Urban rail electromechanical energy-saving prediction control method and system based on adaptive model

Abstract

The invention discloses an urban rail electromechanical energy-saving predictive control method and system based on a self-adaptive model. The system constructs an external environment, equipment and building energy consumption model of a subway station ventilation and air conditioning system through an Informer time sequence algorithm and a Trnsys-Fluent joint simulation model. According to the requirements of the energy-saving system for improving human body comfort and reducing building operation energy consumption, a target cost function is established, a target cost function value is calculated by simulating data such as temperature and humidity in a station and equipment and building energy consumption in combination with prediction data of an external environment and a simulation model, and an optimal solution is searched by using a multi-target particle swarm algorithm. Finally, the optimal solution is issued to a subway station ventilation air-conditioning system, and the purposes of energy conservation and carbon reduction are achieved while the comfortable experience of passengers is ensured. According to the innovative system, the energy efficiency level of the subway station is remarkably improved, and powerful technical support is provided for energy conservation, emission reduction and intelligent management of urban rail transit.

Description

Technical Field

[0001] The invention relates to a subway air conditioning control system, and in particular to an urban rail electromechanical energy-saving prediction control method and system based on an adaptive model. Background Art

[0002] As the scale of urban rail transit operations in my country continues to expand, the energy consumption of air conditioning systems has become increasingly prominent, accounting for 30% to 40% of total energy consumption. Improving the operating efficiency of air conditioning systems is crucial to reducing the operating costs of subway stations and promoting the transformation of the rail transit industry towards a green and low-carbon direction.

[0003] Since the passenger flow in subway stations varies significantly over time, the cooling demand fluctuates rapidly, which requires the control system to be able to adjust the operating parameters of air-conditioning equipment in real time. At present, subway stations generally use model-free, rule-based control methods to adjust the operating parameters of air-conditioning equipment. However, these methods lack global optimization considerations for the air-conditioning system and are difficult to achieve optimal control effects. Summary of the invention

[0004] Purpose of the invention: In view of the above problems, the present invention provides a method and system for predictive control of urban rail electromechanical energy saving based on an adaptive model, which can adaptively adjust the parameters of the controlled equipment in the air-conditioning system in the subway, solving the problems of high energy consumption of the electromechanical system in subway stations, large fluctuations in ambient temperature in public areas, and high energy consumption in equipment areas.

[0005] Technical solution: The technical solution adopted by the present invention is a predictive control method for urban rail electromechanical energy saving based on an adaptive model, comprising:

[0006] Step 1: Build the external environment model of the subway station based on the Informer time series model; build the subway station ventilation and air-conditioning system equipment and building energy consumption model based on the Trnsys-Fluent model, set the prediction interval step size and number of iterations, and initialize the model parameters, control variables, environmental data, and boundary condition data based on the subway station ventilation and air-conditioning system.

[0007] Step 2: Initialize the control quantity and initialize the environmental data as the input of the Informer time series model. The Informer time series model reads the external environment model of the subway station to predict the environmental data, and passes the predicted environmental data to the Trnsys-Fluent model.

[0008] Step 3, input the predicted environmental data and the initialization control quantity into the Trnsys-Fluent model, start the Trnsys kernel in the Trnsys-Fluent model, and start the simulation after the Trnsys kernel reads the ventilation and air-conditioning system equipment and the building energy consumption model of the subway station and initializes the model parameters; then the Trnsys kernel starts the Fluent kernel, and the Fluent kernel reads the fluid model and the initialization boundary condition data in the subway station to perform fluid simulation calculations, and returns the calculation results to the Trnsys kernel to complete the remaining calculations. The Trnsys kernel calculates the hourly energy consumption of the electromechanical system equipment of the subway station and outputs the cost function value;

[0009] Step 4, start the multi-objective particle swarm optimization algorithm MOPSO, and the MOPSO algorithm calculates the optimal control amount according to the current cost function value;

[0010] Step 5, use the model parameters of the real-time subway station ventilation and air-conditioning system equipment and the building energy consumption model as the initialization model parameters for the next calculation, use the actually measured environmental data as the initialization environmental data for the next calculation, use the optimal control amount as the initial control amount for the next calculation, and use the updated boundary condition data as the initial boundary condition data for the next calculation, and return to step 2 for a loop calculation.

[0011] Furthermore, the external environment model of the subway station is constructed based on the Informer time series model. According to the historical data of temperature, humidity, passenger flow, energy consumption, etc. collected by the urban rail ventilation and air-conditioning system, the Informer time series model is used to train, verify, test and deploy reasoning on the pre-processed historical data, and the predicted data is passed to the Trnsys-Fluent model as input, which includes the following steps:

[0012] Step 1: Data collection and preprocessing: collect historical data such as temperature, humidity, passenger flow, and energy consumption of the ventilation and air-conditioning system of urban rail stations, and preprocess the historical data (data cleaning, standardization, data set division, etc.) to convert it into a data set that can be used by the model;

[0013] Step 2, model training. According to the characteristics and requirements of the dataset, select the appropriate Informer model configuration, including hyperparameters such as time step, learning rate, and number of model layers. Input the data in the training set into the model for training according to the input shape required by the model. When the number of loop iterations is less than the target number of iterations, send the data into the model for forward propagation and backward propagation, and use the optimizer to update the model parameters. At the same time, record the loss value and other indicators during the training process. Evaluate the model performance on the validation set after each training cycle. After training is completed, use the test set to evaluate the final performance of the model. After parameter adjustment and optimization, save the optimal weight model;

[0014] Deploy the trained model to the server, update the data within the set prediction time domain every day, and set up a monitoring program to continuously monitor the error between the model prediction value and the actual collected value. When the monitoring program evaluates that the error value is too large, notify the person in charge of the energy-saving system to correct the model in time;

[0015] Part of the data predicted by the Informer algorithm is used for data analysis pages such as data display and anomaly detection, and the other part is input into Trnsys as initialization parameters to participate in simulation calculations.

