Solar central heating intelligent regulation and control system based on big data in alpine region

Abstract

The invention relates to the technical field of data analysis, in particular to an alpine region solar central heating intelligent regulation and control system based on big data, which comprises an acquisition processing unit, an analysis control unit and a decision execution unit. A data storage module in the analysis control unit adopts a distributed architecture to store data, predicts a heating demand and a solar heat collection trend by applying a hybrid algorithm fusing a deep learning neural network and a physical model, generates a control instruction and transmits the control instruction to a heat source adjusting module, and the heat source adjusting module adopts an incremental conductance method to adjust solar heating. And the valve driving module adjusts the opening degree of the valve based on a PID algorithm, and the regulation and control precision and the energy utilization efficiency of the heating system are improved.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a solar centralized heating intelligent control system based on big data in high-cold areas. Background Art

[0002] Data analysis is an important technology. In high-altitude and cold regions, solar heating systems face the problem of difficulty in precise regulation based on complex environments and actual needs. The climatic conditions in high-altitude and cold regions are complex and changeable, including low temperatures, strong winds, unstable solar radiation and other factors. These factors lead to unstable heat production efficiency of solar collectors, and heating demand also changes continuously.

[0003] The climate in high-altitude and cold regions is complex, and the solar heating system is affected by many factors. The temperature, flow and pressure data are complex and large in volume. Traditional control methods often rely on simple mathematical models and algorithms. Simple mathematical models and algorithms have limited data analysis capabilities and cannot accurately capture the nonlinear characteristics and dynamic changes in the data. On the other hand, simple mathematical models and algorithms lack intelligent decision-making mechanisms, and it is difficult to adjust the heating strategy in real time according to the data. This leads to the inability to accurately control the heating system. For example, heat cannot be allocated according to the actual needs of different regions, resulting in uneven indoor temperature. In order to solve this technical problem, we provide a solar centralized heating intelligent control system based on big data in high-altitude and cold regions. Summary of the invention

[0004] The purpose of the present invention is to provide a solar centralized heating intelligent control system based on big data in high-altitude cold areas to solve the problems raised in the above-mentioned background technology.

[0005] The present invention predicts the heating demand and heat generation trend of the solar thermal collector in the future, distributes the heat source using an intelligent allocation algorithm, and controls the valve opening size through a PID algorithm to assist the transmission of the heat source.

[0006] To achieve the above purpose, a solar centralized heating intelligent control system based on big data in high-cold areas is provided, including a collection and processing unit, an analysis and control unit, and a decision-making execution unit;

[0007] The collection and processing unit uses sensors of solar collectors, heat storage devices and heating pipeline nodes to collect temperature data, flow data and pressure data, and sets an opening sensor and an action time recorder at the valve to collect the operation data of the heating system in real time;

[0008] The analysis and control unit includes a data storage module, a data analysis module and a control generation module. The data storage module is used to store the data collected by the collection and processing unit. The data analysis module uses a hybrid algorithm that integrates deep learning neural networks and physical models to predict changes in heating demand and heat generation trends of solar collectors based on historical data and real-time data. The control generation module generates control instructions based on the prediction results and the preset intelligent control strategy.

[0009] The decision-making execution unit includes a valve driving module and a heat source regulation module. The heat source regulation module adjusts solar heating with a maximum power point tracking algorithm based on control instructions, and stores excess heat in a heat storage device through an intelligent distribution algorithm. When the heating demand exceeds a preset threshold, the valve of the heat storage device is started through the valve driving module, and heat is distributed and the valve opening is adjusted according to the thermal balance model of each area.

[0010] As a further improvement of the present technical solution, the data storage module adopts a distributed storage architecture to store data, as follows:

[0011] The collected temperature, pressure and flow data are sharded according to the time dimension, and the sharded data are distributed to a number of storage nodes, where the storage nodes are storage devices located at different physical locations, and each storage node stores the temperature, pressure and flow data of one time dimension;

[0012] A metadata index is constructed, wherein the metadata includes the time period, data type, and storage location to which the data belongs, and a distributed consistency protocol is used to update and delete data on the same node.

[0013] As a further improvement of the technical solution, the deep learning neural network in the data analysis module adopts a convolutional neural network structure in the hybrid algorithm to extract spatial features in temperature and flow data, as follows:

[0014] Arrange the temperature and flow data according to time series and spatial position to form a two-dimensional data matrix, normalize the formatted data, and map the data values ​​to the [0, 1] interval;

[0015] The convolution kernel of the deep learning neural network is defined as 3×3, and the convolution operation is performed by sliding on the data matrix. That is, at each position, the convolution kernel performs multiplication and addition operations with the corresponding data window to generate a new eigenvalue. After the convolution operation, each convolution layer generates a feature map.

[0016] A pooling layer is set after the convolution layer, and the maximum pooling operation is selected to be performed on the feature map. The feature map is downsampled according to the preset pooling window size. After passing through the convolution layer and the pooling layer, the spatial features of the data are extracted. Finally, these features are integrated through the fully connected layer to form a vector representation for prediction.

