Modeling and control method and system for industrial plant air conditioner load optimization prediction

Abstract

The invention discloses a modeling and control method and system for industrial factory building air conditioner load optimization prediction. The method comprises the steps that indoor and outdoor parameter data of an air conditioner system in each area of an industrial factory building are collected; according to a thermal comfort index-predicted average value pwv model and the parameter data, determining the thermal comfort of each area; according to the thermal comfort degree state of each area of the industrial factory building, a preset event triggering control strategy and a controller, the control strategy of the current air conditioner and the state of the controller are determined; according to the current control strategy of the air conditioner and the state of the controller, an intermittent control strategy of the controller is determined, and the intermittent control strategy is divided into a control interval and a non-control interval; and inputting the data of the control interval and the non-control interval into a pre-constructed physical information neural network model to obtain the predicted energy consumption of each area of the industrial factory building. And the air-conditioning system load prediction level is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of commercial user power load power consumption detection and control strategy, and in particular to a modeling and control method and system for optimizing and predicting air conditioning load in industrial plants. Background Art

[0002] With the acceleration of industrialization, the demand for energy by industrial users is growing. At the same time, global warming and the improvement of environmental protection awareness have forced factories to reduce energy consumption and emissions. Therefore, how to achieve energy conservation and emission reduction while ensuring production efficiency has become an important issue facing modern factories. Advanced control strategies can greatly improve the energy efficiency of factory users. By accurately controlling the power consumption of equipment and optimizing energy distribution and scheduling, energy management strategies can help factories maximize energy utilization and thus reduce production costs.

[0003] Due to the continuous development of electronic information technology, automation control technology, Internet of Things technology and artificial intelligence, the performance and functions of energy-saving controllers have also been significantly improved. Modern energy-saving controllers usually use microprocessors as core control units, which have the characteristics of high precision, high reliability and easy expansion. At the same time, through linkage control with other intelligent devices, energy-saving controllers can achieve more complex energy management tasks, such as real-time monitoring of energy consumption and intelligent adjustment of equipment operating parameters. These technological advances provide strong support for the widespread application of energy-saving controllers.

[0004] There are many energy-saving control strategies for load prediction of central air-conditioning systems in industrial plants, such as continuous control strategies such as PID control, adaptive control, sliding film control, neural network control, and event-triggered control. These control strategies can significantly improve the load prediction level of the factory. It is worth noting that the central air-conditioning system mainly provides a comfortable environment for the staff in the factory, so the air-conditioning system does not have to work continuously. As an intermittent control action, the periodic intermittent control strategy does not need to perform continuous control actions, and can be a very flexible control strategy according to the actual needs of the users of the factory. However, the current intermittent control theory and technical methods are mainly based on mathematical model design controllers, but due to the large number of factory air-conditioning load types and the variable operation model replication, the system intermittent energy-saving control strategy based entirely on physical modeling is difficult to effectively predict the load characteristics of the control system. Therefore, it is urgent to propose a new intermittent control strategy for central air-conditioning systems, so as to obtain a satisfactory and comfortable environment for the staff in the industrial plant, and to achieve energy-saving effects, thereby reducing carbon emissions. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a modeling and control method for optimizing and predicting air conditioning load in industrial plants, the method comprising the following steps:

[0006] Collect indoor and outdoor parameter data of air conditioning systems in each area of ​​industrial plants;

[0007] Determine the thermal comfort level of each area according to the thermal comfort index-predicted mean value PWV model and the parameter data;

[0008] Determine the current air conditioning control strategy and controller status according to the thermal comfort status of each area of ​​the industrial plant and the pre-set event-triggered control strategy and controller;

[0009] Determine an intermittent control strategy of the controller according to the current control strategy of the air conditioner and the controller state, and divide the intermittent control strategy into a control interval and a non-control interval;

[0010] The data of the control interval and the non-control interval are input into a pre-built physical information neural network model to obtain the predicted energy consumption of each area of ​​the industrial plant.

[0011] Furthermore, the indoor and outdoor parameter data of the air conditioning system in each area of ​​the industrial plant are collected, including:

[0012] Divide the industrial areas according to the different exit locations of industrial plants;

[0013] Data is collected for each area according to the preset time.

