Heat supply whole network collaborative dynamic optimization regulation and control method, system and equipment and storage medium

Abstract

The invention discloses a heat supply whole network collaborative dynamic optimization regulation and control method, system and equipment and a storage medium, relates to the technical field of heat supply control, and aims at solving the technical problems that in the prior art, thermal load prediction data is inaccurate, and heat transfer distribution cannot be optimized by a heat supply system, so that operation energy consumption is large. According to the technical scheme, the method mainly comprises the steps that building load prediction data and heating station load prediction data are acquired, and a load prediction model is constructed; training the load prediction model, and optimizing the model and pipe network data by using an ARX active regression algorithm; a heat supply whole network physical simulation model is constructed, and heat supply whole network thermal load data is obtained through calculation by using an MILP mixed integer linear programming algorithm; an optimal control strategy is obtained by optimizing a heat supply whole network physical simulation model and introducing a real-time feedback mechanism and an overall cooperative control module, optimization parameters of the heat supply whole network physical simulation model are automatically adjusted by using the optimal control strategy, and the heat supply whole network is continuously in an optimal operation state through automatic repeated iteration.

Description

Technical Field

[0001] The invention relates to the technical field of heating control, and in particular to a method for optimizing the control of energy consumption operation of a heating system. Background Art

[0002] Building energy consumption is one of the three major energy consumptions in my country, and the energy consumption of the central heating system accounts for the largest proportion of building energy consumption. The central heating system consists of a heat source, a heat exchange station, a user side, and a heat network connecting each link. In the entire heating system, indoor temperature, as the main indicator affecting indoor comfort, has always been valued and studied by scholars at home and abroad. How to control system parameters such as water supply temperature according to demand so that the indoor temperature is always maintained in a comfortable area is a problem that central heating needs to face. Although the current heating system has met the heating needs to a certain extent, due to the complexity of the heating system itself, the system still has shortcomings and deficiencies that have a negative impact on the overall efficiency and user comfort. The complexity is mainly reflected in four aspects: complex system structure, complex heat transfer process, complex environmental factors, and complex control algorithm.

[0003] The heating system consists of multiple levels, among which heat collection, heat supply, heat exchange and control center are the main components. The function of the heat collection part is mainly to increase the water temperature and reach the set value. During the temperature increase process, the water temperature is monitored in real time through the temperature sensor. When the water temperature reaches the set value, the hot water is sent to the heat exchange part for heat exchange. The function of the heat exchange part is mainly to exchange the hot water in the heat collection part with the return water on the secondary side to increase the water temperature on the secondary side. The water temperature control on the secondary side is achieved by changing the frequency of the inverter and then the heat exchange efficiency. The function of the heating part is mainly to transport the hot water on the secondary side after heat exchange to the user through a pipeline to achieve heating. The control center performs overall information processing and controls the temperature of the heating system in real time.

[0004] The current control model of the heating system often adopts a static model, which cannot reflect the hysteresis, dynamic changes and radiation of the heating system. Due to its open-loop control characteristics, the shortcomings of the heating system itself, such as long cycle and lack of immediate feedback, are exposed. In particular, the heat transfer process has significant nonlinear characteristics. Factors such as changes in heat load, changes in fluid flow resistance, and heat loss make it difficult to accurately predict the thermal behavior of the system. It takes a certain amount of time for heat to be transferred from the heat source to the user side, that is, it takes a certain amount of time for the system to adjust the water supply temperature to be reflected on the user side. This temperature transfer delay phenomenon causes the user to be unable to feel the effect of temperature adjustment in a short time. In addition, the water and pipes in the pipe network have a certain thermal capacity. Extra time is required to heat or cool the pipes and water during the heat transfer process. After the heating system adjusts the water supply temperature, the distribution of heat in the pipe network needs to gradually reach a new equilibrium state. These phenomena will further increase the lag time of heat transfer and affect the system's rapid response capability. The structural complexity and thermal transfer complexity of the heating system itself are difficult to change. Changing the pipe network structure is extremely costly and involves large-scale infrastructure transformation, including pipeline laying, equipment replacement, and system redesign, which is both time-consuming and requires a lot of money. The complexity of heat transfer stems from the complex physical laws involved, such as heat conduction, heat capacity, convection heat transfer, and radiation heat dissipation. These basic physical properties are also unchangeable.