[0017] As a further improvement of the technical solution, the construction method of the physical model in the data analysis module is as follows:

[0018] The physical model is constructed based on the energy balance principle, and the expression of energy balance is Q in -Q loss =Q ed ; where Q in Expressed as the input of solar radiation energy, Q loss Expressed as heat loss, Q ed Expressed as final heat energy;

[0019] For Q in The average solar altitude angle and atmospheric transparency in high-altitude cold areas are obtained through meteorological observation data, and then substituted into the astronomical parameters and atmospheric transmittance model for calculation;

[0020] For Q loss The heat loss caused by heat conduction and convection is taken into account. For heat conduction loss, a heat conduction loss model is established according to the thermal conductivity of the collector material and the structural size. For convection loss, natural convection and forced convection are considered, and a convection loss model is established based on the principles of fluid mechanics and the temperature difference between the ambient temperature and the collector surface temperature. The heat loss is calculated by combining the heat conduction and convection loss models with the ambient temperature parameters.

[0021] As a further improvement of the technical solution, the data analysis module predicts the change of heating demand and the heat generation trend of the solar thermal collector based on historical data and real-time data, as follows:

[0022] The weight of the deep learning neural network in heating demand is determined by verifying the prediction accuracy of the deep learning neural network on historical data, and the weight of the physical model in heating demand is evaluated by comparing the calculation results of the physical model with the actual measured changes in heating demand;

[0023] For the deep neural network, the historical data containing temperature and flow data are used to train the deep learning neural network, obtain the potential relationship between spatial characteristics and heating demand changes, and output the heating demand change trend value based on the potential relationship. For the physical model, according to Q edAs well as the actual operation of the current heating system, derive the trend value of the heating demand change, multiply the deep neural network trend value by its weight to obtain the neural network weighted value, then multiply the physical model trend value by its weight to obtain the physical model weighted value, add the neural network weighted value and the physical model weighted value to obtain the final heating demand change prediction value;

[0024] The final heat generation trend prediction value of the solar thermal collector is obtained by the same operation method as above.

[0025] As a further improvement of the present technical solution, the control generation module generates control instructions based on the prediction results and the preset intelligent control strategy, as follows:

[0026] Based on the predicted value of heating demand change and the predicted value of heat generation trend of solar collectors, the heating demand and heat generation capacity of solar energy in high-cold areas are determined;

[0027] According to the preset time-sharing and partitioning strategy, determine the time period and corresponding partitioning rules of the current time, and set priorities according to the importance of different areas;

[0028] Finally, according to the changes in heating demand and zoning strategy, valve control instructions are generated to adjust the valve opening of the heating pipe, and heat source adjustment instructions are generated in combination with the heat production trend of the solar collector and energy supplement demand.

[0029] As a further improvement of the technical solution, the heat source regulation module adopts the incremental conductance method when adjusting solar heating with the maximum power point tracking algorithm, as follows:

[0030] The solar radiation intensity is obtained through the sensor at the solar collector, and the change state of the solar radiation intensity is determined according to the preset rules;

[0031] If the rate of change of solar radiation intensity exceeds the set first threshold value in several consecutive cycles, it is determined that the solar radiation intensity changes quickly, and the step size is increased to α. Then, an incremental conductivity calculation is performed based on the increased step size, and the working voltage of the solar collector is adjusted according to the calculation result. If the rate of change of solar radiation intensity is lower than the second threshold value in several consecutive sampling cycles, it is determined that the solar radiation intensity is in a stable state, and the step size is reduced to β. The incremental conductivity calculation is also performed and the working voltage of the solar collector is adjusted. The step size is used to adjust and track the maximum power point.

[0032] As a further improvement of the technical solution, the heat source regulation module stores the excess heat in the heat storage device through an intelligent allocation algorithm, as follows:

[0033] The data including the outlet of the solar collector, the branches of the heating pipeline and the inside of the heat storage device are obtained from the collection and processing unit, and the solar energy input heat is calculated according to the energy input formula of the solar collector. At the same time, the total heat demand is calculated according to the heat demand of each area, and the excess heat is obtained according to the solar energy input heat and the total heat demand;

[0034] Based on the temperature sensor data in the heat storage device, a temperature distribution model is constructed to determine the temperature values ​​and remaining heat storage capacity of different areas. The areas are sorted according to the temperature gradient, and the areas with the largest temperature gradient are given priority. According to the allocation plan determined by the intelligent allocation algorithm, the control system adjusts the valve opening connecting the various areas of the heat storage device. The valve opening is adjusted according to the set flow target value and the actual flow feedback value through the control algorithm.