[0016] Furthermore, the building load module is constructed using Fluent, which uses numerical simulation methods to numerically calculate the flow field information such as temperature, humidity, and velocity distribution in all areas of the station, including the following steps:

[0017] Step 1: Preprocessing: Based on the CAD drawing of the subway station and combined with the measured data on site, a three-dimensional model is established through SpaceClaim, including the main building structures such as the station top air vents, return air vents, entrances and exits, columns, isolation walls, shield doors, elevators and other structural features. Extract the fluid volume, set up the model structure tree, and use Fluent Meshing to divide the preprocessed flow field area into polyhedral grids; in order to improve the calculation accuracy, reduce the overall calculation scale, and increase the solution speed, adjust the grid density for small-sized structures such as air vents and return air vents, and appropriately increase the polyhedral grid volume for the remaining areas, but the maximum grid side length should not exceed 1 / 3 of the minimum building structure length; reduce the total number of grids by adjusting the grid division as above. In principle, the optimized grid skewness is no more than 0.7, and the minimum orthogonal quality is as large as possible. 0.1;

[0018] Step 2: Setting parameters: According to the computational fluid model constructed in step 1, various parameters are set, including physical parameters for calculation, parameters related to the Fluent solution algorithm, various physical properties of solids and fluids, calculation iteration parameters, model boundary conditions and calculation accuracy, etc., and the calculation result output structure is set;

[0019] Step 3: Model accuracy test: First, select a period of typical measured environmental data, set the initial values ​​of each environment, and then start the simulation. Use the general drawing and analysis software TECPLOT to visualize the calculation results; analyze the distribution of the flow field in the station based on the calculation results, as well as the temperature changes in the area where the actual temperature sensor is located and the changes in the overall average temperature of the environment, and compare with the measured data to determine whether the calculation results meet the requirements.

[0020] Trnsys and Fluent are jointly simulated, and the two are dynamically combined. That is, in each time step of Trnsys, Trnsys starts the Fluent kernel and transmits the initial conditions required for calculation such as supply air temperature, humidity and wind speed through the interface program. At the same time, local batch files are set in advance for Fluent to read. After Fluent completes the initialization settings, it performs calculations and returns the calculation results (temperature and humidity in the area where the sensor is located, average temperature and humidity in the station, return air outlet temperature and humidity and wind speed, etc.) to Trnsys through the interface to complete the calculation of the remaining modules. After completion, it enters the simulation of the next time step.

[0021] Furthermore, the cost function and constraints are set, and the difference between the actual temperature and humidity in the public area of ​​the station and the equipment management room and the set temperature and humidity is minimized, and the total energy consumption of the ventilation and air-conditioning system equipment is minimized as the optimization goal. The predictive control cost function is established, and the constraints such as the working conditions of the equipment are determined;

[0022] The objective function expression of the temperature and humidity error model is:

[0023]

[0024] Where F1(t) represents the temperature and humidity error cost function, T amb and T set are the measured and set temperatures, RH amb and RH set are the measured and set humidity, respectively, i is the weight parameter, n is the total number of devices;

[0025]

[0026] Where F2(t) represents the total energy consumption cost function, P i (·) is the power of the i-th device in the ventilation and air-conditioning system, μ j is the weight parameter, i is the device number, j is the number of the weight parameter corresponding to the i-th device, n is the total number of devices, t is the time, f is the device frequency, and ΔT is the sampling period;

[0027]

[0028] Where P1 is the input power of the chiller, Q e is the cooling capacity, T eo is the chilled water outlet temperature, T ci is the cooling water return temperature, A, B, C are the parameters obtained by fitting the measured power, cooling capacity, evaporation and condensation temperature data;

[0029] P2=(a0+a1f+a2f 2 +…)P r

[0030] Where P2 is the pump / fan power, P r is the effective power of the water pump / fan, f is the motor frequency, a0, a1, a2, ... are all coefficients obtained by fitting the measured equipment power and equipment frequency data;

[0031] The constraints are as follows:

[0032]

[0033]

[0034] In the formula, f max With f min are the upper and lower limits of the original motor operating frequency of the equipment, Q1, Q2, Q3 are the rated air volumes of the air-conditioning unit, the exhaust fan and the fresh air fan respectively, f is the air-conditioning fan frequency, g1(f) is the minimum frequency constraint, and g2(f) is the maximum frequency constraint.