[0005] The current heating system has many defects and deficiencies. Most systems set the water supply temperature for the next day based on the outdoor temperature and weather forecast through steady-state adjustment curves and manual verification. When the outdoor meteorological conditions change rapidly, the system cannot adjust the water supply temperature in time, resulting in large fluctuations in indoor temperature, which seriously affects the living comfort and causes energy waste. There are many control variables in the heating system, such as water supply temperature, return water temperature, flow rate, pressure, etc. There are complex coupling relationships between these variables, and comprehensive consideration is required to achieve precise control. Due to the dynamic changes in user heat demand and external environmental conditions, the control system needs to adjust parameters in real time, which also puts high demands on the real-time performance and response speed of the control algorithm. This makes it difficult for even traditional closed-loop PID control to cope with the actual complex situation of the heating system and meet the control requirements. In addition to meteorological environmental factors, the heat demand of different users and buildings varies greatly. The geographical environment, terrain conditions, building structure, etc. of the heating network will affect the transfer and distribution of heat. The heating system needs to be adjusted according to actual needs, which increases the difficulty of operation and regulation control. Summary of the invention

[0006] The present invention solves the technical problems in the prior art that the thermal load prediction data is inaccurate and the heating system cannot optimize the heat transfer distribution, resulting in high operating energy consumption.

[0007] The present invention proposes a method for coordinated dynamic optimization and control of the whole heating network, including a heating system, a plurality of heating stations, a pipe network, and heating equipment. The method comprises the following steps:

[0008] S1. Real-time data collection, the data includes the operation data of the entire heating network and the meteorological data of the area where the entire heating network is located; the operation data includes the pipe network parameters, the real-time operation status of the heating station and the operation parameters of the heating equipment;

[0009] S2. Inputting the collected data into the control system of the heating system, and preprocessing the collected data to obtain building load forecast data and thermal power station load forecast data;

[0010] S3. constructing a load forecasting model using the building load forecasting data and the thermal power station load forecasting data, and training the load forecasting model based on the historical load data to obtain a trained load forecasting model;

[0011] S4. Optimize the trained load prediction model and the pipe network parameters, and use the ARX self-regression algorithm to calculate the pipe network parameter prediction data; the pipe network parameter prediction data includes the pipe network pressure and the secondary side temperature of the pipe network;

[0012] S5. Construct a physical simulation model of the entire heating network; solve the physical simulation model of the entire heating network using the MILP mixed integer linear programming algorithm based on the network parameter prediction data to obtain thermal load data of the entire heating network;

[0013] S6. Optimize the physical simulation model of the entire heating network; obtain the optimal control strategy of the heating system by introducing a real-time feedback mechanism and an overall collaborative control module, use the optimal control strategy to automatically adjust the optimization parameters of the physical simulation model of the entire heating network, and automatically repeat the iteration to ensure that the entire heating network is continuously in the optimal operating state.

[0014] Further optimization, the process of preprocessing the collected data in S2 includes:

[0015] Inputting the meteorological data and the historical load data into a machine learning algorithm to obtain the building load forecast data;

[0016] The building load prediction data and the real-time operating status of the thermal power station are input into a machine learning algorithm to calculate the thermal power station load prediction data.

[0017] Further optimized, the building load prediction data includes building thermal load data, and the building thermal load data is based on the meteorological conditions of the area where the building is located; the meteorological data includes the temperature, humidity and wind speed of the external environment of the area where the heating network is located.

[0018] Further optimization, in S4, the trained load prediction model and the pipe network parameters are optimized including online parameter identification and hysteresis effect modeling and dynamic response mechanism; the pipe network parameters include pipe network pressure and pipe network temperature;

[0019] The online parameter identification is to adjust the load prediction model parameters and continuously iterate through real-time monitoring of the real-time operating status and pipeline network parameters of the thermal power station in the operating data collected until the dynamic state and accurate value of the heating system are within a preset range; the hysteresis effect modeling and dynamic response mechanism is to add delay factors and dynamic characteristics to the load prediction model and pipeline network parameters.

[0020] Further optimization, the network parameter prediction data obtained by using the ARX self-regression algorithm in S4 is:

[0021]

[0022] in, represents the load forecast result at time t for time t+1; q(t) represents the load at time t; T a (t) represents the outdoor temperature at time t; I(t) represents a scalar indicating whether time t is a holiday, which is 1 if it is, and 0 if it is not; a~d represent parameters; a1 represents the parameter of the load item at time t; a2 represents the parameter of the load item at time t-1; a 24 Represents the parameter of the load item at time t-23; a 25 Represents the parameter of the load item at time t-24; a 26 represents the parameter of the load item at time t-26; b1 represents the parameter of the outdoor temperature item at time t; b2 represents the parameter of the outdoor temperature item at time t-1; b 24 b represents the parameter of the outdoor temperature item at time t-23; 25 represents the parameter of the outdoor temperature item at time t-24; c 1,i represents the parameter of the sine correction term; c 2,i Represents the parameter of the cosine correction term; d 2,i The parameter representing the holiday sinusoidal correction term; d 2,i Parameters representing the holiday cosine correction term.

[0023] Further optimization, the method for constructing the physical simulation model of the entire heating network in S5 is: constructing a system operation simulation environment by simulating the overall physical process of the heating system, and evaluating the operating characteristics of the system in the operation simulation environment under the optimal control strategy; the operating characteristics include the thermodynamic characteristics of the heating station, the thermodynamic characteristics of the pipeline network, the equipment operation characteristics and the system dynamic response.