[0035] As a further improvement of the technical solution, the heat source regulating module regulates the valve opening through the valve driving module, as follows:

[0036] The valve driving module regulates the valve opening based on the PID control algorithm, and performs initialization settings on the PID algorithm, wherein the initialization settings include setting the initial values ​​of the proportional coefficient, the integral coefficient, and the differential coefficient. The valve driving module continuously receives the actual flow value fed back by the flow sensor. At the same time, the valve driving module stores a preset flow target value, which is calculated based on the intelligent allocation algorithm and is used to guide the flow rate that needs to be achieved in the current heat storage process, and to filter the collected actual flow value;

[0037] Calculate the error e(t), which is obtained by the difference between the flow target value and the actual flow value, set the proportional link P, integral link I and differential link D, the size of the proportional link error produces a control effect, and its output u P (t) = K P e(t);K P The function of is to make the change of valve opening proportional to the error. The integral link I is used to eliminate the steady-state error of the valve drive module. Its output The differential link D adjusts the control signal according to the rate of change of the error, and its output

[0038] The PID control algorithm adds the outputs of the proportional, integral and differential links to obtain the final control signal. The valve drive module adjusts the valve opening according to the size and direction of the control signal.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] In a solar centralized heating intelligent control system based on big data in high-cold areas, a hybrid algorithm that integrates deep learning neural networks and physical models in the data analysis module is used to improve the accuracy of predicting changes in heating demand and heat generation trends of solar collectors. The control generation module adjusts the valve opening and generates heat source adjustment instructions in combination with the time-sharing and zoning strategy. The heat source adjustment module uses the incremental conductance method to adaptively step when adjusting solar heating, which can better track the maximum power point. It can also reasonably store excess heat in a heat storage device through an intelligent allocation algorithm, and use a PID control algorithm to accurately adjust the valve opening to achieve efficient heating control. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is an overall block diagram of the present invention.

[0042] The meaning of each number in the figure is:

[0043] 1. Acquisition and processing unit; 2. Analysis and control unit; 21. Data storage module; 22. Data analysis module; 23. Control generation module; 3. Decision execution unit; 31. Valve drive module; 32. Heat source regulation module. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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] In high-altitude and cold regions, the solar heating system has complex climate, resulting in unstable heat production efficiency of solar collectors and constantly changing heating demand. At the same time, traditional control methods rely on simple models and algorithms, with limited data analysis capabilities, unable to capture the nonlinear characteristics and dynamic changes of data, and lack of intelligent decision-making mechanisms, making it difficult to accurately control the heating system.

[0046] The present invention provides a solar centralized heating intelligent control system based on big data in high-cold areas. Figure 1 As shown, it includes a collection and processing unit 1, an analysis and control unit 2, and a decision execution unit 3;

[0047] In order to fully understand the operating conditions of the heating system, including obtaining relevant temperature data such as the heat production efficiency of the solar collector, the heat storage of the heat storage device, the heat transfer loss of the heating pipeline and the heating temperature of each area, determine the flow data of the heat transfer rate and distribution, and the pressure data reflecting the operating pressure conditions of the system. At the same time, the valve opening degree and action time are accurately grasped through the valve opening sensor and the action time recorder, providing accurate data support for subsequent analysis, regulation and optimization of the system, ensuring safe, reliable and efficient operation of the system.

[0048] The collection and processing unit 1 uses sensors of solar collectors, heat storage devices and heating pipeline nodes to collect temperature data, flow data and pressure data, and sets an opening sensor and an action time recorder at the valve to collect the operation data of the heating system in real time;

[0049] The heating system will generate a large amount of temperature, pressure and flow data during operation, especially in the case of long-term operation and multiple monitoring points. The use of a distributed storage architecture can store data in multiple nodes, avoiding the problem of insufficient capacity of a single storage device and effectively handling the storage needs of large amounts of data.

[0050] The data storage module 21 uses a distributed storage architecture to store data, as follows:

[0051] The collected temperature, pressure and flow data are sharded according to the time dimension. For example, the data can be sharded by hour, day or week, and the data of different time periods are separated. The sharded data are distributed to several storage nodes located in different physical locations. Each storage node stores the temperature, pressure and flow data of one time dimension.

[0052] Build a metadata index. The metadata records that a certain temperature data belongs to the time period of 9:00-10:00 a.m., which is temperature type data and is stored on node A. When new data needs to be updated to a certain node, the consistency protocol is used to ensure that all relevant nodes can correctly update the data status to avoid data inconsistency, providing a good data foundation for subsequent data analysis and system regulation.

[0053] The analysis and control unit 2 includes a data storage module 21, a data analysis module 22 and a control generation module 23. The data storage module 21 is used to store the data collected by the collection and processing unit 1 to provide a basis for subsequent analysis and control. The data analysis module 22 uses a hybrid algorithm that integrates deep learning neural networks and physical models to predict changes in heating demand and heat generation trends of solar collectors based on historical data and real-time data. A single method is difficult to accurately grasp the changes in complex heating systems. The hybrid algorithm can combine the advantages of both to improve prediction accuracy. The control generation module 23 generates control instructions based on the prediction results and the preset intelligent control strategy, so that the operation of the heating system can be reasonably adjusted according to the actual needs and future trends of the system to achieve efficient and energy-saving heating goals;

[0054] Temperature and flow data have spatial distribution characteristics in the heating system. For example, the temperature and flow may vary in different areas. This spatial difference is crucial for accurately predicting heating demand and heat generation trends of solar collectors. The convolutional neural network structure can effectively mine the spatial features in the data. By sliding the convolution kernel on the data matrix, it can capture the associations and patterns between data at different locations.