[0035] Furthermore, the multi-objective particle swarm optimization algorithm (MOPSO) is used to solve the Pareto optimal solution set, which includes the following steps:

[0036] Step 1: Initialize the speed and position of all particles and calculate the objective function values ​​of all particles;

[0037] Step 2: Store the non-inferior solutions in an external document to form a Pareto optimal solution set;

[0038] Step 3: Sort by dense distance to determine the optimal individual and population;

[0039] Step 4: Update the inertia weight, particle position vector and velocity vector;

[0040] Step 5: Calculate the fitness function value of each particle after the update and update the Pareto optimal solution set;

[0041] Step 6: Update individual optimality and population optimality;

[0042] Step 7 determines whether the convergence condition is met: if so, output the result and the number of iterations; otherwise, return to step 3.

[0043] The present invention proposes an urban rail electromechanical energy-saving prediction control system based on an adaptive model, including an adaptive MPC controller; the adaptive MPC controller includes:

[0044] The Informer environmental data prediction module stores the external environmental model of the subway station built based on the Informer time series model. The optimal control quantity calculated last time and the environmental data obtained by actual measurement are used as the input of the Informer time series model. The Informer time series model reads the external environmental model of the subway station to predict the environmental data, and passes the predicted environmental data to the Trnsys-Fluent model.

[0045] The Trnsys-Fluent joint dynamic simulation module stores the subway station ventilation and air-conditioning system equipment and building energy consumption model based on the subway station ventilation and air-conditioning system and constructed based on the Trnsys-Fluent model. The Trnsys-Fluent joint dynamic simulation module inputs the predicted environmental data and the optimal control quantity calculated last time into the Trnsys-Fluent model, starts the Trnsys kernel in the Trnsys-Fluent model, and the Trnsys kernel reads the subway station ventilation and air-conditioning system equipment and building energy consumption model and the model parameters calculated last time and starts the simulation; then the Trnsys kernel starts the Fluent kernel, and the Fluent kernel reads the fluid model in the subway station and the initialization boundary condition data to perform fluid simulation calculation, and returns the calculation results to the Trnsys kernel to complete the remaining calculations, obtain the updated boundary condition data and the hourly energy consumption of the metro station electromechanical system equipment, and output the cost function value;

[0046] MOPSO solver, which stores the multi-objective particle swarm optimization algorithm MOPSO. The MOPSO algorithm calculates the optimal control amount according to the current cost function value;

[0047] The adaptive MPC controller sends the optimal control quantity to each controlled device.

[0048] Furthermore, the subway station ventilation and air conditioning system equipment and building energy consumption model includes:

[0049] Input module, used to input environmental data predicted by the Informer algorithm;

[0050] The building load module is built based on Fluent and is used to simulate the flow field distribution in the station through numerical simulation methods, and output the temperature, humidity and wind speed distribution in the station;

[0051] The wind system module, including the variable air volume fan module, the air duct module, and the surface cooler module, is used to simulate the air supply volume, temperature, humidity, and equipment energy consumption data of the subway station wind system;

[0052] The water system module includes the chiller module, water pump module, cooling tower module, water pipe module, and confluence and diversion module, which is used to simulate the supply and return water temperature, supply and return water flow, cooling capacity, and equipment energy consumption data of the subway station water system;

[0053] The control module is used to generate control quantities and send them to the control devices in the wind and water systems to achieve global control of the subway station wind and water systems.

[0054] Beneficial effects: The present invention simulates the ventilation and air-conditioning system of the subway station through Trnsys, and uses Fluent to construct a building load module instead of a building module, which can not only accurately simulate the operating status and energy consumption trend of the ventilation and air-conditioning system, but also clearly feedback the air flow and temperature and humidity distribution of the station environment; introduces the Informer time series model to provide external environmental parameters, and improves the efficiency of solving the cost function under certain conditions; establishes a multi-objective optimization problem based on minimizing the environmental temperature and humidity error and minimizing the system energy consumption, and uses a multi-objective particle swarm optimization algorithm to calculate the Pareto optimal solution set and continuously update it online, and applies the Pareto optimal solution to the actual production environment, so as to achieve efficient and stable operation of the ventilation and air-conditioning system equipment, and provide a more comfortable riding environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a structural block diagram of the energy-saving prediction control system of the present invention;

[0056] Figure 2 It is a flow chart of the prediction results of the monitoring model of the present invention;

[0057] Figure 3 It is a structural schematic diagram of the subway station ventilation and air-conditioning system model based on Trnsys of the present invention. DETAILED DESCRIPTION

[0058] The present invention will be further described below in conjunction with the accompanying drawings.

[0059] Example 1

[0060] The urban rail electromechanical energy-saving prediction and control method based on an adaptive model described in the present invention is an energy-saving control method for the overall air-conditioning ventilation and air-conditioning system of a subway station during the air-conditioning season, including: constructing an external environment model of the subway station based on the Informer timing model, and constructing a subway station ventilation and air-conditioning system equipment and building energy consumption model based on the Trnsys-Fluent joint dynamic simulation model; obtaining predicted values ​​of outdoor temperature and humidity, passenger flow, equipment heat generation and control quantity through the external environment model of the subway station based on the measured data of ambient temperature and humidity and the energy consumption of the urban rail station ventilation and air-conditioning system; and conducting simulation predictions with the subway station ventilation and air-conditioning system equipment and building energy consumption models. Figure 1The figure shows the structure block diagram of the energy-saving prediction control system corresponding to this method. The entire simulation prediction calculation cycle is:

[0061] Step 1: Build the external environment model of the subway station based on the Informer time series model; build the subway station ventilation and air-conditioning system equipment and building energy consumption model based on the Trnsys-Fluent model, set the prediction interval step size and number of iterations, and initialize the model parameters, control variables, environmental data, and boundary condition data based on the subway station ventilation and air-conditioning system.