[0024] Further optimization, the step of solving the physical simulation model of the heating network based on the pipe network parameter prediction data using the MILP mixed integer linear programming algorithm in S5 includes: inputting the pipe network parameter prediction data into the physical simulation model of the heating network; setting the optimization parameters including the switching state of the heating equipment, the heating temperature, the flow rate and the pressure difference; setting the optimization objective function to the lowest operating energy consumption; the MILP mixed integer linear programming algorithm is:

[0025] minc T x

[0026] Ax≤b

[0027]

[0028] Among them: x is the decision variable vector, including the device switch state integer variable and the heating temperature and flow rate continuous variables; c is the objective function coefficient vector; A is the constraint matrix; b is the constraint vector; represents a set of integer variables, represents a set of continuous variables.

[0029] The present invention proposes a coordinated dynamic optimization control system for the entire heating network, wherein the entire heating network includes a heating system, multiple heating stations, a pipe network, and multiple heating equipment, and is characterized in that the control system includes the following modules:

[0030] Data acquisition module: used to collect data in real time, the data includes the operation data of the whole heating network and the meteorological data of the area where the whole heating network is located; the operation data includes the pipe network parameters, the real-time operation status of the heating station and the operation parameters of the heating equipment;

[0031] Data preprocessing module: used to input the collected data into the control system of the heating system, and also used to preprocess the collected data to obtain building load forecast data and heating station load forecast data;

[0032] A load forecasting model building module: used to build a load forecasting model using the building load forecasting data and the thermal power station load forecasting data, and also used to train the load forecasting model based on the historical load data to obtain a trained load forecasting model;

[0033] Load forecasting model optimization module: used to optimize the trained load forecasting model and the pipe network parameters, and also used to calculate the pipe network parameter prediction data using the ARX organic self-regression algorithm; the pipe network parameter prediction data includes the pipe network pressure and the pipe network secondary side temperature;

[0034] Heating network physical simulation model construction module: used to construct the heating network physical simulation model; based on the pipe network parameter prediction data, the MILP mixed integer linear programming algorithm is used to solve the heating network physical simulation model to obtain the heating network thermal load data;

[0035] An optimal solution module for energy consumption operation; used to optimize the physical simulation model of the entire heating network; obtain the optimal control strategy of the heating system by introducing a real-time feedback mechanism and an overall collaborative control module, and automatically adjust the optimization parameters of the physical simulation model of the entire heating network using the optimal control strategy, and automatically repeat iterations to keep the entire heating network in the optimal operating state.

[0036] The present invention also proposes a control device for a method for collaborative dynamic optimization and control of the entire heating network, wherein the device includes a processor and a memory, wherein the memory is used to store computer programs, and when the microprocessor reads the computer program on the memory, it executes any one of the methods for collaborative dynamic optimization and control of the entire heating network proposed by the present invention.

[0037] The present invention also proposes a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, any one of the methods for collaborative dynamic optimization and control of the entire heating network proposed by the present invention is implemented.

[0038] Beneficial technical effects of the present invention:

[0039] The present invention aims at the technical problem of inaccurate thermal load prediction data in the prior art. By constructing a thermal load prediction model, the building load prediction is realized, and the thermal load demand of the building with reference to meteorological conditions is predicted; the thermal power station load prediction is realized, and the load situation of each thermal power station at a future moment is predicted; accurate thermal power station load prediction data and building load prediction data are formed, thereby ensuring the accurate prediction of thermal load prediction data; and the ARX self-regression algorithm is used to predict the pipe network parameters, thereby realizing the precise control of thermal load prediction data.

[0040] The present invention aims to solve the technical problem in the prior art that the heating system cannot optimize the heat transfer distribution, resulting in high operating energy consumption. By simulating the physical process of the entire heating system, including the thermodynamic characteristics of the heating station and the pipeline network, the equipment operating characteristics, and the dynamic response of the system, a physical simulation model of the entire heating network is constructed. The optimal control strategy is tested in the virtual environment of the model, the optimization parameters of the model are automatically adjusted, and automatic repeated iterations are performed to ensure that the heating system is continuously in the optimal operating state: the energy consumption is minimized and the operating cost is minimized, thereby achieving coordinated dynamic optimization and regulation of the entire heating network and meeting the heating needs of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a load forecasting model architecture diagram described in Implementation Method 2;

[0042] Figure 2 This is a curve diagram of the load forecast data calculated by the ARX organic self-regression algorithm as described in Implementation Mode 5; the ordinate load in the figure is the load, the curve k-step ahead measured load k=1 is the load predicted one step ahead at the current moment, and the curve k-step ahead predicted load k=1 is the load predicted one step ahead at the current moment;

[0043] Figure 3 This is the architecture diagram of the physical simulation model of the entire heating network described in Implementation Method 6. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] Embodiment 1: A method for coordinated dynamic optimization and control of the entire heating network described in this embodiment includes:

[0046] S1. Real-time data collection, the data includes the operation data of the entire heating network and the meteorological data of the area where the entire heating network is located; the operation data includes the pipe network parameters, the real-time operation status of the heating station and the operation parameters of the heating equipment;