[0055] The deep learning neural network in the data analysis module 22 adopts a convolutional neural network structure in the hybrid algorithm. The convolutional neural network structure can automatically learn and extract spatial features in temperature and flow data. Compared with the traditional feature extraction method, it does not require manual design of complex feature engineering, reduces the impact of human factors on feature extraction, and improves the accuracy and objectivity of feature extraction. It is used to extract spatial features in temperature and flow data, as follows:

[0056] Arrange the temperature and flow data according to time series and spatial position to form a two-dimensional data matrix. For example, for a heating system with multiple zones, the temperature and flow data collected at different time points can be arranged according to the zone position and time sequence, so as to construct a data matrix reflecting the spatiotemporal characteristics of the system. The formatted data is normalized and the data values ​​are mapped to the [0, 1] interval, which helps to improve the training efficiency and stability of the neural network, so that data of different magnitudes can be processed at the same scale.

[0057] The convolution kernel of the deep learning neural network is defined as 3×3. For example, for temperature and flow data, a 3x3 convolution kernel may be selected, and then the convolution operation is performed by sliding on the data matrix. That is, at each position, the convolution kernel performs multiplication and addition operations with the corresponding data window to generate a new eigenvalue. By continuously sliding the convolution kernel, the entire data matrix can be convolved. After the convolution operation, each convolution layer generates a feature map;

[0058] A pooling layer is set after the convolution layer, and the maximum pooling operation is selected to be performed on the feature map. The feature map is downsampled according to the preset pooling window size. For example, if the feature map is an 8x8 matrix, after a 2x2 pooling window operation, a 4x4 downsampling matrix will be obtained. The pooling operation can reduce the dimension of the data while retaining the main features of the data. After the convolution layer and the pooling layer, the spatial features of the data are extracted. Finally, these features are integrated through the fully connected layer to form a vector representation for prediction. The fully connected layer synthesizes each feature and maps the feature to a prediction vector according to the weights obtained through training. It is used for subsequent predictions of changes in heating demand and heat generation trends of solar collectors, thereby providing a more accurate basis for predictions of changes in heating demand and heat generation trends of solar collectors.

[0059] The purpose of constructing a physical model based on the energy balance principle is to describe the energy conversion and transfer process in the solar heating system from the physical essence. By clarifying the relationship between solar radiation energy input, heat loss and final thermal energy, the working principle and performance of the system can be understood more accurately.

[0060] The construction method of the physical model in the data analysis module 22 is that the physical model is constructed based on the energy balance principle and the actual physical process. Its various parameters and formulas have clear physical meanings, which makes the model easy to understand and interpret, and can provide a clear physical basis for the analysis and optimization of the system, as follows:

[0061] The physical model is constructed based on the principle of energy balance. The expression of energy balance is Q in -Q loss =Q ed ; where Q in Expressed as the input of solar radiation energy, Q loss Expressed as heat loss, Q ed Expressed as the final thermal energy, the model comprehensively considers multiple factors such as solar radiation energy input, heat conduction and convection loss, which helps to more accurately predict the performance of the system and changes in heating demand, and improve the control effect of the system.

[0062] For Q in The average solar altitude angle θ and atmospheric transparency τ in the high-cold region are obtained through meteorological observation data, and then substituted into the astronomical parameters and atmospheric transmittance model for calculation, that is, Q in =I0×τ×sin(θ); where I0 is the solar constant, which represents the solar radiation energy received per unit area per unit time outside the Earth's atmosphere, perpendicular to the direction of the sun's rays.

[0063] For Q lossThe heat loss caused by heat conduction and convection is considered. For the heat conduction loss, a heat conduction loss model is established according to the thermal conductivity χ of the collector material and the thickness d and area S of the collector. The heat conduction loss can be expressed as Where ΔT is the temperature difference on both sides of the collector. For convection loss, natural convection and forced convection are considered, and according to the principles of fluid mechanics and the ambient temperature T hj and collector surface temperature T s The temperature difference is used to establish the convection loss model, and then the convection loss Q is calculated. dl =S×(T s -T hj ) ; Combine the heat conduction and convection loss models with the ambient temperature parameters to calculate the heat loss Q loss =Q red +Q dl ; The physical model can more accurately calculate the solar radiation energy input and heat loss, thereby predicting the final thermal energy of the system, that is, the actual heating capacity of the heating system, which helps to plan the heating strategy in advance and avoid insufficient heating or energy waste.

[0064] Deep learning neural networks and physical models each have their own advantages and disadvantages. Neural networks can automatically learn complex patterns in data, but may lack physical interpretability. Physical models are based on physical principles and have clear physical meanings, but may not be able to fully capture the complex nonlinear relationships in the data. By combining the two and determining the weights, their advantages can be combined to improve the accuracy of predictions.

