Boiler model dynamic coupling method and system based on Internet of Things

Abstract

The invention discloses a boiler model dynamic coupling method and system based on the Internet of Things, and belongs to the technical field of boiler monitoring, and the dynamic coupling method comprises the following specific steps: analyzing the interface requirements of CFD model output and one-dimensional mechanism model input, and establishing a unified data format standard, a communication protocol and the like. Error accumulation caused by mismatching of input and output of the model can be avoided, the supporting capacity for high-frequency transmission and multi-dimensional data is improved, the operation efficiency of the multi-level coupling model is improved, the change of the operation state of the boiler can be responded in real time, the workload of developers in interface analysis and protocol design is reduced, and the development cost is reduced. It is ensured that the dynamic coupling model keeps precision and stability in long-time operation, prediction misalignment caused by environment disturbance or boiler aging is avoided, coupling unbalance caused by excessive performance optimization of a single-level model is avoided, the economical efficiency and safety of boiler operation are improved, the workload of manual parameter adjustment is reduced, and the method is suitable for large-scale popularization and application. And the automation level of industrial boiler management is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of boiler monitoring, and in particular to a boiler model dynamic coupling method and system based on the Internet of Things. Background Art

[0002] Industrial boilers play a vital role in energy production and chemical industry, and their operating efficiency and safety directly affect the production cost and environmental impact of enterprises. With the rapid development of industrial intelligence and automation technology, boilers, as important equipment in industrial production and energy supply, are facing many challenges such as efficient operation, pollution control, and intelligent monitoring. With the increase of complexity and dynamic change characteristics of parameters (such as temperature, pressure, combustion efficiency, etc.) during boiler operation, traditional control methods based on experience and single models have gradually become insufficient. On the one hand, CFD models can accurately capture the key fluid dynamics characteristics of boiler combustion due to their high precision and fine-grained simulation capabilities; on the other hand, one-dimensional mechanism models have unique advantages in system-level dynamic simulation due to their high computational efficiency and strong physical interpretability. However, it is difficult to balance accuracy and real-time performance when using any model alone; therefore, it is particularly important to invent a dynamic coupling method for boiler models based on the Internet of Things.

[0003] The existing dynamic coupling method for boiler models has the problem of error accumulation due to the mismatch between model input and output, and has poor support for high-frequency transmission and multi-dimensional data, which reduces the operating efficiency of multi-level coupling models and places a heavy workload on developers in interface analysis and protocol design. In addition, the existing dynamic coupling method for boiler models cannot guarantee the accuracy and stability of the dynamic coupling model during long-term operation, and is prone to prediction inaccuracies due to environmental disturbances or boiler aging, reducing the economy and safety of boiler operation. Summary of the invention

[0004] The purpose of the present invention is to solve the defects in the prior art and provide a boiler model dynamic coupling method and system based on the Internet of Things.

[0005] In order to achieve the above object, the present invention adopts the following technical solution: The dynamic coupling method of boiler model based on the Internet of Things includes: 1) Analyze the interface requirements for CFD model output and one-dimensional mechanism model input, and establish a unified data format standard and communication protocol; 2) Based on the unified data format standard and communication protocol, collect the matching data of CFD model output and one-dimensional mechanism model input, and train the error analysis model based on the matching data; 3) The output data of the CFD model is transferred to the error analysis model, and the CFD output data is dynamically adjusted according to the deviation information provided by the error analysis model, and the corrected data is input into the one-dimensional mechanism model to obtain the output data of the one-dimensional mechanism model; 4) Based on the output data of the two one-dimensional mechanism models, the fine-grained combustion simulation tasks handled by the CFD model and the macroscopic heat transfer and flow characteristics tasks handled by the one-dimensional mechanism model are set, and a multi-level coupling model of CFD and one-dimensional mechanism model is constructed through high- and low-level functional division of labor; 5) Collect historical data generated during the operation of the multi-level coupling model and optimize the parameters of the multi-level coupling model. At the same time, regularly evaluate the prediction accuracy of the multi-level coupling model and update the parameters; 6) Based on the prediction output of the multi-level coupling model after parameter update, dynamically adjust the boiler operating parameters, analyze the operating data, detect potential fault signs, generate early warning information and recommend optimization strategies when abnormalities occur.