[0062] Step 2: Initialize the control quantity and initialize the environmental data as the input of the Informer time series model. The Informer time series model reads the external environment model of the subway station to predict the environmental data, and passes the predicted environmental data to the Trnsys-Fluent model.

[0063] Step 3, input the predicted environmental data and the initialization control quantity into the Trnsys-Fluent model, start the Trnsys kernel in the Trnsys-Fluent model, and start the simulation after the Trnsys kernel reads the ventilation and air-conditioning system equipment and the building energy consumption model of the subway station and initializes the model parameters; then the Trnsys kernel starts the Fluent kernel, and the Fluent kernel reads the fluid model and the initialization boundary condition data in the subway station to perform fluid simulation calculations, and returns the calculation results to the Trnsys kernel to complete the remaining calculations. The Trnsys kernel calculates the hourly energy consumption of the electromechanical system equipment of the subway station and outputs the cost function value;

[0064] Step 4, start the multi-objective particle swarm optimization algorithm MOPSO, and the MOPSO algorithm calculates the optimal control amount according to the current cost function value;

[0065] Step 5: Send the optimal control quantity to each controlled device;

[0066] Step 6, use the model parameters of the real-time subway station ventilation and air-conditioning system equipment and the building energy consumption model as the initialization model parameters for the next calculation, use the actually measured environmental data as the initialization environmental data for the next calculation, use the optimal control amount as the initial control amount for the next calculation, and use the updated boundary condition data as the initial boundary condition data for the next calculation, and return to step 2 for a loop calculation.

[0067] The external environment model of the subway station is constructed based on the Informer time series model. According to the historical data of temperature, humidity, passenger flow, energy consumption, etc. collected by the urban rail ventilation and air-conditioning system, the Informer time series model is used to train, verify, test and deploy reasoning on the pre-processed historical data to obtain the external environment model of the subway station. The data predicted by the external environment model of the subway station is passed to the Trnsys-Fluent model as one of the inputs. The details are as follows:

[0068] (1) Data collection and preprocessing: collect historical data such as temperature, humidity, passenger flow, and energy consumption of the ventilation and air-conditioning system of urban rail stations, and preprocess the historical data (data cleaning, standardization, data set division, etc.) to convert them into data sets that can be used by the model; extract data features, including time features (year, month, day, hour, minute, quarter, week, whether it is a weekend, whether it is a holiday, etc.) and correlation features (temperature, humidity, passenger flow, etc.);

[0069] (2) Model training: According to the characteristics and requirements of the dataset, select the appropriate Informer model configuration, including hyperparameters such as time step, learning rate, and number of model layers. Input the data in the training set into the model for training according to the input shape required by the model. When the number of loop iterations is less than the target number of iterations, send the data into the model for forward propagation and backward propagation, and use the optimizer to update the model parameters. At the same time, record the loss value and other indicators during the training process. Evaluate the model performance on the validation set after each training cycle. After training is completed, use the test set to evaluate the final performance of the model. After parameter adjustment and optimization, save the optimal weight model;

[0070] Taking the hourly prediction model as an example, the model input data shape is (B, L, N), and the output data shape is (C, N), where B represents the batch size, L represents the input sequence length, N represents the number of features, and C represents the output sequence length; the model meets the required evaluation criteria: the accuracy rate is greater than 95%;

[0071] (3) Figure 2 As shown, the trained model is deployed on the server, the data in the set prediction time domain is updated, and a monitoring program is set to continuously monitor the error between the model prediction value and the actual collection value. When the monitoring program evaluates that the error value is too large, the interaction between the prediction model and the simulation model is cut off. Subsequently, the historical environmental data is used as the simulation model input, and the person in charge of the energy-saving system is notified to correct the model in time.

[0072] (4) Part of the data predicted by the Informer time series model is input into the Trnsys-Fluent model as environmental parameters, and the other part is used for functions such as data display and anomaly detection.

[0073] Based on the ventilation and air-conditioning system of subway stations, a model of the ventilation and air-conditioning system equipment and building system of subway stations was constructed based on Trnsys. The load submodule in the model was reconstructed using Fluent, and combined with the data predicted by the Informer algorithm, the hourly energy consumption of the ventilation and air-conditioning system equipment of subway stations throughout the year was simulated and calculated.

[0074] The ventilation and air-conditioning system of subway stations is divided into wind system and water system. The wind system is further divided into the ventilation and air-conditioning system of the public area of ​​the station (large system) and the ventilation and air-conditioning system of the equipment management room of the station (small system). The wind system equipment includes air-conditioning units, exhaust fans, air valves, air ducts, related sensors, dynamic balancing valves, etc. A station has multiple small systems depending on the location. The water system equipment includes chillers, chilled water pumps, cooling water pumps, cooling towers, water valves, water pipes, related sensors, etc.

[0075] Based on the above system, a Trnsys-Fluent joint dynamic simulation model is constructed, which consists of input module, building load module, wind system module, water system module, control module and calculation output module;

[0076] (1) The input module consists of the Informer timing model and the Trnsys input file. The Trnsys input file contains some static parameters. The Informer timing model predicts parameters such as temperature, humidity, and control quantity, providing initialization parameters for the simulation model.

