consumption as optimization targets, the optimal scheduling scheme for the air-conditioning system in the transportation building is determined, the intelligent scheduling and efficient operation of the air-conditioning system are realized, and the relationship between comfort and energy consumption is balanced. Parameters can be independently set and adjusted according to actual needs, and the method has strong flexibility and customizability.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] A scheduling optimization method for a rail transit station air conditioning system based on a multi-objective method comprises the following steps:

[0008] P1. Collect architectural parameter information of rail transit stations, build a 3D model of the rail transit station recognized by EnergyPlus through SketchUp software, import the 3D model of the rail transit station and meteorological data files into EnergyPlus software to build the air conditioning system, and set the basic structural parameters of the rail transit station;

[0009] P2. Modify the key parameters of the air conditioning system, obtain the data generated by the simulation operation of the EnergyPlus software, aggregate the data generated by the simulation operation with the real-time monitoring data and perform pre-processing operations to form a data set for intelligent optimization model training;

[0010] P3. Build an intelligent optimization model for the air-conditioning system. Divide the data set preprocessed in step P2 into a test set and a validation set in proportion. Use the test set data for iterative training. Use the validation set data to calculate the error evaluation index for evaluation. After the model is usable, use the intelligent optimization model for real-time simulation and adjust the air-conditioning system until the air-conditioning system scheduling plan is determined.

[0011] P4. Obtain real-time monitoring data through indoor and outdoor temperature sensor monitoring equipment and input it into the intelligent optimization model. After autonomously setting the target temperature range and air conditioning setting temperature range, output the air conditioning system scheduling plan.

[0012] In step P1, the rail transit station is a rail transit building, and the building parameter information includes geometric information of the building, passenger flow, and station area;

[0013] The EnergyPlus software imports the 3D model of the transportation station and the meteorological data file, and sets the basic structural parameters of the rail transit station;

[0014] The basic structural parameters include the building's envelope, room orientation, geographic information, occupant density, and power-consuming equipment information.

[0015] In step P1, the meteorological data are meteorological parameters and weather phenomena for each hour in a year.

[0016] In step P2, the key parameters include the air conditioning cooling setting temperature and the air conditioning heating setting temperature, the preprocessing includes data cleaning, data standardization, and data conversion into a format for model training; the real-time monitoring data includes indoor and outdoor temperature, humidity, passenger flow and CO2 concentration.

[0017] In step P3, the steps of real-time simulation of the intelligent optimization model are as follows:

[0018] W1. Write an intelligent optimization algorithm in Python and set the initial relevant environmental parameters and air conditioning system related parameters;

[0019] The relevant parameters include indoor and outdoor temperature, humidity, passenger flow and CO2 concentration;

[0020] The air conditioning system related parameters include the target temperature range for human comfort, the air conditioning cooling setting temperature range and the air conditioning heating setting temperature range;

[0021] W2. Based on the preprocessed data set obtained in step P2, a temperature and energy consumption prediction model is constructed using a deep learning algorithm, and the model is iteratively trained. After the training is completed, the model is verified and evaluated;

[0022] Use the deep learning algorithm and the preprocessed data set to train the indoor temperature and energy consumption prediction model, iteratively update the parameters of the temperature prediction model and the energy consumption prediction model until the preset number of iterations is reached. After the training is completed, the effect of the prediction model is evaluated by the root mean square error and the mean absolute error. After the model is usable, enter step W3;

[0023] W3, substitute the target temperature and the minimum value of the air conditioning setting temperature range, use the temperature and energy consumption prediction model to predict the indoor temperature and energy consumption, and adjust the air conditioning system according to the indoor temperature until the set target temperature is reached, and output the total energy consumption under the current setting;

[0024] W4. Based on step W2, traverse the air conditioning setting temperature range and the target temperature range in turn, adjust the target temperature value and the air conditioning temperature setting value and compare them with the previous ones until the air conditioning setting scheduling plan is found.