[0014] Furthermore, the thermal comfort level of each area is determined according to the thermal comfort index-predicted mean value PWV model and the parameter data. The specific formula is:

[0015] η pmv =(0.303exp(-0.036M)+0.028)L

[0016] Among them, η pmv is thermal comfort, and M is metabolic rate.

[0017] Furthermore, according to the thermal comfort status of each area of ​​the industrial plant and the preset event-triggered control strategy and controller, the current air-conditioning control strategy and controller status are determined, including:

[0018] When the recommended value of thermal comfort is between -0.5 and 0.5, it is the best thermal comfort state, and between -1 and 1, it is the secondary thermal comfort state;

[0019] Depending on the thermal comfort status, the controller settings are as follows:

[0020]

[0021] Among them, when the state trajectory enters the E1(t) area, a new control strategy is adopted. When the state trajectory enters the E2(t) area, the control strategy of the previous moment is continued. When the state trajectory enters the E3(t) area, the control is stopped.

[0022]

[0023] Among them, E1(t) indicates that the indoor temperature is in the uncomfortable area, and the controller is in working state at this time; E2(t) indicates that the indoor temperature is in the appropriate comfortable area, and E3(t) indicates that the indoor temperature is in the optimal comfortable area, and the controller is in the stopped state at this time; when the indoor temperature trajectory enters the E3 area from the E2 area, the control is stopped;

[0024] The event trigger control strategy is:

[0025] When the state trajectory enters area E1 from area E2, the control is activated and the activation time sequence is defined by the event trigger strategy:

[0026] Furthermore, it also includes:

[0027] Calculate the energy consumption of each sampling point: pmv |≥1, trigger control:

[0028]

[0029] Among them, t r is the indoor temperature, Indicates the air supply quality, c p represents the specific heat capacity of air, t s Indicates the supply air temperature. When 0.5<|η pmv |<1, the controller remains unchanged and the energy consumption is the energy consumption of the previous moment. pmv |≤0.5, the control is stopped and the energy consumption is 0.

[0030] Furthermore, it also includes:

[0031] Construct the following multi-zone air conditioning system physical equations:

[0032]

[0033] Among them, t r is the indoor temperature, t m is the total thermal mass temperature of the building, C r and C m is the specific heat capacity of indoor air and lumped thermal mass, R ra is the thermal resistance between the room and the outside air, R rmis the thermal resistance between the room and the thermal mass, α is the solar amplitude absorption coefficient;

[0034] By calculating the physical equations of the multi-zone air-conditioning system, the parameters used to construct the physical information neural network model are obtained.

[0035] Furthermore, according to the current control strategy of the air conditioner and the controller state, the intermittent control strategy of the controller is determined, and the intermittent control strategy is divided into a control interval and a non-control interval, including:

[0036] When the controller stops controlling, the energy consumption is 0, and this part is divided into the non-control interval. When the controller is running, the energy consumption is not 0, and this part is divided into the control interval.

[0037] Furthermore, the data of the control interval and the non-control interval are input into a pre-built physical information neural network model to obtain the predicted energy consumption of each area of ​​the industrial plant, including:

[0038] The data of the control interval is brought into the pre-built physical information neural network model to obtain the predicted room temperature and energy consumption, and the predicted control interval data is combined with the non-control interval data in chronological order to obtain the final prediction result, and the mean absolute error MAE is calculated by the following formula to ensure the accuracy of the model;

[0039]

[0040] Where M is the number of sampling points, u i Represents the true value of energy consumption, Represents the predicted value of energy consumption.

[0041] Furthermore, it also includes:

[0042] The energy consumption predicted by the physical information neural network model is calculated according to the following formula to obtain the energy consumption of all areas in one day under the intermittent control strategy;

[0043]

[0044] Among them I e is the total energy consumption per day, is the energy consumption at time t, and m is the number of sampling times.

[0045] The present invention also provides a modeling and control system for optimizing and predicting air conditioning load in industrial plants, comprising:

[0046] Data acquisition module, used to collect indoor and outdoor parameter data of air conditioning systems in each area of ​​industrial plants;

[0047] A thermal comfort determination module, used to determine the thermal comfort of each area according to the thermal comfort index-prediction mean value PWV model and the parameter data;

[0048] A strategy and state determination module, used to determine the control strategy and controller state of the current air conditioner according to the thermal comfort state of each area of ​​the industrial plant and the pre-set event trigger control strategy and controller;

[0049] An interval division module, used to determine the intermittent control strategy of the controller according to the current control strategy of the air conditioner and the controller state, and divide the intermittent control strategy into a control interval and a non-control interval;

[0050] The prediction module is used to input the data of the control interval and the non-control interval into a pre-built physical information neural network model to obtain the predicted energy consumption of each area of ​​the industrial plant.