[0077] (2) The building load module is constructed using Fluent, which uses numerical simulation methods to simulate the airflow field distribution conditions such as temperature, humidity, and velocity distribution in the station. The specific steps are as follows:

[0078] Preprocessing, based on the CAD drawing of the subway station, combined with the measured data on site, a three-dimensional model is established through SpaceClaim, including the main building structures such as the station top air vents, return air vents, entrances and exits, columns, isolation walls, shield doors, elevators and other structural features. The fluid volume is extracted, the model structure tree is set, and Fluent Meshing is used to divide the preprocessed flow field area into polyhedral grids. In order to improve the calculation accuracy, reduce the overall calculation scale, and increase the solution speed, the grid density is adjusted for small-sized structures such as air vents and return air vents, and the polyhedral grid volume is appropriately increased for the remaining areas, but the maximum grid side length should not exceed 1 / 3 of the minimum building structure length. The total number of grids is reduced by adjusting the grid division as above. In principle, the optimized grid skewness is not greater than 0.7, and the minimum orthogonal quality is as large as possible. 0.1;

[0079] Set parameters. According to the computational fluid model constructed in step 1, set various parameters for it, including physical parameters for calculation, parameters related to the Fluent solution algorithm, various physical properties of solids and fluids, calculation iteration parameters, model boundary conditions and calculation accuracy, etc.; set the calculation result output structure;

[0080] To test the accuracy of the model, we first selected a period of typical measured environmental data, set the initial values ​​of each environment, and then started the simulation. We visualized the calculation results using the general drawing and analysis software TECPLOT; based on the calculation results, we analyzed the distribution of the flow field in the station, as well as the temperature changes in the area where the actual temperature sensor is located and the changes in the overall average temperature of the environment, and compared them with the measured data to determine whether the accuracy of the calculation results meets the requirements;

[0081] Trnsys and Fluent are jointly simulated. The two are dynamically combined. That is, in each time step of Trnsys, Trnsys starts the Fluent kernel and transmits the initial conditions required for calculation such as supply air temperature, humidity and wind speed through the interface program. At the same time, the local batch file is set in advance for Fluent to read. After Fluent completes the initialization settings, it calculates and returns the calculation results (temperature and humidity in the area where the sensor is located, average temperature and humidity in the station, temperature and humidity at the return air outlet and wind speed, etc.) to Trnsys through the interface to complete the calculation of the remaining modules. After completion, it enters the simulation of the next time step;

[0082] (3) The air system module consists of the variable air volume fan model, air duct model, and surface cooler model in Trnsys. It simulates the subway station air conditioning units, return exhaust fans, air ducts and other equipment, and calculates data such as air supply volume, temperature, humidity, and equipment energy consumption;

[0083] (4) The water system module consists of a chiller module, a water pump module, a cooling tower module, a water pipe module, a water collector and a water distributor. It simulates the water system chiller, chilled water pump, cooling water pump, cooling tower, water pipe and other equipment of the subway station, and calculates the data such as the supply and return water temperature, supply and return water flow, cooling capacity, and equipment energy consumption;

[0084] (5) The control module is composed of a controller, which selects the control quantity matrix according to the optimal solution set and sends it to the control equipment in the Trnsys system to simulate the control effect. Finally, it is applied to the actual subway station to realize the global control of the ventilation and air-conditioning system.

[0085] (6) The calculation output module includes a calculator, an integrator, etc. for calculating the equipment operation data, and displays the energy-saving control effect in a visual form through online drawing, data display and data storage in the display module.

[0086] Set the cost function and constraints, take the minimum difference between the actual temperature and humidity in the station public area and equipment management room and the set temperature and humidity, and the minimum total energy consumption of ventilation and air conditioning system equipment as the optimization goal, establish the predictive control cost function, and determine the constraints such as the working conditions of the equipment;

[0087] The objective function expression of the temperature and humidity error model is:

[0088]

[0089] Where F1(t) represents the temperature and humidity error cost function, T amb and T set are the measured and set temperatures, RH amb and RH set are the measured and set humidity, respectively, i is the weight parameter, n is the total number of devices;

[0090]

[0091] Where F2(t) represents the total energy consumption cost function, P i (·) is the power of the i-th device in the ventilation and air-conditioning system, μ j is the weight parameter, i is the device number, j is the number of the weight parameter corresponding to the i-th device, n is the total number of devices, t is the time, f is the device frequency, and ΔT is the sampling period;

[0092]

[0093] Where P1 is the input power of the chiller, Q e is the cooling capacity, T eo is the chilled water outlet temperature, T ci is the cooling water return temperature, A, B, C are the parameters obtained by fitting the measured power, cooling capacity, evaporation and condensation temperature data;

[0094] P2=(a0+a1f+a2f+…)P r

[0095] Where P2 is the pump / fan power, P r is the effective power of the water pump / fan, f is the motor frequency, a0, a1, a2, ... are all coefficients obtained by fitting the measured equipment power and equipment frequency data;

[0096] The constraints are as follows:

[0097]

[0098]

[0099] In the formula, f max With f min are the upper and lower limits of the original motor operating frequency of the equipment, Q1, Q2, Q3 are the rated air volumes of the air-conditioning unit, the exhaust fan and the fresh air fan respectively, f is the air-conditioning fan frequency, g1(f) is the minimum frequency constraint, and g2(f) is the maximum frequency constraint.