[0025] In step W3, the indoor temperature is adjusted by the air conditioning system as follows:

[0026] W3.1, the intelligent optimization model first determines whether the outdoor temperature is greater than the current target temperature, if not, it goes to step W3.2, if yes, it goes to step W3.3;

[0027] W3.2, determine whether the indoor temperature is close to the target temperature, if not, proceed to step W3.4, if yes, proceed to step W3.5;

[0028] W3.3, determine whether the indoor temperature is close to the target temperature, if not, go to step W3.6, if yes, go to step W3.5;

[0029] W3.4, determine whether the indoor temperature is greater than the target temperature, if not, the air conditioning heating temperature is +1, if yes, the air conditioning heating temperature is -1;

[0030] W3.5, keep the current settings;

[0031] W3.6, determine whether the indoor temperature is lower than the target temperature, if not, the air conditioning temperature is -1, if yes, the air conditioning temperature is +1;

[0032] W3.7. After adjusting the air conditioner setting temperature based on steps W3.1-W3.6, call the temperature and energy consumption prediction model again for prediction.

[0033] In step W4, the optimal air conditioning setting scheduling scheme is found as follows:

[0034] Determine whether the air-conditioning temperature setting range has been traversed. If not, the air-conditioning setting temperature is increased by 1, and the current target temperature and the air-conditioning setting temperature are substituted and step W3 is performed again. If so, continue to determine whether the target temperature range has been traversed. If not, the target temperature is increased by 1, and the current target temperature and the air-conditioning setting temperature are substituted and step W3 is performed again. If so, output the air-conditioning setting temperature and target temperature corresponding to the minimum energy consumption to obtain the air-conditioning system scheduling plan.

[0035] The beneficial effects of the present invention are:

[0036] 1. Use SketchUp software to build a 3D model of a rail transit station that can be recognized by EnergyPlus, and build an air-conditioning system to set key environmental parameters. By modifying the key parameters of the air-conditioning system, a large amount of data generated by multiple simulation runs of the EnergyPlus software is obtained and aggregated with real-time monitoring data to form a data set. Then, the optimal air-conditioning system scheduling plan is determined based on the preprocessed data set and the intelligent optimization model.

[0037] 2. Based on simulation and deep learning algorithms, it solves the dual-objective optimization problem of human comfort and indoor energy consumption. Through real-time scheduling and feedback mechanisms, it adapts to dynamic changes in indoor and outdoor environments, and can adjust parameters according to actual needs. It has strong flexibility and customizability, fully considers human comfort and energy consumption, and realizes intelligent scheduling and efficient operation of the air-conditioning system. While improving user experience, it brings significant advantages in building energy conservation and environmental protection.

[0038] 3. Through the adjustment of key parameters, it plays a vital role in indoor energy consumption and human comfort. Within the self-set parameter range, simulation is carried out one by one according to the value step size, so as to obtain a large amount of data to fill the gaps in the actual collected data. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flow chart of the method of the present invention;

[0040] Figure 2 Schematic diagram of the 3D model of a rail transit station built using SketchUp software;

[0041] Figure 3 Flowchart of the intelligent optimization algorithm used in the intelligent optimization model. DETAILED DESCRIPTION

[0042] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0043] Example 1

[0044] like Figures 1 to 2 As shown, a scheduling optimization method for a multi-objective rail transit station air conditioning system comprises the following steps:

[0045] P1. Collect architectural parameter information of rail transit stations, build a 3D model of rail transit stations that can be recognized by EnergyPlus through SketchUp software, import the 3D model of rail transit stations and local meteorological data files into EnergyPlus software to build air conditioning systems, and set basic structural parameters of rail transit stations;

[0046] P2. Modify the key parameters of the air conditioning system, obtain a large amount of data generated by the simulation operation of the EnergyPlus software (the EP software simulates the environment of the target station every 10 minutes for a total of one year, and obtains a total of 8760*6 data). The data generated by the simulation operation is aggregated with the real-time monitoring data and pre-processed to form a data set for intelligent optimization model training;

[0047] P3. Build an intelligent optimization model for the air-conditioning system. Divide the data set preprocessed in step P2 into a test set and a validation set in proportion. Use the test set data for iterative training. Use the validation set data to calculate the error evaluation index for evaluation. After the model is usable, use the intelligent optimization model for real-time simulation and adjust the air-conditioning system until the optimal air-conditioning system scheduling plan is determined (i.e., a plan that satisfies both human comfort and low indoor energy consumption).