[0047] S2. Inputting the collected data into the control system of the heating system, and preprocessing the collected data to obtain building load forecast data and thermal power station load forecast data;

[0048] S3. constructing a load forecasting model using the building load forecasting data and the thermal power station load forecasting data, and training the load forecasting model based on the historical load data to obtain a trained load forecasting model;

[0049] S4. Optimize the trained load prediction model and the pipe network parameters, and use the ARX self-regression algorithm to calculate the pipe network parameter prediction data; the pipe network parameter prediction data includes the pipe network pressure and the secondary side temperature of the pipe network;

[0050] S5. Construct a physical simulation model of the entire heating network; solve the physical simulation model of the entire heating network using the MILP mixed integer linear programming algorithm based on the network parameter prediction data to obtain thermal load data of the entire heating network;

[0051] S6. Optimize the physical simulation model of the entire heating network; obtain the optimal control strategy of the heating system by introducing a real-time feedback mechanism and an overall collaborative control module, use the optimal control strategy to automatically adjust the optimization parameters of the physical simulation model of the entire heating network, and automatically repeat the iteration to ensure that the entire heating network is continuously in the optimal operating state.

[0052] In this embodiment, by collecting data, processing data, building load prediction models and physical simulation models of the entire heating network, the accuracy of the thermal power station load and building load prediction data is improved, and the heat transfer and distribution of the heating system are optimized; by combining the MILP mixed integer linear programming algorithm, the heating efficiency is maximized and the resource utilization is optimized, thereby reducing the energy consumption of the system operation; through thermal load prediction and optimization, the heating temperature and pressure difference are accurately controlled, thereby meeting the heating needs of users, reducing energy consumption, and improving economic benefits. The optimal operating state refers to the automatic adjustment of the heating system control strategy at a certain moment to minimize the system's operating energy consumption, thereby obtaining the control optimization parameters when the entire heating network is coordinated and dynamically optimized to achieve the optimal energy consumption operation.

[0053] Implementation method two, this implementation method is a further example of S2 in a method for coordinated dynamic optimization and control of the entire heating network described in implementation method one, and the process of preprocessing the collected data in S2 includes: inputting the meteorological data and the historical load data into a machine learning algorithm to obtain the building load prediction data; inputting the building load prediction data and the real-time operating status of the thermal power station into a machine learning algorithm to calculate the thermal power station load prediction data.

[0054] This implementation further defines the process of preprocessing the collected data. During the preprocessing process, the building load forecast analyzes meteorological data and historical load data, and uses the machine learning algorithm long short-term memory network model (LSTM) to predict the building's heat load demand with reference to meteorological conditions. The building's parameters under each meteorological condition include the temperature, humidity and wind speed of the external environment of the building area. This ensures accurate prediction of future heating demand, thereby providing basic data for subsequent optimization. The thermal power station load forecast data includes the load data of all thermal power stations for the next 24 hours. The thermal power station load forecast is based on the results of the building load forecast, combined with the real-time operating status of the thermal power station in the operating data, to predict the load conditions of each thermal power station at a future moment. This is based on a machine learning algorithm, and through the training of a large amount of historical data, an accurate load forecast model is established, such as Figure 1 shown.

[0055] Implementation method three, this implementation method is a further example of S2 in a method for coordinated dynamic optimization and control of the entire heating network described in implementation method two, the building load prediction data includes building thermal load data, and the building thermal load data is based on the meteorological conditions of the area where the building is located; the meteorological data includes the temperature, humidity and wind speed of the external environment of the area where the entire heating network is located.

[0056] This implementation further limits the building load prediction data and meteorological data, and realizes the prediction of the building's heat load demand with reference to meteorological conditions, thereby ensuring accurate prediction of future heating demand.

[0057] Embodiment 4, this embodiment is a further example of S4 in a method for coordinated dynamic optimization and control of the whole heating network described in embodiment 1, wherein the optimization of the trained load prediction model and the pipe network parameters in S4 includes online parameter identification and hysteresis effect modeling and dynamic response mechanism; the pipe network parameter prediction data includes pipe network pressure and pipe network secondary side temperature;

[0058] The online parameter identification is to adjust the load prediction model parameters and continuously iterate through real-time monitoring of the real-time operating status and pipeline network parameters of the thermal power station in the operating data collected until the dynamic state and accurate value of the heating system are within a preset range; the hysteresis effect modeling and dynamic response mechanism is to add delay factors and dynamic characteristics to the load prediction model and pipeline network parameters.

[0059] This embodiment further defines a method for optimizing the trained load prediction model and the pipeline network parameters, ensures the accuracy of the model through online parameter identification, and enables the heating system to continuously adjust the model parameters to keep them consistent with the actual situation by real-time monitoring of the operating data of the thermal power station and the pipeline network.