[0100] The multi-objective particle swarm optimization algorithm (MOPSO) is used to solve the Pareto optimal solution set, which includes the following steps:

[0101] (1) Initialize the speed and position of all particles and calculate the objective function values ​​of all particles;

[0102] (2) Store the non-inferior solutions in an external document to form a Pareto optimal solution set;

[0103] (3) Sort by density distance to determine the optimal individual and population;

[0104] (4) Update the inertia weight, particle position vector and velocity vector;

[0105] (5) Calculate the fitness function value of each particle after the update and update the Pareto optimal solution set;

[0106] (6) Update individual optimality and population optimality;

[0107] (7) Determine whether the convergence condition is met: if so, output the result and the number of iterations; otherwise, return to (3).

[0108] Using the MOPSO optimization controller to control the environmental control system includes the following steps:

[0109] (1) In the current control time domain, the MOPSO optimizer starts the hybrid model and obtains the optimal solution as the optimal control quantity at the current moment through simulation calculation;

[0110] (2) Send the optimal control quantity to the PLC through the network communication protocol to control the corresponding equipment;

[0111] (3) The current environmental parameters are collected through the environmental control system and used as the initial values ​​for the next calculation, which are returned to the MOPSO optimizer and return to (1);

[0112] While satisfying the comfort of subway passengers, it reduces the operating energy consumption and operating costs of the ventilation and air-conditioning system for subway operations, and improves the social benefits of energy conservation and emission reduction.

[0113] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0114] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0115] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0116] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0117] Example 2

[0118] The urban rail electromechanical energy-saving prediction control system based on the adaptive model of the present invention has a structural schematic diagram as shown in FIG. Figure 1 As shown, the adaptive MPC controller includes: Trnsys-Fluent joint dynamic simulation module, Informer environmental data prediction module and MOPSO solver.

[0119] The Informer environmental data prediction module stores the external environmental model of the subway station built based on the Informer time series model. The optimal control quantity calculated last time and the environmental data obtained by actual measurement are used as the input of the Informer time series model. The Informer time series model reads the external environmental model of the subway station to predict the environmental data, and passes the predicted environmental data to the Trnsys-Fluent model.

[0120] The Trnsys-Fluent joint dynamic simulation module stores a subway station ventilation and air-conditioning system based on the Trnsys-Fluent model to build a subway station ventilation and air-conditioning system equipment and building energy consumption model. The Trnsys-Fluent joint dynamic simulation module inputs the predicted environmental data and the optimal control quantity calculated last time into the Trnsys-Fluent model, starts the Trnsys kernel in the Trnsys-Fluent model, and the Trnsys kernel reads the subway station ventilation and air-conditioning system equipment and building energy consumption model and the model parameters calculated last time to start simulation; then the Trnsys kernel starts the Fluent kernel, the Fluent kernel reads the fluid model in the subway station and the initialization boundary condition data to perform fluid simulation calculation, and returns the calculation results to the Trnsys kernel to complete the remaining calculations, obtain updated boundary condition data and the hourly energy consumption of the metro station electromechanical system equipment, and output the cost function value;

[0121] MOPSO solver stores the multi-objective particle swarm optimization algorithm MOPSO. The MOPSO algorithm calculates the optimal control amount based on the current cost function value.

[0122] The adaptive MPC controller sends the optimal control quantity to each controlled device.

[0123] like Figure 3 As shown, based on the subway station ventilation and air-conditioning system, a subway station ventilation and air-conditioning system equipment and building system model is constructed based on Trnsys. The load submodule in the model is reconstructed using Fluent, and combined with the data predicted by the Informer algorithm, the hourly energy consumption of the subway station ventilation and air-conditioning system equipment throughout the year is simulated and calculated.

[0124] The ventilation and air-conditioning system of subway stations is divided into wind system and water system. The wind system is further divided into the ventilation and air-conditioning system of the public area of ​​the station (large system) and the ventilation and air-conditioning system of the equipment management room of the station (small system). The wind system equipment includes air-conditioning units, exhaust fans, air valves, air ducts, related sensors, dynamic balancing valves, etc. A station has multiple small systems depending on the location. The water system equipment includes chillers, chilled water pumps, cooling water pumps, cooling towers, water valves, water pipes, related sensors, etc.

[0125] Based on the above system, a Trnsys-Fluent joint dynamic simulation model is constructed, which consists of an input module, a building load module, a wind system module, a water system module, a control module and a calculation output module.

[0126] The input module consists of the Informer timing model and the Trnsys input file. The Trnsys input file contains some static parameters. The Informer timing model predicts parameters such as temperature, humidity, and control quantity, providing initialization parameters for the simulation model.

[0127] The building load module is constructed using Fluent, which uses numerical simulation methods to simulate the airflow field distribution conditions such as temperature, humidity, and velocity distribution in the station;

[0128] The air system module consists of the variable air volume fan model, air duct model, and surface cooler model in Trnsys. It simulates the subway station air conditioning unit, exhaust fan, air duct, dynamic balancing valve and other equipment, and calculates the air supply volume, temperature, humidity, equipment energy consumption and other data.