[0065] The data analysis module 22 predicts the change of heating demand and the heat generation trend of the solar thermal collector based on historical data and real-time data, as follows:

[0066] Determine the weight of deep neural network in heating demand by verifying the prediction accuracy of deep learning neural network on historical data NN By comparing the calculation results of the physical model with the actual measured heating demand change value, the weight of the physical model in the heating demand is evaluated. PM ;

[0067] For deep neural networks, historical data containing temperature and flow data are used to train the deep learning neural network, obtain the potential relationship between spatial characteristics and changes in heating demand, and output the trend value of heating demand changes based on the potential relationship For the physical model, according to Q ed As well as the actual operation of the current heating system, derive the trend value of heating demand changes Multiply the deep neural network trend value by its weight to get the neural network weighted value Then multiply the trend value of the physical model by its weight to get the weighted value of the physical model Add the weighted values ​​of the neural network and the physical model to get the final predicted value of heating demand change

[0068] The final prediction value of the heat generation trend of the solar thermal collector is obtained by the same operation method as above, that is, the weights of the neural network and the physical model in the prediction of the heat generation trend of the solar thermal collector are determined respectively, the trend values ​​of each are calculated, and then the weighted sum is obtained to obtain the final prediction value of the heat generation trend of the solar thermal collector. in and They are the heat generation trend values ​​of the solar collector calculated by the neural network and the physical model respectively. By combining the two methods and reasonably determining the weights, it is possible to more accurately predict changes in heating demand and the heat generation trend of the solar collector. More accurate prediction results provide a reliable basis for heating control strategies. For example, the valve opening and heat source output can be adjusted in advance according to the predicted changes in heating demand to avoid insufficient heating or energy waste, thereby improving the energy utilization rate and heating effect of the heating system.

[0069] The control generation module 23 generates control instructions based on the prediction results and the preset intelligent control strategy, as follows:

[0070] Suppose the current heating demand in a certain area is Q GN , the predicted heating demand change rate is Δr GN , then the predicted heating demand can be expressed as GN yc =Q GN (1+Δr GN );

[0071] Assume that the current heat generation capacity of the solar collector is Q JR , the predicted heat generation trend change rate is Δr JR , then the predicted solar thermal capacity can be expressed as JR yc =Q JR (1+Δr JR );

[0072] According to the preset time-sharing and zoning strategy, determine which time period of the day the current time is in, such as daytime, nighttime, or winter and non-winter seasons of the year, and the corresponding zoning rules. For example, during the daytime in winter, divide the commercial area and the residential area into different zones, and set priorities according to their importance. The commercial area has a higher priority because of the frequent activities of people; the residential area has a higher priority according to the population density.

[0073] According to the changes in heating demand and the zoning strategy, valve control instructions are generated to adjust the opening degree of the valves in the heating pipelines. If the heating demand in a certain area increases and the priority is high, the opening degree of the valves in the heating pipelines of that area is increased to increase the heat supply; conversely, if the heating demand decreases, the opening degree of the valves is decreased.

[0074] Let the initial opening degree of the valve be k0, the change rate of heating demand be Δr GN , the area priority be p, and the valve opening degree adjustment coefficient be γ. Then the adjusted valve opening degree k can be expressed as k = k0(1 + pγΔr GN ); when Δr GN > 0 and p = 0.8, compared with the area where p = 0.3, the valve opening degree increases. When Δr GN < 0, k < k0, and the valve opening degree decreases.

[0075] Combined with the heat production trend of the solar collector and the energy supplement demand, heat source adjustment instructions are generated. Let the current total output power of the heat source be P total , the heat production power of the solar collector be P solar , and the heating demand power be P GN . If P solar ≥P GN , then the adjusted total output power P total of the heat source can be expressed as P total = P solar ; if P solar < P GN , then the maximum power point tracking algorithm is started to adjust the solar heating, and the heat is reasonably distributed according to the actual demand, making the heating process more efficient, reducing energy waste, meeting the heating demands of different regions and time periods at the same time, improving the heating effect, and contributing to the development of heating technology towards a more efficient and intelligent direction.

[0076] To improve the solar energy utilization efficiency and adapt to environmental changes, the solar heating is adjusted through the maximum power point tracking algorithm to balance the heat supply and demand, optimize the energy utilization, store the excess heat in the heat storage device to meet the additional heating demand, and reasonably distribute the heat and adjust the valve opening degree according to the heat balance model to ensure the indoor temperature is comfortable and improve the heating quality, system control accuracy and energy utilization efficiency.

[0077] The decision execution unit 3 includes a valve drive module 31 and a heat source adjustment module 32. The heat source adjustment module 32 adjusts the solar heating based on the control instructions by the maximum power point tracking algorithm and stores the excess heat in the heat storage device through the intelligent distribution algorithm. When the heating demand exceeds the preset threshold, the valve of the heat storage device is started through the valve drive module 31, and the heat is distributed and the valve opening degree is adjusted according to the heat balance model of each region.