[0051] The present invention provides a modeling and control method and system for optimizing the prediction of air-conditioning load in industrial plants, and proposes a physical information neural network gap optimization control strategy based on data modeling for the prediction of central air-conditioning load in factories, thereby effectively improving the load prediction level of air-conditioning systems, greatly reducing energy costs while ensuring comfortable operation of factory personnel, and effectively reducing the amount of energy load data in factories. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flow chart of a modeling and control method for optimizing and predicting air conditioning load in industrial plants provided by an embodiment of the present invention;

[0053] Figure 2 is a diagram of steps involved in an embodiment of the present invention;

[0054] Figure 3 is a workflow diagram involved in an embodiment of the present invention;

[0055] Figure 4 is a control interval involved in the embodiment of the present invention;

[0056] Figure 5 The embodiment of the present invention relates to the intermittent control of plant area 1 based on the improved physical information neural network;

[0057] Figure 6 is an energy consumption comparison diagram of area 1 involved in an embodiment of the present invention;

[0058] Figure 7 It is a structural schematic diagram of a modeling and control system for optimizing and predicting air conditioning load in industrial plants provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0059] Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present invention, so the present invention is not limited to the specific implementation disclosed below.

[0060] Example 1

[0061] like Figure 1 As shown, an embodiment of the present invention provides a modeling and control method for optimizing and predicting air conditioning load in industrial plants, comprising the following steps:

[0062] Step S101, collecting indoor and outdoor parameter data of the air conditioning system in each area of ​​the industrial plant.

[0063] Specific steps are as follows Figure 2 As shown, parameter data is collected, including indoor temperature, outdoor temperature and other parameter information, and the data is preprocessed. Factories often contain multiple air outlets. The factory is divided into areas according to the different air outlet locations of industrial plants (one air outlet corresponds to one area), and then data is collected for each area according to the preset time (data can be collected every 5 minutes).

[0064] Step S102, determining the thermal comfort of each area according to the thermal comfort index-prediction mean value PWV model and the parameter data.

[0065] The data of each area is brought into the thermal comfort index-predicted mean value PMV model (an evaluation index that takes into account the thermal comfort feeling of the human body) to calculate the thermal comfort η pmv ,

[0066] The specific formula is:

[0067] η pmv =(0.303exp(-0.036M)+0.028)L (1)

[0068] Among them, η pmv is thermal comfort, M is metabolic rate (valued at 60 W·m -2 ).

[0069] The expression of L is as follows:

[0070]

[0071] In formula 2, M is the metabolic rate (valued at 1.2 met), W is the metabolic power (valued at 0), and t a is the air temperature (decision variable), p a is the water vapor pressure (mainly depends on the decision variable t a, as shown in formula 3), t cl is the clothing surface temperature (depending on various factors such as formula 4), h c is the convective heat transfer coefficient (as shown in Equation 5), is the mean radiation temperature (decision variable), f cl is the clothing area coefficient (the value is 1.1).

[0072]

[0073] Where Φ is the relative humidity (the value is 60%)

[0074]

[0075] where h c The expression is as shown in formula 5

[0076]

[0077] where v ar Relative air velocity (value is 0.18m·s -2 )

[0078] Step S103, determining the current air conditioning control strategy and controller status according to the thermal comfort status of each area of ​​the industrial plant and the preset event-triggered control strategy and controller.