[0060] The process of online parameter identification is a process of continuous iteration and optimization, which ensures the dynamic adaptability of the model and the accuracy of prediction. By optimizing the operation mode of the heating system, reducing energy waste and lowering operating costs, the overall efficiency and stability of the system can be improved. Through online parameter identification, the real-time operation status data of the thermal power station and the pressure and temperature data of the pipeline network in the collected operation data can be monitored in real time. By adjusting the parameters of the thermal prediction model and continuously iterating, the dynamic adaptability of the model and the accuracy of prediction are ensured.

[0061] In order to solve the problem that the heating system cannot reflect the hysteresis and dynamic elements, the system introduces hysteresis effect modeling and dynamic response mechanism, adds delay factors and dynamic characteristics to the existing prediction model, and enhances the structure of ARX from regression algorithm to more accurately reflect the actual situation of the heating system. By reducing the influence of the complexity of environmental factors, real-time monitoring and data analysis, environmental changes can be predicted in advance, and system operating parameters can be adjusted in time to adapt to different climatic conditions and user needs. By introducing hysteresis effect modeling and dynamic response mechanism into the heating system, delay factors and dynamic characteristics can be added to the existing prediction model, and the structure of ARX from regression algorithm can be enhanced to more accurately reflect the actual situation of the heating system, solving the problem that the heating system cannot reflect the hysteresis and dynamic elements.

[0062] Implementation mode 5, this implementation mode is a further example of S4 in a method for coordinated dynamic optimization and control of the whole heating network described in implementation mode 1 or 4. The pipe network parameter prediction data obtained by using the ARX self-regression algorithm in S4 is:

[0063]

[0064] in, represents the load forecast result at time t for time t+1; q(t) represents the load at time t; T a (t) represents the outdoor temperature at time t; I(t) represents a scalar indicating whether time t is a holiday, which is 1 if it is, and 0 if it is not; a~d represent parameters; a1 represents the parameter of the load item at time t; a2 represents the parameter of the load item at time t-1; a 24 Represents the parameter of the load item at time t-23; a 25 Represents the parameter of the load item at time t-24; a 26 represents the parameter of the load item at time t-26; b1 represents the parameter of the outdoor temperature item at time t; b2 represents the parameter of the outdoor temperature item at time t-1; b 24 b represents the parameter of the outdoor temperature item at time t-23; 25 represents the parameter of the outdoor temperature item at time t-24; c 1,i represents the parameter of the sine correction term; c 2,i Represents the parameter of the cosine correction term; d 2,i The parameter representing the holiday sinusoidal correction term; d 2,i Parameters representing the holiday cosine correction term.

[0065] In this embodiment, the ARX autoregressive algorithm is used to predict the secondary side temperature of the pipeline network. By using the ARX (AutoRegressive with eXogenous inputs) autoregressive algorithm to predict the pipeline network parameters, accurate temperature prediction is the prerequisite for achieving precise control. The optimization control strategy improves the dynamic response capability of the system, reduces temperature fluctuations, and improves user comfort through advanced control algorithms, such as model predictive control (MPC) and intelligent control technology. After obtaining accurate building load and thermal power station load forecasts, the heating system uses the ARX autoregressive algorithm to predict relevant parameters on the secondary side of the pipeline network, such as temperature and flow. By considering the current temperature, historical temperature, and external inputs (such as heating temperature and flow), a regression model is established to predict the pipeline network parameters. This step is crucial for subsequent optimization control, because accurate temperature prediction is the prerequisite for achieving precise control.

[0066] Load forecasting is generally based on statistical models, and commonly used methods include regression analysis, time series analysis, support vector machines, random forests, neural networks, LSTM neural networks, and the like. Unlike load simulation based on physical models, these methods all use historical data for analysis and modeling, and predict the air conditioning and heating load in the future based on the predicted meteorological parameters, current and historical loads, and other key characteristic values ​​that affect load changes. In the optimization control of the heating system, it is necessary to predict the load of the thermal power station and parameters such as the temperature and flow of the pipeline network in order to subsequently calculate the overall optimality of the system. The present invention adopts a classic time series analysis model-the ARX derived regression algorithm considering time-varying parameters. The model can well reflect the periodic daily changes, time delays, and climate impact characteristics of the load. Compared with other machine learning algorithms, this method shows better adaptability and requires less data.

[0067] This implementation method selects historical load, outdoor temperature, holidays and other data as characteristic values. It can also add characteristic values ​​such as wind speed and radiation intensity according to the actual load characteristics and influencing factors. The model uses two third-order Fourier expansions to represent the periodic variation characteristics of weekdays and holidays. The above model is for a time-invariant system, while the actual load model may change slowly over time, such as the impact of changes in characteristics that are not considered. In order to simulate the time-varying system, the system parameters a~d can be dynamically estimated and corrected without changing the system model. This paper adopts the memory-fading recursive least squares algorithm (RLS) for dynamic parameter identification to quickly adapt to changes in model parameters.