[0048] The data set division ratio is set according to actual needs; the two performance indicators for evaluating the scheduling plan of the indoor air-conditioning system of rail transit stations include: human comfort and indoor energy consumption (human discomfort hours and building energy consumption). Human comfort is a nonlinear equation established based on factors such as the average indoor temperature and indoor relative humidity of transportation buildings, and indoor energy consumption is the energy consumption generated by the overall operation of the air-conditioning system per hour;

[0049] P4. Obtain real-time monitoring data through indoor and outdoor temperature sensor monitoring equipment and input it into the intelligent optimization model. After autonomously setting the target temperature range and air conditioning setting temperature range, output an optimal air conditioning system scheduling plan that meets human comfort and has low indoor energy consumption.

[0050] In step P1, the rail transit station is a rail transit building, and the building parameter information includes geometric information of the building, passenger flow, and station area; and the 3D model of the rail transit station is saved in the .idf format recognizable by EnergyPlus using the Legacy OpenStudio plug-in;

[0051] The EnergyPlus software imports the 3D model of the transportation station and the meteorological data file of the area where it is located, and sets the basic structural parameters of the rail transit station;

[0052] The basic structural parameters include the building's envelope, room orientation, geographic information, occupant density, and power-consuming equipment information.

[0053] The meteorological data refers to the meteorological data provided by the EnergyPlus official website for simulating actual building operating conditions.

[0054] Build the air-conditioning system and set key environmental parameters according to the actual operation of the rail transit station, including: air-conditioning system parameters, lighting equipment parameters, personnel activity time planning data, air-conditioning system heating temperature and air-conditioning system cooling temperature.

[0055] The air conditioning system parameters include: cooling and heating temperature setting values, allocation of temperature control areas, operation mode and scheduling strategy of the air conditioning system.

[0056] In step P1, the meteorological data are meteorological parameters and weather phenomena for each hour in a year.

[0057] In step P2, the key parameters include the air conditioning cooling setting temperature and the air conditioning heating setting temperature, the preprocessing includes data cleaning, data standardization, and data conversion into a format suitable for model training; the real-time monitoring data includes indoor and outdoor temperature, humidity, passenger flow and CO2 concentration.

[0058] The air conditioning system is a comprehensive system, including two aspects: an air conditioning refrigeration system and an air conditioning heating system.

[0059] Among them, the refrigeration cycle and heating cycle are the core parts of the system, which are used to achieve cooling and heating effects respectively. In addition, it also includes the air supply and return air system, which is used to circulate air and adjust indoor air flow. In order to meet the specific needs of the building, auxiliary equipment is independently defined according to the actual situation, including: fans, water pumps, boilers, and the selection and layout of these equipment are determined according to the characteristics and needs of the building.

[0060] like Figure 3 As shown, in step P3, the steps of real-time simulation of the intelligent optimization model are as follows:

[0061] W1. Write an intelligent optimization algorithm in Python and set the initial relevant environmental parameters and air conditioning system related parameters;

[0062] The relevant parameters include indoor and outdoor temperature, humidity, passenger flow and CO2 concentration;

[0063] The air-conditioning system related parameters include: the target temperature range when the human body is comfortable, the air-conditioning cooling setting temperature range, and the air-conditioning heating setting temperature range; the target temperature range when the human body is comfortable is set to 22-26°C, the air-conditioning cooling setting temperature range is 21-27°C, and the air-conditioning heating setting temperature range is 17-23°C.