[0079] According to the recommended value of PMV in ISO7730, when the recommended value of thermal comfort is between -0.5 and 0.5, it is the best thermal comfort state, and between -1 and 1, it is the secondary thermal comfort state;

[0080] Depending on the thermal comfort status, the controller settings are as follows:

[0081]

[0082] Among them, when the state trajectory enters the E1(t) area, a new control strategy is adopted (such as formula 7); when the state trajectory enters the E2(t) area, the control strategy of the previous moment is continued; when the state trajectory enters the E3(t) area, the control is stopped;

[0083]

[0084] Among them, E1(t) indicates that the indoor temperature is in the uncomfortable area, and the controller is in working state at this time; E2(t) indicates that the indoor temperature is in the appropriate comfortable area, and E3(t) indicates that the indoor temperature is in the optimal comfortable area, and the controller is in the stopped state at this time; when the indoor temperature trajectory enters the E3 area from the E2 area, the control is stopped;

[0085] The event trigger control strategy is:

[0086] When the state trajectory enters area E1 from area E2, the control is activated and the activation time sequence is defined by the event trigger strategy:

[0087] Calculate the energy consumption of each sampling point: pmv |≥1, trigger control:

[0088]

[0089] Among them, t r is the indoor temperature, Indicates the air supply quality, c p represents the specific heat capacity of air, t s Indicates the supply air temperature. When 0.5<|η pmv |<1, the controller remains unchanged and the energy consumption is the energy consumption of the previous moment. pmv |≤0.5, the control is stopped and the energy consumption is 0.

[0090] Construct the following multi-zone air conditioning system physical equations:

[0091]

[0092] Among them, t r is the indoor temperature, t m is the total thermal mass temperature of the building, C r and C m is the specific heat capacity of indoor air and lumped thermal mass, R ra is the thermal resistance between the room and the outside air, R rm is the thermal resistance between the room and the thermal mass, α is the solar amplitude absorption coefficient;

[0093] By calculating the physical equations of the multi-zone air-conditioning system, the parameters used to construct the physical information neural network model are obtained.

[0094] Step S104: determining an intermittent control strategy of the controller according to the current control strategy of the air conditioner and the controller state, and dividing the intermittent control strategy into a control interval and a non-control interval.

[0095] When the controller stops controlling, the energy consumption is 0, and this part is divided into the non-control interval. When the controller is running, the energy consumption is not 0, and this part is divided into the control interval.

[0096] Step S105, inputting the data of the control interval and the non-control interval into a pre-built physical information neural network model to obtain the predicted energy consumption of each area of ​​the industrial plant.

[0097] Design physical information neural network (specific network parameter settings are shown in Table 1)

[0098]

[0099] Table 1 Network setting parameters

[0100] The network inputs time, indoor temperature, outdoor temperature and energy consumption, and the network outputs the predicted indoor temperature and estimated predicted energy consumption. The loss function (Equation 10) is mainly composed of two parts: one is the physical information error term, which extracts the physical information in Equation 8 and substitutes the physical quantity predicted by the network into the corresponding physical law (Equation 9) to calculate the residual to constitute this part of the loss function, thereby ensuring physical consistency; the other part of the loss function is mainly the data error, which is the difference between the network prediction output and the actual observed data.

[0101]

[0102] Loss = L reg +L p h ys (10)

[0103] Where L reg Represents data error, t r,i and u phys r,i represents the indoor temperature and energy consumption at the i-th moment, and represents the predicted indoor temperature and predicted energy consumption at the i-th moment. phys represents the physical error term, T m,i represents the total thermal mass temperature of the building, represents the predicted total thermal mass temperature of the building at the i-th moment.

[0104] The data of the control interval is brought into the pre-built physical information neural network model to obtain the predicted room temperature and energy consumption, and the predicted control interval data is combined with the non-control interval data in chronological order to obtain the final prediction result, and the mean absolute error MAE is calculated by the following formula to ensure the accuracy of the model;

[0105]

[0106] Where M is the number of sampling points, u i Represents the true value of energy consumption, Represents the predicted value of energy consumption.

[0107] The energy consumption predicted by the physical information neural network model is calculated according to the following formula to obtain the energy consumption of all areas in one day under the intermittent control strategy;

[0108]

[0109] Among them I e is the total energy consumption per day, is the energy consumption at time t, and m is the number of sampling times.

[0110] Example 2

[0111] The following is a detailed description with reference to the accompanying drawings: Figure 3 This example mainly simulates the working environment of a factory in summer, simulates 5 days of data, divides the factory into 6 areas, and takes area 1 (30m×40m×9m) as an example. The main steps are as follows:

[0112] Step 1: Divide the factory building (80m×60m×9m) into 6 areas, and collect data from area 1 every 5 minutes.