[0078] The heat source regulation module 32 uses the incremental conductance method when adjusting solar heating using the maximum power point tracking algorithm, as follows:

[0079] The solar radiation intensity is obtained through the sensor at the solar collector, and the change state of the solar radiation intensity is determined according to the preset rules;

[0080] If the rate of change of solar radiation intensity exceeds the set first threshold value ΔI1 within several consecutive cycles, it is determined that the solar radiation intensity changes quickly, and the step size is increased to ΔV1. Then, an incremental conductance calculation is performed based on the increased step size. The core formula of the incremental conductance calculation is based on the maximum power point tracking principle. For a solar collector, its output power is P=VI, where V is voltage and I is current. At the maximum power, When calculated by the incremental conductance method, The current variation ΔI is calculated by combining the formula with the increased step size ΔV1, and then the operating voltage of the solar collector is adjusted.

[0081] If the rate of change of solar radiation intensity is lower than the second threshold value ΔI2 within several consecutive sampling periods, it is determined that the solar radiation intensity is in a stable state, and the step size is reduced to ΔV2. The incremental conduction is calculated in the same way, that is, according to the stable state The current change ΔI is calculated in combination with the reduced step size ΔV2, and the operating voltage of the solar collector is adjusted. The step size is used to adjust and track the maximum power point. The step size is quickly adjusted according to the changing state of solar radiation intensity, so as to track the maximum power point more effectively. When the solar radiation intensity changes rapidly, increasing the step size can adapt to the changes more quickly, so that the solar collector can be adjusted to a working state close to the maximum power point in a short time; when the solar radiation intensity is stable, reducing the step size can improve the tracking accuracy, ensuring that the solar collector works stably near the maximum power point, so that the solar collector can make better use of solar energy resources, and maintain a high energy conversion efficiency under different changes in solar radiation intensity, thereby increasing the proportion of solar energy in the heating system and reducing dependence on other energy sources.

[0082] The heat source regulation module 32 stores the excess heat in the heat storage device through the intelligent distribution algorithm, as follows:

[0083] The data collected from the collection and processing unit 1 include the solar collector outlet temperature T out 、Flow Q out Temperature T of each branch of the heating pipe fz 、Flow Q fz And the internal temperature of the thermal storage device T xr , and calculate the solar energy input heat according to the energy input formula of the solar collector, that is, Q JR=c×m×(T out -T in ), where c is the specific heat capacity, m is the mass of the medium, T in is the inlet temperature of the collector. At the same time, the total heat demand is calculated based on the heat demand of each area. Where i represents the region number, m i is the medium quality of each area, T set is the set temperature, T room，i For the indoor temperature of each area, the excess heat Q is obtained according to the solar energy input heat and the total heat demand dy , i.e. Q dy =Q JR -Q xq ; Reasonable calculation and utilization of excess heat, storing it in a heat storage device, and releasing it when needed, improves the energy efficiency of the entire heating system and reduces the need for additional energy sources.

[0084] According to the temperature sensor data in the thermal storage device, a temperature distribution model is constructed. Assume that there are n areas in the thermal storage device, and the temperatures are T1, T2, ..., T n , the remaining heat storage capacities are C1, C2, ..., C n , sort the areas according to the temperature gradient, give priority to the area with the largest temperature gradient, and according to the allocation plan determined by the intelligent allocation algorithm, the control system adjusts the valve opening of each area connected to the heat storage device. The control algorithm adjusts the valve opening according to the set flow target value and the actual flow feedback value. The heat storage area is reasonably selected through the temperature distribution model, avoiding local overheating or overcooling of the heat storage device due to unreasonable heat storage, extending the service life of the heat storage device, and the reasonable storage and utilization of excess heat enables the heating system to maintain a stable heating capacity when the solar energy is unstable or the heating demand fluctuates, avoiding large fluctuations in indoor temperature.

[0085] In the heating system, the flow rate will change dynamically with time and various factors. The integral and differential links of the PID control algorithm can adapt to this dynamic change, adjust the valve opening in time, and ensure the stable operation of the system.

[0086] The heat source regulating module 32 regulates the valve opening through the valve driving module 31, as follows:

[0087] The valve drive module 31 regulates the valve opening based on the PID control algorithm and initializes the PID algorithm. The initialization setting includes setting the initial values ​​of the proportional coefficient, the integral coefficient and the differential coefficient. The valve drive module 31 continuously receives the actual flow value fed back by the flow sensor. At the same time, the valve drive module 31 stores a preset flow target value. The flow target value is calculated based on the intelligent allocation algorithm and is used to guide the flow rate that needs to be achieved in the current heat storage process, and to filter the collected actual flow value.

[0088] By properly adjusting the proportional, integral and differential coefficients, the control accuracy of the valve opening can reach a very high level, meeting the heating system's requirements for precise flow control.

[0089] Calculate the error e(t), which is obtained by the difference between the flow target value and the actual flow value. Set the proportional link P, integral link I and differential link D. The size of the proportional link error produces a control effect, and its output u P (t) = K P e(t);K P The function of is to make the change of valve opening proportional to the error. The integral link I is used to eliminate the steady-state error of the valve drive module 31. Its output The differential link D adjusts the control signal according to the rate of change of the error, and its output Since the PID control algorithm takes into account the accumulation and rate of change of errors, it can maintain good stability when there are steady-state errors and dynamic changes in the system. Even in a complex heating system environment, it can ensure stable control of the valve opening, thereby ensuring stable operation of the system.