[0064] The air conditioning system related parameters include the target temperature range for human comfort, the air conditioning cooling setting temperature range and the air conditioning heating setting temperature range;

[0065] W2. Based on the preprocessed data set obtained in step P2, a temperature and energy consumption prediction model is constructed using a deep learning algorithm, and the model is iteratively trained. After the training is completed, the model is verified and evaluated;

[0066] Use the deep learning algorithm and the preprocessed data set to train the indoor temperature and energy consumption prediction model, iteratively update the parameters of the temperature prediction model and the energy consumption prediction model until the preset number of iterations is reached. After the training is completed, the performance of the prediction model is evaluated by the root mean square error and the mean absolute error. After the model is usable, enter step W3;

[0067] W3, substitute the target temperature and the minimum value of the air conditioning setting temperature range, use the temperature and energy consumption prediction model to predict the indoor temperature and energy consumption, and adjust the air conditioning system according to the indoor temperature until the set target temperature is reached, and output the total energy consumption under the current setting;

[0068] Substitute the target temperature and the minimum value of the air conditioner setting temperature range, call the temperature prediction model to predict the indoor temperature under the current air conditioning system parameter settings hour by hour, and call the energy consumption prediction model to predict the load under the relevant parameter settings of the air conditioning system hour by hour, then compare the difference between each predicted temperature and the target temperature, and judge whether the current predicted temperature reaches the set target temperature. If not, adjust the air conditioner setting temperature, accumulate the energy consumption and make a prediction at the next moment until the indoor temperature reaches the set target temperature, and record all the energy consumption generated in the process and the number of uncomfortable hours required to reach the target temperature; if so, output the total energy consumption under the current setting;

[0069] W4. Based on step W2, traverse the air conditioning setting temperature interval and the target temperature interval in turn, adjust the target temperature value and the air conditioning temperature setting value and compare them with the previous ones until the optimal air conditioning setting scheduling plan is found.

[0070] In step W3, the indoor temperature is adjusted by the air conditioning system as follows:

[0071] W3.1, the intelligent optimization model first determines whether the outdoor temperature is greater than the current target temperature, if not, it goes to step W3.2, if yes, it goes to step W3.3;

[0072] W3.2, determine whether the indoor temperature is close to the target temperature, if not, proceed to step W3.4, if yes, proceed to step W3.5;

[0073] W3.3, determine whether the indoor temperature is close to the target temperature, if not, go to step W3.6, if yes, go to step W3.5;

[0074] W3.4, determine whether the indoor temperature is greater than the target temperature, if not, the air conditioning heating temperature is +1, if yes, the air conditioning heating temperature is -1;

[0075] W3.5, keep the current settings;

[0076] W3.6, determine whether the indoor temperature is lower than the target temperature, if not, the air conditioning temperature is -1, if yes, the air conditioning temperature is +1;

[0077] W3.7. After adjusting the air conditioner setting temperature based on steps W3.1-W3.6, call the temperature and energy consumption prediction model again for prediction.

[0078] In step W4, the optimal air conditioning setting scheduling scheme is found as follows:

[0079] Determine whether to traverse the air-conditioning temperature setting range. If not, increase the air-conditioning setting temperature by 1, substitute the current target temperature and the air-conditioning setting temperature and perform step W3 again. If yes, continue to determine whether to traverse the target temperature range. If not, increase the target temperature by 1, substitute the current target temperature and the air-conditioning setting temperature and perform step W3 again. If yes, output the air-conditioning setting temperature and target temperature corresponding to the minimum energy consumption to obtain the optimal air-conditioning system scheduling plan.

[0080] The calculation error evaluation index includes root mean square error and mean absolute error; the data set division ratio is set according to actual needs;

[0081] The root mean square error is a statistic used to measure the difference between the model prediction value and the actual observation value. The calculation formula of the root mean square error RMSE is as follows:

[0082]

[0083] In formula (1), n ​​is the total number of observations; Y i is the i-th actual observation value; is the i-th predicted value;

[0084] The mean absolute error is a statistic used to measure the average error between the predicted value and the actual observed value. The calculation formula of the mean absolute error MAE is as follows:

[0085]

[0086] In formula (2), n is the total number of observations; Y i is the i-th actual observation value; is the i-th predicted value; |·| is the absolute value.

[0087] The temperature and energy consumption prediction model is constructed as follows:

[0088] First, the structure and parameters of the temperature and energy consumption prediction model are defined, that is, the network layer and the number of neurons of the model are set, the training set is trained using the back propagation algorithm, the parameters of the model are updated through iteration, and the appropriate loss function and optimizer (that is, the loss function and optimizer with the best fitting effect) are selected according to the actual situation, as well as the appropriate learning rate and batch size are set.