[0129] The water system module consists of a chiller module, a water pump module, a cooling tower module, a water pipe module, a water collector and a water distributor module, etc. It simulates the water system chillers, chilled water pumps, cooling water pumps, cooling towers, water pipes and other equipment in subway stations, and calculates data such as the supply and return water temperature, supply and return water flow, cooling capacity, and equipment energy consumption;

[0130] The control module is composed of a controller. The control quantity matrix is ​​obtained through calculation and sent to the control equipment in the Trnsys system to simulate the control effect and verify whether the current solution is the optimal solution. Finally, it is applied to the actual subway station to realize the global control of the ventilation and air-conditioning system.

[0131] The calculation output module includes a calculator, an integrator, etc. for calculating the equipment operation data, and displays the energy-saving control effect in a visual form through online drawing, data display and data storage by the display module.

Claims

1. A predictive control method for urban rail electromechanical energy saving based on an adaptive model, characterized in that: include: Step 1: Build the external environment model of the subway station based on the Informer time series model; build the subway station ventilation and air-conditioning system equipment and building energy consumption model based on the Trnsys-Fluent model, set the prediction interval step size and number of iterations, and initialize the model parameters, control variables, environmental data, and boundary condition data based on the subway station ventilation and air-conditioning system. Step 2: Initialize the control quantity and initialize the environmental data as the input of the Informer time series model. The Informer time series model reads the external environment model of the subway station to predict the environmental data, and passes the predicted environmental data to the Trnsys-Fluent model. Step 3, input the predicted environmental data and the initialization control quantity into the Trnsys-Fluent model, start the Trnsys kernel in the Trnsys-Fluent model, and start the simulation after the Trnsys kernel reads the ventilation and air-conditioning system equipment and building energy consumption model of the subway station and initializes the model parameters; then the Trnsys kernel starts the Fluent kernel, and the Fluent kernel reads the fluid model in the subway station and the initialization boundary condition data to perform fluid simulation calculations, and returns the calculation results to the Trnsys kernel to complete the remaining calculations, obtain the updated boundary condition data and the hourly energy consumption of the electromechanical system equipment of the subway station, and output the cost function value; Step 4, start the multi-objective particle swarm optimization algorithm MOPSO, and the MOPSO algorithm calculates the optimal control amount according to the current cost function value; Step 5, use the model parameters of the real-time subway station ventilation and air-conditioning system equipment and the building energy consumption model as the initialization model parameters for the next calculation, use the actually measured environmental data as the initialization environmental data for the next calculation, use the optimal control amount as the initial control amount for the next calculation, and use the updated boundary condition data as the initial boundary condition data for the next calculation, and return to step 2 for a loop calculation.

2. The urban rail electromechanical energy-saving prediction control method based on the adaptive model according to claim 1 is characterized in that: The Trnsys-Fluent model is to reconstruct the building load submodule in the Trnsys kernel using the Fluent kernel, including the following steps: Step 21, based on the CAD drawing of the subway station and combined with the on-site measured data, a three-dimensional model is established through SpaceClaim, and the structural model includes the air supply vents, return vents, entrances and exits, isolation walls and elevators in the station; the fluid volume is extracted, and the polyhedral mesh is divided into convection field areas using FluentMeshing to obtain the fluid model in the subway station; Step 22, based on the constructed subway station fluid model, Fluent simulates and calculates the airflow field distribution through numerical simulation methods. The calculation results include the temperature, humidity and wind speed distribution in the station, and the calculation results are returned to the Trnsys kernel to complete the remaining calculations.

3. The urban rail electromechanical energy-saving prediction control method based on an adaptive model according to claim 1 is characterized in that: The external environment model of the subway station is constructed based on the Informer time series model, including: according to the historical data of temperature, humidity, passenger flow, and energy consumption collected by the urban rail ventilation and air-conditioning system, the pre-processed historical data is trained, verified, and tested using the Informer time series model to obtain the external environment model of the subway station.

4. The urban rail electromechanical energy-saving prediction control method based on the adaptive model according to claim 3 is characterized in that: Constructing a subway station external environment model based on the Informer time series model also includes: deploying the trained subway station external environment model to the server, updating the data in the set prediction time domain, and setting a monitoring program to continuously monitor the error between the predicted value and the actual collected value of the subway station external environment model. When the monitoring program evaluates that the error value is too large, correcting the hyperparameters of the subway station external environment model.

5. The urban rail electromechanical energy-saving prediction control method based on an adaptive model according to claim 1 is characterized in that: The reconstruction of the building load submodule in the Trnsys kernel by using the Fluent model includes: Trnsys and Fluent are mixed simulated in a dynamic combination manner: in each time step, the Trnsys model provides the boundary condition data required by the Fluent model, and the boundary condition data include the ambient temperature, the supply air volume in the station, and the supply air temperature and humidity. The Fluent model calculates according to the boundary condition data provided by the Tmsys model, and returns the calculation results to the Trnsys model for simulation of the next time step; the calculation results include the return air volume, return air temperature and humidity in the station.