[0090] The PID control algorithm adds the outputs of the three links of proportional, integral and differential to obtain the final control signal, that is, u(t) = u P (t)+u I (t)+u D (t); the valve driving module 31 adjusts the valve opening k according to the size and direction of the control signal. If the control signal is positive, the valve opening increases, and if the control signal is negative, the valve opening decreases, that is, the final valve opening k = σ × u(t); wherein σ is a proportional constant. By accurately adjusting the valve opening through the PID control algorithm, precise control of the flow rate during the heat storage process can be achieved, ensuring that the heat can be stored and distributed according to the predetermined plan, thereby improving the accuracy and stability of heat storage.

[0091] In the present invention, multi-source data is collected by the acquisition processing unit, the data storage module in the analysis control unit adopts a distributed architecture to store data, and a hybrid algorithm integrating deep learning neural network and physical model is used to predict heating demand and solar energy collection trend, and control instructions are generated and transmitted to the heat source regulation module. The heat source regulation module adopts the incremental conductance method to adjust solar energy heating and stores excess heat in the heat storage device. The valve drive module adjusts the valve opening based on the PID algorithm to improve the control accuracy and energy utilization efficiency of the heating system.

[0092] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A solar central heating intelligent control system based on big data in high-cold areas, characterized by: It comprises a collection and processing unit (1), an analysis and control unit (2) and a decision execution unit (3); The collection and processing unit (1) uses sensors of solar collectors, heat storage devices and heating pipeline nodes to collect temperature data, flow data and pressure data, and sets an opening sensor and an action time recorder at the valve to collect the operation data of the heating system in real time; The analysis control unit (2) comprises a data storage module (21), a data analysis module (22) and a control generation module (23), wherein the data storage module (21) is used to store data collected by the collection and processing unit (1), the data analysis module (22) uses a hybrid algorithm that integrates a deep learning neural network and a physical model to predict changes in heating demand and heat generation trends of solar collectors based on historical data and real-time data, and the control generation module (23) generates control instructions based on the prediction results and a preset intelligent control strategy; The decision execution unit (3) comprises a valve drive module (31) and a heat source adjustment module (32). The heat source adjustment module (32) adjusts solar heating using a maximum power point tracking algorithm based on a control instruction, and stores excess heat in a heat storage device through an intelligent allocation algorithm. When the heating demand exceeds a preset threshold, the valve of the heat storage device is activated through the valve drive module (31), and heat is distributed and the valve opening is adjusted according to a heat balance model of each area.

2. According to claim 1, a solar centralized heating intelligent control system based on big data in high-cold areas is characterized by: The data storage module (21) uses a distributed storage architecture to store data, specifically as follows: The collected temperature, pressure and flow data are sharded according to the time dimension, and the sharded data are distributed to a number of storage nodes, where the storage nodes are storage devices located at different physical locations, and each storage node stores the temperature, pressure and flow data of one time dimension; A metadata index is constructed, wherein the metadata includes the time period, data type, and storage location to which the data belongs, and a distributed consistency protocol is used to update and delete data on the same node.

3. According to claim 1, a solar centralized heating intelligent control system based on big data in high-cold areas is characterized by: The deep learning neural network in the data analysis module (22) adopts a convolutional neural network structure in the hybrid algorithm to extract spatial features from the temperature and flow data, as follows: Arrange the temperature and flow data according to time series and spatial position to form a two-dimensional data matrix, normalize the formatted data, and map the data values ​​to the [0, 1] interval; The convolution kernel of the deep learning neural network is defined as 3×3, and the convolution operation is performed by sliding on the data matrix. That is, at each position, the convolution kernel performs multiplication and addition operations with the corresponding data window to generate a new eigenvalue. After the convolution operation, each convolution layer generates a feature map. A pooling layer is set after the convolution layer, and the maximum pooling operation is selected to be performed on the feature map. The feature map is downsampled according to the preset pooling window size. After passing through the convolution layer and the pooling layer, the spatial features of the data are extracted. Finally, these features are integrated through the fully connected layer to form a vector representation for prediction.

4. According to claim 1, a solar centralized heating intelligent control system based on big data in high-cold areas is characterized by: The construction method of the physical model in the data analysis module (22) is as follows: The physical model is constructed based on the energy balance principle, and the expression of energy balance is Q in -Q loss =Q ed ; where Q in Expressed as the input of solar radiation energy, Q loss Expressed as heat loss, Q ed Expressed as final heat energy; For Q in The average solar altitude angle and atmospheric transparency in high-altitude cold areas are obtained through meteorological observation data, and then substituted into the astronomical parameters and atmospheric transmittance model for calculation; For Q loss The heat loss caused by heat conduction and convection is taken into account. For heat conduction loss, a heat conduction loss model is established according to the thermal conductivity of the collector material and the structural size. For convection loss, natural convection and forced convection are considered, and a convection loss model is established based on the principles of fluid mechanics and the temperature difference between the ambient temperature and the collector surface temperature. The heat loss is calculated by combining the heat conduction and convection loss models with the ambient temperature parameters.