[0113] Step 2: Bring the data in area 1 into the PMV calculation model to calculate the thermal comfort, and divide the data into control interval and non-control interval according to the intermittent control strategy mentioned in the present invention, such as Figure 4 As shown;

[0114] (y=0 indicates the non-control stage, y=1 indicates the control stage) The control duration is about 12 hours, while the control duration of continuous control is 24 hours. The energy consumption of each sampling point is calculated according to formula (7).

[0115] Step 4: Extract physical information according to the above model (Formula 8), the specific values ​​are as follows:

[0116]

[0117] Step 5: Preprocess the 1440 sampling data of the first four days of measurement, divide the data into control interval and non-control interval according to the control strategy of the intermittent controller, and bring the data of the control interval (time, indoor temperature, outdoor temperature, energy consumption) into the neural network for training. Then bring the data of the fifth day into the trained model to predict the indoor temperature and energy consumption of each area.

[0118] Step 6: Data integration. Combine the predicted data of the control interval of the fifth day (test set) in region 1 with the original data of the non-control interval in chronological order to obtain the final prediction results. Calculate the MAE (mean absolute error) to ensure that the model is more accurate. The results are as follows: Figure 5 As shown:

[0119] Results comparison: Figure 6The comparison effect diagram of the continuous control and the control strategy of the present invention in area 1 is shown in FIG. According to formula (11), the energy consumption of the continuous control and the intermittent control can be calculated:

[0120]

[0121] The sampling interval is once every 5 minutes, so a total of 288 samples are required; Formula (13) and Formula (14) respectively obtain the energy consumption results of continuous control and intermittent control for one day, and Formula (15) shows that intermittent control saves about 27.03% energy than continuous control.

[0122] Example 3

[0123] Based on the same inventive concept, the present invention provides a modeling and control system for optimizing and predicting air conditioning load in industrial plants, such as Figure 7 As shown, including:

[0124] The data collection module 710 is used to collect indoor and outdoor parameter data of the air conditioning system in each area of ​​the industrial plant;

[0125] A thermal comfort determination module 720, for determining the thermal comfort of each area according to the thermal comfort index-prediction mean value PWV model and the parameter data;

[0126] A strategy and state determination module 730, for determining the control strategy and controller state of the current air conditioner according to the thermal comfort state of each area of ​​the industrial plant and the preset event-triggered control strategy and controller;

[0127] An interval division module 740 is used to determine the intermittent control strategy of the controller according to the current control strategy of the air conditioner and the controller state, and divide the intermittent control strategy into a control interval and a non-control interval;

[0128] The prediction module 750 is used to input the data of the control interval and the non-control interval into a pre-built physical information neural network model to obtain the predicted energy consumption of each area of ​​the industrial plant.

[0129] The present invention provides a modeling and control method and system for optimizing the prediction of air-conditioning load in industrial plants, and proposes a physical information neural network gap optimization control strategy based on data modeling for the prediction of central air-conditioning load in factories, thereby effectively improving the load prediction level of air-conditioning systems, greatly reducing energy costs while ensuring comfortable operation of factory personnel, and effectively reducing the amount of energy load data in factories.

[0130] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks 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 produce 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.

[0132] 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 comprising 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.

[0133] 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. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalents that do not depart from the spirit and scope of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A modeling and control method for optimizing and predicting air conditioning load in industrial plants, characterized in that: include: Collect indoor and outdoor parameter data of air conditioning systems in each area of ​​industrial plants; Determine the thermal comfort level of each area according to the thermal comfort index-predicted mean value PWV model and the parameter data; Determine the current air conditioning control strategy and controller status according to the thermal comfort status of each area of ​​the industrial plant and the pre-set event-triggered control strategy and controller; Determine an intermittent control strategy of the controller according to the current control strategy of the air conditioner and the controller state, and divide the intermittent control strategy into a control interval and a non-control interval; The data of the control interval and the non-control interval are input into a pre-built physical information neural network model to obtain the predicted energy consumption of each area of ​​the industrial plant.

2. The method according to claim 1, characterized in that Collect indoor and outdoor parameter data of air conditioning systems in each area of ​​industrial plants, including: Divide the industrial areas according to the different exit locations of industrial plants; Data is collected for each area according to the preset time.

3. The method according to claim 1, characterized in that According to the thermal comfort index-predicted mean value PWV model and the parameter data, the thermal comfort of each area is determined. The specific formula is: or pmv =(0.303exp(-0.036M)+0.028)L Among them, η pmv is thermal comfort, and M is metabolic rate.