[0006] A further improvement of the present invention is that the specific steps of establishing a unified data format standard and communication protocol in step 1) are as follows: Q1.1: Collect the data types of CDF model output and the data types required for one-dimensional mechanism model input, and build the feature set of CFD model output data and the one-dimensional mechanism model input parameter set based on the collected data sets, and detect the correlation data between the two sets to generate the corresponding key data set; Q1.2: Generate candidate sets of all available data formats and communication protocols based on the key data set, and construct a decision graph based on the candidate sets of data formats and communication protocols, where the nodes in the decision graph represent candidate combinations of formats and protocols. Randomly generate multiple groups of search entities, initialize the initial node position of each search entity on the decision graph, set the initial pheromone concentration of each node, and calculate the transfer probability of each search entity based on the pheromone concentration and combination matching degree of the node where each search entity is currently located and all unvisited nodes; Q1.3: Each search agent randomly selects a node according to the transition probability to select the next path. The search agent visits the nodes in the graph in turn until a complete path is constructed to generate a complete candidate solution. The total fitness of each path is calculated, and the pheromone concentration of the corresponding node is updated according to the total fitness of each path. The path construction is repeated according to the pheromone update until the maximum number of iterations or the pheromone change value reaches the preset threshold; Q1.4: Collect the constructed complete paths and calculate the total fitness of each path. Select the path with the highest pheromone concentration as the final data format and communication protocol combination. Apply the selected data format and communication protocol to simulate the data transmission from CFD output to one-dimensional mechanism model input, and evaluate the data transmission delay, reliability and error.

[0007] A further improvement of the present invention is that the specific steps of training the error analysis model based on matching data in step 2) are as follows: Q2.1: Preprocess the collected CFD model output data and one-dimensional mechanism model input data, and extract the key features of the output data and input data through feature dimensionality reduction. At the same time, record the actual measurement or experimental data as the reference truth value. Based on the extracted key features of the output data, key features of the input data and the reference truth value, obtain the deviation between the CFD output and the one-dimensional model input, and construct the model data set; Q2.2: Divide the model data set into a training set and a validation set, set the input layer and output layer of the error analysis model, add a hidden layer, use the ReLU nonlinear function as the hidden layer activation function, and then initialize the weights and biases of the error analysis model; Q2.3: Input the training set data into the error analysis model, and calculate through the linear transformation and activation function of each hidden layer in the model, and finally output the feature transmission value to the output layer, and calculate the error prediction value, and calculate the loss value of the deviation between the error prediction value and the true error through the mean square error loss function; Q2.4: Calculate the gradient of the loss value with respect to each layer parameter in the error analysis model according to the chain rule, propagate the gradient from the output layer to the input layer layer by layer, and update the weights and biases through the gradient descent algorithm. Repeat multiple rounds of training until the loss function value converges to the preset range. Calculate the prediction error of the trained model through the validation set and evaluate the performance indicators. If the performance indicators do not meet the preset standards, adjust the learning rate and batch size hyperparameters, and deploy the final error analysis model to the actual detection platform.

[0008] A further improvement of the present invention is that the specific steps of constructing a multi-level coupling model of CFD and a one-dimensional mechanism model through high-level and low-level functional division in step 4) are as follows: Q3.1: Based on the functional information of the CFD model and the one-dimensional mechanism model, determine the functional positioning and hierarchical division of the CFD model and the one-dimensional mechanism model in the overall coupling model, unify the formats of various types of data of the two groups of models based on a unified data format, and set up the communication interface between the two groups of models according to the selected communication protocol; Q3.2: Divide the CFD model into high-level levels to provide detailed local information, and the one-dimensional mechanism model into low-level levels to handle the overall thermodynamic dynamics. Then, convert the CFD local output into the one-dimensional mechanism model input through an aggregation function, and feed the global results of the one-dimensional mechanism model back to the CFD model as the boundary conditions or initial value corrections of the CFD model to form a multi-level coupling model with data interaction between levels; Q3.3: Set the synchronization time step of the multi-level coupling model manually or automatically to make the data update frequency of the CFD and one-dimensional models consistent. Then build the communication between the error analysis model and the multi-level coupling model, and use the error analysis model to predict the deviation between the CFD output and the one-dimensional model input, and correct the CFD output in real time.

[0009] A further improvement of the present invention is that the specific steps of regularly evaluating the prediction accuracy of the multi-level coupling model and updating the parameters in step 5) are as follows: Q4.1: Collect the output results of the multi-level coupling model in real time, and evaluate the deviation between the model output and the actual data through the mean square error. If the prediction error is higher than the preset threshold, the model parameters need to be updated and a population consisting of multiple model parameter vectors is randomly initialized; Q4.2: Set up multiple groups of individuals. The current position of each group of individuals represents a candidate model parameter vector in the population. Use a multi-level coupling model to predict each group of model parameter vectors and calculate the fitness value. The model parameter vector with the lowest fitness value is the optimal individual. Set up a set of coefficient vectors and generate a random number between 0 and 1 in each round of iteration. Q4.3: If the generated random number is greater than the preset probability parameter, the coefficient and step size for controlling the position update are calculated based on the coefficient vector, and the remaining individuals are judged according to the size of the direction to determine whether they are shrinking around the optimal solution. If the coefficient is less than 1, it means that the remaining individuals are surrounding the optimal individual, and the positions of the remaining individuals are updated according to the optimal individual, that is, the parameter values ​​of each individual are adjusted. If the coefficient is ≥1, it means that the remaining individuals are far away from the optimal individual. Through random search, random positions are selected in the population space, and the positions of the remaining individuals are updated. After each iteration, the coefficient vector is updated based on the linear decreasing rule. Q4.4: If the generated random number is less than or equal to the preset probability parameter, the distance between the positions of the remaining individuals in the population and the position of the optimal individual is calculated, and the spiral motion formula is used to simulate the movement law of each individual along the spiral trajectory to approach the optimal individual, and the individual position is updated; Q4.5: Repeatedly select the best individual and iteratively update the position until the fitness value change converges to the preset threshold range. Then compare the fitness values ​​of each group of individuals and output the model parameter vector with the lowest fitness value as the optimal parameter configuration of the multi-level coupling model, and replace the original parameter configuration. At the same time, monitor the subsequent operation status of the multi-level coupling model in real time to perform real-time parameter updates.