[0068] Figure 2is a prediction result curve, and the forgetting coefficient alpha is 0.97. According to the figure, after the initial 100 hours of parameter identification, the prediction error gradually decreases and tends to be stable. Compared with the traditional machine learning algorithm, this method shows better adaptability and requires less data. The final CVRMSE root mean square error coefficient of variation is 8.4%. Compared with other traditional algorithms, the prediction accuracy can be improved by 7-10%. In the heating system, based on the current and historical temperature data and external input data (such as heating temperature and flow), the application of ARX self-regression algorithm can predict the temperature and other parameters of the secondary side of the pipe network, which is crucial for the optimization control of the heating system.

[0069] Implementation method six. This implementation method is a further example of S5 in a method for coordinated dynamic optimization and control of the entire heating network described in implementation method one. The method for constructing a physical simulation model of the entire heating network in S5 is: constructing a system operation simulation environment by simulating the overall physical process of the heating system, and evaluating the operating characteristics of the system in the operation simulation environment under the optimal control strategy; the operating characteristics include thermodynamic characteristics of the thermal power station, thermodynamic characteristics of the pipeline network, equipment operation characteristics and system dynamic response.

[0070] This embodiment further defines a method for constructing a physical simulation model of the entire heating network. The physical simulation model of the entire heating network is an important component of the system. Figure 3 As shown in the figure, the model simulates the physical process of the entire heating system to evaluate the operating effect of the system under different control strategies. The simulation model not only takes into account the thermodynamic characteristics of the heating station and the pipe network, but also includes the operating characteristics of the equipment and the dynamic response of the system. Through the simulation model, the system can test different control strategies in a virtual environment to find the optimal solution.

[0071] Implementation method seven, this implementation method is a further example of S5 in a method for collaborative dynamic optimization and control of a heating network described in implementation method one, and the steps of solving the physical simulation model of the heating network based on the pipe network parameter prediction data using the MILP mixed integer linear programming algorithm in S5 are as follows: input the pipe network parameter prediction data into the physical simulation model of the heating network; set the optimization parameters including the switching state of the heating equipment of the heat source equipment, and the heating temperature, flow rate and pressure difference of the pipe section medium, including: the equipment refers to heat source equipment such as heat pump system, boiler system, etc., and the flow rate, temperature and pressure difference refer to the medium flow rate, temperature and pressure difference from the heat source to the heat exchange station pipe section or from the heat exchange station to the terminal pipe section in the heat network; set the optimization objective function to minimize the operating energy consumption; the MILP mixed integer linear programming algorithm is:

[0072] minc T x

[0073] Ax≤b

[0074]

[0075] Among them: x is the decision variable vector, including the device switch state integer variable and the heating temperature and flow rate continuous variables; c is the objective function coefficient vector; A is the constraint matrix; b is the constraint vector; represents a set of integer variables, represents a set of continuous variables.

[0076] This embodiment further defines the step of S5. On the basis of obtaining the prediction results and the simulation model, the MILP mixed-integer linear programming algorithm is used to globally optimize the heating system. The optimization goal is to achieve the optimal operation of the system by adjusting the switch state, power, supply temperature and pressure difference. The MILP mixed-integer linear programming algorithm can handle complex multivariable optimization problems, find the optimal control strategy, and ensure that the heating system has the lowest energy consumption and the lowest operating cost under the premise of meeting user needs. MILP mixed-integer linear programming is used to optimize the following aspects: equipment switch state (integer variable), heating temperature, flow (continuous variable) and pressure difference (continuous variable). Through the optimization and control method of the present invention, accurate temperature control and heat distribution can improve the dynamic response ability of the system, reduce temperature fluctuations, and improve user comfort; predict environmental changes in advance, and adjust system operating parameters in time to adapt to different climatic conditions and user needs; optimize the operation mode of the heating system, while meeting the heating demand, minimize energy consumption and operating costs, reduce energy waste, reduce operating costs, ensure that the heating system can stably and efficiently meet the heating needs of users, and improve user satisfaction.

[0077] In this embodiment, first, the existing MILP mixed integer linear programming solver (such as CPLEX, Gurobi) is used to solve and establish the mathematical model of the return water temperature system; in view of the complexity of the heating network reaction, mechanism modeling is difficult to achieve. The kernel uses the step response curve method to establish a mathematical model for return water temperature control. The advantage of this method is that it is simple and easy to implement. When the system is stable, by adjusting the electric regulating valve on the return water pipe, the valve opening is increased from 70% to 72.4%, and the secondary network return water temperature rises from 33.7°C to 35.3°C. The test data is read by the PLC controller. Since the temperature control system has a delay, the operation process of this system can be approximated as a first-order inertia and pure lag link, and the object transfer function is solved. The following formula is the transfer function expression of the controlled object.

[0078] G(s)=K0e -τs / (T0s+1)

[0079] Where: K0 is the amplification factor of the process, T0 is the time constant of the inertia link, and τ is the pure lag time. Then solve K0 according to the following formula.