[0089] By calculating the root mean square error (RMSE), the squares of the prediction errors are averaged and then squared to obtain a non-negative error measurement value. The smaller the RMSE, the smaller the difference between the model's predicted value and the actual observed value, and the better the model's prediction effect.

[0090] By using SketchUp software to build a 3D model of a rail transit station that can be recognized by EnergyPlus, and setting up key environmental parameters for the air-conditioning system, a large amount of data generated by multiple simulation runs of the EnergyPlus software is obtained by modifying the key parameters of the air-conditioning system. The data set is aggregated with real-time monitoring data to form a data set, and then the optimal air-conditioning system scheduling plan is determined based on the preprocessed data set and the intelligent optimization model.

[0091] Based on simulation and deep learning algorithms, the dual-objective optimization problem of human comfort and indoor energy consumption is solved. Through real-time scheduling and feedback mechanisms, it adapts to dynamic changes in indoor and outdoor environments, and can adjust parameters according to actual needs. It has strong flexibility and customizability, fully considers human comfort and energy consumption, and realizes intelligent scheduling and efficient operation of the air-conditioning system. While improving user experience, it brings significant advantages in building energy conservation and environmental protection.

[0092] The adjustment of key parameters plays a vital role in indoor energy consumption and human comfort. Within the independently set parameter range, simulations are performed one by one according to the value step size, so as to obtain a large amount of data to fill the gaps in the actual collected data.

[0093] In summary, the reduction of building energy consumption is often accompanied by an increase in human discomfort time, that is, human comfort and energy consumption reduction are contradictory. The EnergyPlus software is used to simulate the operating conditions of real rail transit stations, and the large amount of data generated by the operation is combined with the actual collected data as a training set. The deep learning algorithm and the above data set are used to train the intelligent optimization model, and the real-time monitoring data is combined to realize the intelligent scheduling of the indoor air-conditioning system of transportation buildings.

[0094] By changing the key parameters of the air conditioner and performing multiple simulations to obtain a large amount of data, the data generated by the simulation operation of the software is combined with the actual collected data as the data set of the intelligent optimization model, thereby enriching the data set while reducing dependence on the actual operating status of the building.

[0095] Based on simulation and deep learning algorithms, the dual-objective optimization problem of human comfort and indoor energy consumption is solved. Through real-time scheduling and feedback mechanisms, it adapts to dynamic changes in indoor and outdoor environments, and can adjust parameters according to actual needs. It has strong flexibility and customizability, fully considers human comfort and energy consumption, and realizes intelligent scheduling and efficient operation of the air-conditioning system. While improving user experience, it brings significant advantages in building energy conservation and environmental protection.

[0096] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A scheduling optimization method for a rail transit station air conditioning system based on multiple objectives, characterized in that: The following steps are involved: P1. Collect architectural parameter information of rail transit stations, build a 3D model of the rail transit station recognized by EnergyPlus through SketchUp software, import the 3D model of the rail transit station and meteorological data files into EnergyPlus software to build the air conditioning system, and set the basic structural parameters of the rail transit station; P2. Modify the key parameters of the air conditioning system, obtain the data generated by the simulation operation of the EnergyPlus software, aggregate the data generated by the simulation operation with the real-time monitoring data and perform pre-processing operations to form a data set for intelligent optimization model training; P3. Build an intelligent optimization model for the air-conditioning system. Divide the data set preprocessed in step P2 into a test set and a validation set in proportion. Use the test set data for iterative training. Use the validation set data to calculate the error evaluation index for evaluation. After the model is usable, use the intelligent optimization model for real-time simulation and adjust the air-conditioning system until the air-conditioning system scheduling plan is determined. P4. Obtain real-time monitoring data through indoor and outdoor temperature sensor monitoring equipment and input it into the intelligent optimization model. After autonomously setting the target temperature range and air conditioning setting temperature range, output the air conditioning system scheduling plan.

2. A scheduling optimization method based on a multi-objective rail transit station air conditioning system as claimed in claim 1, characterized in that: In step P1, the rail transit station is a rail transit building, and the building parameter information includes geometric information of the building, passenger flow, and station area; The EnergyPlus software imports the 3D model of the transportation station and the meteorological data file, and sets the basic structural parameters of the rail transit station; The basic structural parameters include the building's envelope, room orientation, geographic information, occupant density, and power-consuming equipment information.