6. The urban rail electromechanical energy-saving prediction control method based on an adaptive model according to claim 1 is characterized in that: The cost functions in the subway station ventilation and air conditioning system equipment and building energy consumption model include the temperature and humidity error model objective function and the energy consumption model objective function; The objective function F1(t) of the temperature and humidity error model is expressed as: F1(t)=λ1||T amb -T set ||+λ2||RH amb -RH set || Where, T amb and T set are the measured temperature and the set temperature, RH amb and RH set are the measured humidity and the set humidity, respectively, λ1 and λ2 are the weight parameters of temperature and humidity, respectively; The energy consumption model objective function F2(t) is expressed as: Where F2(t) represents the total energy consumption cost function, P i (t, f, ΔT) is the power of the i-th device in the ventilation and air-conditioning system, μ j is the weight parameter of the i-th device, t is time, f is device frequency, ΔT is the sampling period; n is the total number of devices in the ventilation and air-conditioning system; the devices in the ventilation and air-conditioning system include chillers, water pumps / fans; The calculation formula for the cooling machine input power is: Where P1 is the input power of the chiller, Q e is the cooling capacity, T eo is the cooling water outlet temperature, T ci is the cooling water return temperature, A, B, C are the coefficient parameters of the chiller input power; The calculation formula for pump / fan power is: <h2 style=";text-align:left;direction:ltr">P2 = (a0 + a1f + a2f<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +…)P<h2 style=";text-align:left;direction:ltr"> r Where P2 is the pump / fan power, P r is the effective power of the water pump / fan, f is the motor frequency, a0, a1, a2, … are the coefficient parameters of the water pump / fan power.

7. The urban rail electromechanical energy-saving prediction control method based on an adaptive model according to claim 1 is characterized in that: The constraints in the subway station ventilation and air conditioning system equipment and building energy consumption model are: In the formula, f max With f min are the upper and lower limits of the original motor operating frequency of the equipment, Q1, Q2, Q3 are the rated air volumes of the air-conditioning unit, exhaust fan and fresh air fan respectively, f is the air-conditioning fan frequency, g1(f) is the minimum frequency constraint, and g2(f) is the maximum frequency constraint.

8. The urban rail electromechanical energy-saving prediction and control method based on an adaptive model according to claim 1 is characterized in that: The MOPSO algorithm calculates the optimal control amount according to the current cost function value, including: Step 31, initializing the speed and position of all particles, and calculating the objective function value of all particles; the objective function is the cost function in the subway station ventilation and air conditioning system equipment and building energy consumption model; Step 32, non-inferior solutions form a Pareto optimal solution set; Step 33, sorting according to the dense distance to determine the optimal individual and population; Step 34, updating the inertia weight, particle position vector and velocity vector; Step 35, calculate the fitness function value of each particle after update, and update the Pareto optimal solution set; Step 36, updating the individual optimum and the population optimum; Step 37, determine whether the convergence condition is met, if so, output the result and the number of iterations, otherwise return to step 33.

9. An urban rail electromechanical energy-saving prediction control system based on an adaptive model, characterized in that: Includes adaptive MPC controller; The adaptive MPC controller includes: The Informer environmental data prediction module stores the external environmental model of the subway station built based on the Informer time series model. The optimal control quantity calculated last time and the environmental data obtained by actual measurement are used as the input of the Informer time series model. The Informer time series model reads the external environmental model of the subway station to predict the environmental data, and passes the predicted environmental data to the Trnsys-Fluent model. The Trnsys-Fluent joint dynamic simulation module stores the subway station ventilation and air-conditioning system equipment and building energy consumption model based on the subway station ventilation and air-conditioning system and constructed based on the Trnsys-Fluent model. The Trnsys-Fluent joint dynamic simulation module inputs the predicted environmental data and the optimal control quantity calculated last time into the Trnsys-Fluent model, starts the Trnsys kernel in the Trnsys-Fluent model, and the Trnsys kernel reads the subway station ventilation and air-conditioning system equipment and building energy consumption model and the model parameters calculated last time and starts the simulation; then the Trnsys kernel starts the Fluent kernel, and the Fluent kernel reads the fluid model in the subway station and the initialization boundary condition data to perform fluid simulation calculation, and returns the calculation results to the Trnsys kernel to complete the remaining calculations, obtain the updated boundary condition data and the hourly energy consumption of the metro station electromechanical system equipment, and output the cost function value; MOPSO solver, which stores the multi-objective particle swarm optimization algorithm MOPSO. The MOPSO algorithm calculates the optimal control amount according to the current cost function value; The adaptive MPC controller sends the optimal control quantity to each controlled device.

10. The urban rail electromechanical energy-saving prediction control system based on the adaptive model according to claim 9 is characterized in that: The subway station ventilation and air conditioning system equipment and building energy consumption model includes: Input module, used to input environmental data predicted by the Informer algorithm; The building load module is built based on Fluent and is used to simulate the flow field distribution in the station through numerical simulation methods, and output the temperature, humidity and wind speed distribution in the station; The wind system module, including the variable air volume fan module, the air duct module, and the surface cooler module, is used to simulate the air supply volume, temperature, humidity, and equipment energy consumption data of the subway station wind system; The water system module includes the chiller module, water pump module, cooling tower module, water pipe module, and confluence and diversion module, which is used to simulate the supply and return water temperature, supply and return water flow, cooling capacity, and equipment energy consumption data of the subway station water system; The control module is used to generate control quantities and send them to the control devices in the wind and water systems to achieve global control of the subway station wind and water systems.

Images