[0080]

[0081] Where: y(∞) is the steady-state value of the system, y(0) is the initial value of the system, and x0 is the step change. When the initial value of the system is 33.7℃ and the steady-state output value is 35.3℃, the electric control valve rises from 70% to 72.4%, that is, x0 is 2.4. Substituting it into the above formula, K0 is calculated to be 0.667.

[0082] Solve for T0 and τ. Assume that the tangent line intersects the X-axis at point τ. Select two points t1 and t2 on the curve so that τ<t1<t2. The corresponding function values ​​are y0(t1) and y0(t2). Substitute them into the following equations and solve for T0 and τ.

[0083]

[0084]

[0085] After experimental analysis, the mathematical model of the system was finally determined as:

[0086]

[0087] Then, the real-time feedback mechanism and overall coordination are carried out; in order to solve the problem of lack of timely feedback and overall coordination, the system introduces a real-time feedback mechanism and an overall coordination control module. By adding feedback loops to each link, the system can evaluate the effect of the current control strategy in real time and make adjustments immediately when deviations are detected. The overall coordination module ensures that the various parts of the system can cooperate with each other during optimization and control, avoiding the problem of local optimality leading to global performance degradation.

[0088] Then, the control strategy is executed. After the optimization is completed, the system applies the optimal control strategy to actual operation. Through the control module, the system can automatically adjust the operating status of the heating equipment, adjust the opening and closing of the valve, and change the heating temperature and pressure difference. The whole process is carried out automatically in the background without manual intervention, ensuring the intelligence and efficiency of the system.

[0089] Finally, feedback and iterative optimization are performed. The system continuously collects new data during actual operation and evaluates the effect of the control strategy in real time through the feedback mechanism. If deviations or situations that do not meet expectations are found, the system will automatically start the iterative optimization process, re-predict, simulate and optimize, and adjust the control strategy to ensure that the system is always in the optimal operating state.

[0090] Although the preferred embodiments of the present disclosure have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present disclosure.

[0091] Obviously, those skilled in the art can make various changes and modifications to the present disclosure without departing from the spirit and scope of the present disclosure. Thus, if these modifications and variations of the present disclosure fall within the scope of the claims of the present disclosure and their equivalents, the present disclosure is also intended to include these modifications and variations.

[0092] Finally, it should be noted that the above implementation modes are only used to illustrate the technical solutions of the present disclosure rather than to limit its protection scope. Although the present disclosure is described in detail with reference to the above implementation modes, ordinary technicians in the relevant field should understand that after reading the present disclosure, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation modes of the invention, but these changes, modifications or equivalent substitutions are all within the protection scope of the disclosed claims to be approved.

Claims

1. A method for coordinated dynamic optimization and control of a heating network, wherein the heating network includes a heating system, multiple heating stations, a pipe network and multiple heating equipment, characterized in that: The method comprises the following steps: S1. Real-time data collection, the data includes the operation data of the entire heating network and the meteorological data of the area where the entire heating network is located; the operation data includes the pipe network parameters, the real-time operation status of the heating station and the operation parameters of the heating equipment; S2. Inputting the collected data into the control system of the heating system, and preprocessing the collected data to obtain building load forecast data and thermal power station load forecast data; S3. constructing a load forecasting model using the building load forecasting data and the thermal power station load forecasting data, and training the load forecasting model based on the historical load data to obtain a trained load forecasting model; S4. Optimize the trained load prediction model and the pipe network parameters, and use the ARX self-regression algorithm to calculate the pipe network parameter prediction data; the pipe network parameter prediction data includes the pipe network pressure and the secondary side temperature of the pipe network; S5. Construct a physical simulation model of the entire heating network; solve the physical simulation model of the entire heating network using the MILP mixed integer linear programming algorithm based on the network parameter prediction data to obtain thermal load data of the entire heating network; S6. Optimize the physical simulation model of the entire heating network; obtain the optimal control strategy of the heating system by introducing a real-time feedback mechanism and an overall collaborative control module, use the optimal control strategy to automatically adjust the optimization parameters of the physical simulation model of the entire heating network, and automatically repeat the iteration to ensure that the entire heating network is continuously in the optimal operating state.

2. A method for coordinated dynamic optimization and control of the entire heating network according to claim 1, characterized in that: The process of preprocessing the collected data in S2 includes: Inputting the meteorological data and the historical load data into a machine learning algorithm to obtain the building load forecast data; The building load prediction data and the real-time operating status of the thermal power station are input into a machine learning algorithm to calculate the thermal power station load prediction data.

3. A method for coordinated dynamic optimization and control of the entire heating network according to claim 2, characterized in that: The building load prediction data includes building thermal load data, and the building thermal load data is based on the meteorological conditions of the area where the building is located; the meteorological data includes the temperature, humidity and wind speed of the external environment of the area where the heating network is located.