3. The scheduling optimization method based on a multi-objective rail transit station air conditioning system according to claim 1, characterized in that: In step P1, the meteorological data are meteorological parameters and weather phenomena for each hour in a year.

4. The scheduling optimization method based on a multi-objective rail transit station air conditioning system according to claim 1, characterized in that: In step P2, the key parameters include the air conditioning cooling setting temperature and the air conditioning heating setting temperature, the preprocessing includes data cleaning, data standardization, and data conversion into a format for model training; the real-time monitoring data includes indoor and outdoor temperature, humidity, passenger flow and CO2 concentration.

5. The scheduling optimization method based on a multi-objective rail transit station air conditioning system according to claim 1, characterized in that: In step P3, the steps of real-time simulation of the intelligent optimization model are as follows: W1. Write an intelligent optimization algorithm in Python and set the initial relevant environmental parameters and air conditioning system related parameters; The relevant parameters include indoor and outdoor temperature, humidity, passenger flow and CO2 concentration; The air conditioning system related parameters include the target temperature range for human comfort, the air conditioning cooling setting temperature range and the air conditioning heating setting temperature range; W2. Based on the preprocessed data set obtained in step P2, a temperature and energy consumption prediction model is constructed using a deep learning algorithm, and the model is iteratively trained. After the training is completed, the model is verified and evaluated; Use the deep learning algorithm and the preprocessed data set to train the indoor temperature and energy consumption prediction model, iteratively update the parameters of the temperature prediction model and the energy consumption prediction model until the preset number of iterations is reached. After the training is completed, the effect of the prediction model is evaluated by the root mean square error and the mean absolute error. After the model is usable, enter step W3; W3, substitute the target temperature and the minimum value of the air conditioning setting temperature range, use the temperature and energy consumption prediction model to predict the indoor temperature and energy consumption, and adjust the air conditioning system according to the indoor temperature until the set target temperature is reached, and output the total energy consumption under the current setting; W4. Based on step W2, traverse the air conditioning setting temperature range and the target temperature range in turn, adjust the target temperature value and the air conditioning temperature setting value and compare them with the previous ones until the air conditioning setting scheduling plan is found.

6. A scheduling optimization method based on a multi-objective rail transit station air conditioning system as claimed in claim 5, characterized in that: In step W3, the indoor temperature is adjusted by the air conditioning system as follows: W3.1, the intelligent optimization model first determines whether the outdoor temperature is greater than the current target temperature, if not, it goes to step W3.2, if yes, it goes to step W3.3; W3.2, determine whether the indoor temperature is close to the target temperature, if not, proceed to step W3.4, if yes, proceed to step W3.5; W3.3, determine whether the indoor temperature is close to the target temperature, if not, go to step W3.6, if yes, go to step W3.5; W3.4, determine whether the indoor temperature is greater than the target temperature, if not, the air conditioning heating temperature is +1, if yes, the air conditioning heating temperature is -1; W3.5, keep the current settings; W3.6, determine whether the indoor temperature is lower than the target temperature, if not, the air conditioning temperature is -1, if yes, the air conditioning temperature is +1; W3.

7. After adjusting the air conditioner setting temperature based on steps A31-A36, the temperature and energy consumption prediction model is called again for prediction.

7. A rail transit station air conditioning system scheduling method based on multi-objective optimization as claimed in claim 5, characterized in that: In step W4, the optimal air conditioning setting scheduling scheme is found as follows: Determine whether the air-conditioning temperature setting range has been traversed. If not, the air-conditioning setting temperature is increased by 1, and the current target temperature and the air-conditioning setting temperature are substituted and step W3 is performed again. If so, continue to determine whether the target temperature range has been traversed. If not, the target temperature is increased by 1, and the current target temperature and the air-conditioning setting temperature are substituted and step W3 is performed again. If so, output the air-conditioning setting temperature and target temperature corresponding to the minimum energy consumption to obtain the air-conditioning system scheduling plan.

Images