4. The method according to claim 1, characterized in that According to the thermal comfort status of each area of ​​the industrial plant and the pre-set event-triggered control strategy and controller, the current air conditioner control strategy and controller status are determined, including: When the recommended value of thermal comfort is between -0.5 and 0.5, it is the best thermal comfort state, and between -1 and 1, it is the secondary thermal comfort state; Depending on the thermal comfort status, the controller settings are as follows: Among them, when the state trajectory enters the E1(t) area, a new control strategy is adopted. When the state trajectory enters the E2(t) area, the control strategy of the previous moment is continued. When the state trajectory enters the E3(t) area, the control is stopped. Among them, E1(t) indicates that the indoor temperature is in the uncomfortable area, and the controller is in working state at this time; E2(t) indicates that the indoor temperature is in the appropriate comfortable area, and E3(t) indicates that the indoor temperature is in the optimal comfortable area, and the controller is in the stopped state at this time; when the indoor temperature trajectory enters the E3 area from the E2 area, the control is stopped; The event trigger control strategy is: When the state trajectory enters area E1 from area E2, the control is activated and the activation time sequence is defined by the event trigger strategy:

5. The method according to claim 4, characterized in that Also includes: Calculate the energy consumption of each sampling point: pmv |≥1, trigger control: Among them, t r is the indoor temperature, Indicates the air supply quality, c p represents the specific heat capacity of air, t s Indicates the supply air temperature. When 0.5<|η pmv |<1, the controller remains unchanged and the energy consumption is the energy consumption of the previous moment. pmv |≤0.5, the control is stopped and the energy consumption is 0.

6. The method according to claim 4, characterized in that Also includes: Construct the following multi-zone air conditioning system physical equations: Among them, t r is the indoor temperature, t m is the total thermal mass temperature of the building, C r and C m is the specific heat capacity of indoor air and lumped thermal mass, R ra is the thermal resistance between the room and the outside air, R rm is the thermal resistance between the room and the thermal mass, α is the solar amplitude absorption coefficient; By calculating the physical equations of the multi-zone air-conditioning system, the parameters used to construct the physical information neural network model are obtained.

7. The method according to claim 1, characterized in that According to the current control strategy and controller state of the air conditioner, an intermittent control strategy of the controller is determined, and the intermittent control strategy is divided into a control interval and a non-control interval, including: When the controller stops controlling, the energy consumption is 0, and this part is divided into the non-control interval. When the controller is running, the energy consumption is not 0, and this part is divided into the control interval.

8. The method according to claim 1, characterized in that The data of the control interval and the non-control interval are input into a pre-built physical information neural network model to obtain the predicted energy consumption of each area of ​​the industrial plant, including: The data of the control interval is brought into the pre-built physical information neural network model to obtain the predicted room temperature and energy consumption, and the predicted control interval data is combined with the non-control interval data in chronological order to obtain the final prediction result, and the mean absolute error MAE is calculated by the following formula to ensure the accuracy of the model; Where M is the number of sampling points, u i Represents the true value of energy consumption, Represents the predicted value of energy consumption.

9. The method according to claim 1 or 8, characterized in that: Also includes: The energy consumption predicted by the physical information neural network model is calculated according to the following formula to obtain the energy consumption of all areas in one day under the intermittent control strategy; Among them I e is the total energy consumption per day, is the energy consumption at time t, and m is the number of sampling times.

10. A modeling and control system for optimizing and predicting air conditioning load in industrial plants, characterized in that: include: Data acquisition module, used to collect indoor and outdoor parameter data of air conditioning systems in each area of ​​industrial plants; A thermal comfort determination module, used to determine the thermal comfort of each area according to the thermal comfort index-prediction mean value PWV model and the parameter data; A strategy and state determination module, used to determine the control strategy and controller state of the current air conditioner according to the thermal comfort state of each area of ​​the industrial plant and the pre-set event triggering control strategy and controller; An interval division module, used to determine the intermittent control strategy of the controller according to the current control strategy of the air conditioner and the controller state, and divide the intermittent control strategy into a control interval and a non-control interval; The prediction module is used to input the data of the control interval and the non-control interval into a pre-built physical information neural network model to obtain the predicted energy consumption of each area of ​​the industrial plant.

Images