4. A method for coordinated dynamic optimization and control of the entire heating network according to claim 1, characterized in that: The optimization of the trained load prediction model and the pipe network parameters in S4 includes online parameter identification, hysteresis effect modeling and dynamic response mechanism; the pipe network parameters include pipe network pressure and pipe network temperature; The online parameter identification is to adjust the load prediction model parameters and continuously iterate through real-time monitoring of the real-time operating status and pipeline network parameters of the thermal power station in the operating data collected until the dynamic state and accurate value of the heating system are within a preset range; the hysteresis effect modeling and dynamic response mechanism is to add delay factors and dynamic characteristics to the load prediction model and pipeline network parameters.

5. A method for coordinated dynamic optimization and control of the entire heating network according to claim 1 or 4, characterized in that: The pipe network parameter prediction data obtained by using the ARX organic regression algorithm in S4 is: in, represents the load forecast result at time t for time t+1; q(t) represents the load at time t; Ta(t represents the outdoor temperature of the building at time t; I(t) represents whether time t is a holiday, which is 1 if it is, and 0 if it is not; a~d represent parameters; a1 represents the parameter of the load item at time t; a2 represents the parameter of the load item at time t-1; a 24 Represents the parameter of the load item at time t-23; a 25 Represents the parameter of the load item at time t-24; a 26 represents the parameter of the load item at time t-26; b1 represents the parameter of the outdoor temperature item at time t; b2 represents the parameter of the outdoor temperature item at time t-1; b 24 Represents the parameter of the outdoor temperature item at time t-23; b 25 represents the parameter of the outdoor temperature item at time t-24; c1,i represents the parameter of the sine correction item; c2,i represents the parameter of the cosine correction item; d 1, i represents the parameter of the holiday sinusoidal correction term; d 2, i represents the parameter of the holiday cosine correction term.

6. A method for coordinated dynamic optimization and control of the entire heating network according to claim 1, characterized in that: The method for constructing a physical simulation model of the entire heating network described in S5 is: constructing a system operation simulation environment by simulating the overall physical process of the heating system, and evaluating the operating characteristics of the system in the operation simulation environment under the optimal control strategy; the operating characteristics include the thermodynamic characteristics of the heating station, the thermodynamic characteristics of the pipeline network, the equipment operation characteristics and the system dynamic response.

7. A method for coordinated dynamic optimization and control of the entire heating network according to claim 6, characterized in that: The step of solving the physical simulation model of the heating network using the MILP mixed integer linear programming algorithm based on the pipe network parameter prediction data in S5 includes: inputting the pipe network parameter prediction data into the physical simulation model of the heating network; setting optimization parameters including the switch state of the heating equipment, the heating temperature, the flow rate and the pressure difference; setting the optimization objective function to minimize the operating energy consumption; the MILP mixed integer linear programming algorithm is: mincTx Ax Among them: x is the decision variable vector, including the integer variable of the switch state of the heating equipment and the continuous variables of the heating temperature and flow rate; c is the objective function coefficient vector; A is the constraint matrix; b is the constraint vector; represents a set of integer variables, represents a set of continuous variables.

8. A heat supply network collaborative dynamic optimization control system, the heat supply network includes a heat supply system, multiple heat stations, a pipe network and multiple heat supply equipment, characterized in that: The control system includes the following modules: Data acquisition module: used to collect data in real time, the data includes the operation data of the whole heating network and the meteorological data of the area where the whole heating network is located; the operation data includes the pipe network parameters, the real-time operation status of the heating station and the operation parameters of the heating equipment; Data preprocessing module: used to input the collected data into the control system of the heating system, and also used to preprocess the collected data to obtain building load forecast data and heating station load forecast data; A load forecasting model building module: used to build a load forecasting model using the building load forecasting data and the thermal power station load forecasting data, and also used to train the load forecasting model based on the historical load data to obtain a trained load forecasting model; Load forecasting model optimization module: used to optimize the trained load forecasting model and the pipe network parameters, and also used to calculate the pipe network parameter prediction data using the ARX organic self-regression algorithm; the pipe network parameter prediction data includes the pipe network pressure and the pipe network secondary side temperature; Heating network physical simulation model construction module: used to construct the heating network physical simulation model; based on the pipe network parameter prediction data, the MILP mixed integer linear programming algorithm is used to solve the heating network physical simulation model to obtain the heating network thermal load data; An optimal solution module for energy consumption operation; used to optimize the physical simulation model of the entire heating network; obtain the optimal control strategy of the heating system by introducing a real-time feedback mechanism and an overall collaborative control module, and automatically adjust the optimization parameters of the physical simulation model of the entire heating network using the optimal control strategy, and automatically repeat iterations to keep the entire heating network in the optimal operating state.

9. A control device for a method for collaborative dynamic optimization and control of a heating network, the device comprising a processor and a memory, wherein: The memory is used to store computer programs, and is characterized in that when the microprocessor reads the computer program on the memory, it executes a method for coordinated dynamic optimization and control of the entire heating network as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, a method for coordinated dynamic optimization and control of the entire heating network as described in any one of claims 1 to 7 is implemented.

Images