5. According to claim 1, a solar centralized heating intelligent control system based on big data in high-cold areas is characterized by: The data analysis module (22) predicts the change in heating demand and the heat generation trend of the solar thermal collector based on historical data and real-time data, as follows: The weight of the deep learning neural network in heating demand is determined by verifying the prediction accuracy of the deep learning neural network on historical data, and the weight of the physical model in heating demand is evaluated by comparing the calculation results of the physical model with the actual measured changes in heating demand; For the deep neural network, the historical data containing temperature and flow data are used to train the deep learning neural network, obtain the potential relationship between spatial characteristics and heating demand changes, and output the heating demand change trend value based on the potential relationship. For the physical model, according to Q ed As well as the actual operation of the current heating system, derive the trend value of the heating demand change, multiply the deep neural network trend value by its weight to obtain the neural network weighted value, then multiply the physical model trend value by its weight to obtain the physical model weighted value, add the neural network weighted value and the physical model weighted value to obtain the final heating demand change prediction value; The final heat generation trend prediction value of the solar thermal collector is obtained by the same operation method as above.

6. According to claim 1, a solar centralized heating intelligent control system based on big data in high-cold areas is characterized by: The control generation module (23) generates control instructions based on the prediction results and the preset intelligent control strategy, specifically as follows: Based on the predicted value of heating demand change and the predicted value of heat generation trend of solar collectors, the heating demand and heat generation capacity of solar energy in high-cold areas are determined; According to the preset time-sharing and partitioning strategy, determine the time period and corresponding partitioning rules of the current time, and set priorities according to the importance of different areas; Finally, according to the changes in heating demand and zoning strategy, valve control instructions are generated to adjust the valve opening of the heating pipe, and heat source adjustment instructions are generated in combination with the heat production trend of the solar collector and energy supplement demand.

7. According to claim 1, a solar centralized heating intelligent control system based on big data in high-cold areas is characterized by: The heat source adjustment module (32) uses an incremental conductance method when adjusting solar heating using a maximum power point tracking algorithm, as follows: The solar radiation intensity is obtained through the sensor at the solar collector, and the change state of the solar radiation intensity is determined according to the preset rules; If the rate of change of solar radiation intensity exceeds the set first threshold value in several consecutive cycles, it is determined that the solar radiation intensity changes quickly, and the step size is increased to α. Then, an incremental conductivity calculation is performed based on the increased step size, and the working voltage of the solar collector is adjusted according to the calculation result. If the rate of change of solar radiation intensity is lower than the second threshold value in several consecutive sampling cycles, it is determined that the solar radiation intensity is in a stable state, and the step size is reduced to β. The incremental conductivity calculation is also performed and the working voltage of the solar collector is adjusted. The step size is used to adjust and track the maximum power point.

8. According to claim 3, a solar centralized heating intelligent control system based on big data in high-cold areas is characterized by: The heat source regulating module (32) stores the excess heat in the heat storage device through an intelligent distribution algorithm, specifically as follows: Acquire data including the outlet of the solar thermal collector, branches of the heating pipeline and the inside of the heat storage device from the collection and processing unit (1), calculate the solar thermal input heat according to the energy input formula of the solar thermal collector, and at the same time calculate the total heat demand according to the heat demand of each area, and obtain excess heat according to the solar thermal input heat and the total heat demand; Based on the temperature sensor data in the heat storage device, a temperature distribution model is constructed to determine the temperature values ​​and remaining heat storage capacity of different areas. The areas are sorted according to the temperature gradient, and the areas with the largest temperature gradient are given priority. According to the allocation plan determined by the intelligent allocation algorithm, the control system adjusts the valve opening connecting the various areas of the heat storage device. The valve opening is adjusted according to the set flow target value and the actual flow feedback value through the control algorithm.

9. According to claim 8, a solar centralized heating intelligent control system based on big data in high-cold areas is characterized by: The heat source regulating module (32) regulates the valve opening through the valve driving module (31), specifically as follows: The valve driving module (31) controls the valve opening based on the PID control algorithm and performs initialization settings on the PID algorithm, wherein the initialization settings include setting initial values ​​of the proportional coefficient, the integral coefficient and the differential coefficient. The valve driving module (31) continuously receives the actual flow value fed back by the flow sensor. At the same time, the valve driving module (31) stores a preset flow target value, which is calculated based on the intelligent allocation algorithm and is used to guide the flow rate that needs to be achieved in the current heat storage process, and performs filtering processing on the collected actual flow value; Calculate the error e(t), which is obtained by the difference between the flow target value and the actual flow value, set the proportional link P, integral link I and differential link D, the size of the proportional link error produces a control effect, and its output u P (t) = K P e(t);K P The function of the integral link I is to make the change of the valve opening proportional to the error. The integral link I is used to eliminate the steady-state error of the valve drive module (31). The differential link D adjusts the control signal according to the rate of change of the error, and its output The PID control algorithm adds the outputs of the three links of proportion, integration and differentiation to obtain the final control signal. The valve drive module (31) adjusts the opening of the valve according to the size and direction of the control signal.

Images