[0010] A further improvement of the present invention is that the specific steps of detecting potential fault signs in step 6) are as follows: Q5.1: Run the multi-level coupling model to predict the output parameters of the boiler and calculate the deviation between the predicted output and the target value. At the same time, adjust the boiler operating parameters according to the deviation. Collect the real-time operating parameters of the boiler from the sensor and calculate the error between the model prediction and the actual operation. Collect the operating states of the boiler and the operating adjustment actions to construct the state space and action space of the boiler operation. S5.2: Update the current boiler state according to the real-time operation data, and calculate the state transition probability of the current state transferring to the next state under different adjustment actions based on the state space and action space of the boiler, and construct the corresponding state transition probability matrix based on the calculated probability; S5.3: Based on the historical state and state transition probability, the probability of potential fault state is accumulated. If the fault probability is higher than the threshold, a fault warning is triggered and the fault type is analyzed. If the fault is a medium or high risk fault, the boiler operation is terminated.

[0011] A further improvement of the present invention is that it also includes: 7) Design multi-scenario experiments to collect boiler operation data, compare the experimental results with the prediction results of the multi-level coupling model, calculate the error index, and optimize the model parameters based on the error analysis results.

[0012] The boiler model dynamic coupling system based on the Internet of Things includes: Model analysis module, which analyzes the interface requirements of CFD model output and one-dimensional mechanism model input, and establishes a unified data format standard and communication protocol; The error analysis model training module collects the matching data of CFD model output and one-dimensional mechanism model input based on the unified data format standard and communication protocol, and trains the error analysis model based on the matching data; The data processing module transfers the output data of the CFD model to the error analysis model, dynamically adjusts the CFD output data according to the deviation information provided by the error analysis model, and inputs the corrected data into the one-dimensional mechanism model to obtain the output data of the one-dimensional mechanism model; The multi-level coupling model construction module sets the fine-grained combustion simulation tasks that the CFD model is responsible for and the macroscopic heat transfer and flow characteristics tasks that the one-dimensional mechanism model handles based on the output data of two sets of one-dimensional mechanism models, and constructs a multi-level coupling model of CFD and one-dimensional mechanism models through high- and low-level functional division of labor; Optimize and update the module, collect historical data generated during the operation of the multi-level coupling model, optimize the parameters of the multi-level coupling model, and regularly evaluate the prediction accuracy of the multi-level coupling model and update the parameters; The detection and early warning module dynamically adjusts boiler operating parameters based on the predicted output of the multi-level coupling model after parameter update, analyzes operating data, detects potential fault signs, generates early warning information when anomalies occur, and recommends optimization strategies.

[0013] A further improvement of the present invention is that, in the model analysis module, the specific steps of establishing a unified data format standard and communication protocol are as follows: Q1.1: Collect the data types of CDF model output and the data types required for one-dimensional mechanism model input, and build the feature set of CFD model output data and the one-dimensional mechanism model input parameter set based on the collected data sets, and detect the correlation data between the two sets to generate the corresponding key data set; Q1.2: Generate candidate sets of all available data formats and communication protocols based on the key data set, and construct a decision graph based on the candidate sets of data formats and communication protocols, where the nodes in the decision graph represent candidate combinations of formats and protocols. Randomly generate multiple groups of search entities, initialize the initial node position of each search entity on the decision graph, set the initial pheromone concentration of each node, and calculate the transfer probability of each search entity based on the pheromone concentration and combination matching degree of the node where each search entity is currently located and all unvisited nodes; Q1.3: Each search agent randomly selects a node according to the transition probability to select the next path. The search agent visits the nodes in the graph in turn until a complete path is constructed to generate a complete candidate solution. The total fitness of each path is calculated, and the pheromone concentration of the corresponding node is updated according to the total fitness of each path. The path construction is repeated according to the pheromone update until the maximum number of iterations or the pheromone change value reaches the preset threshold; Q1.4: Collect the constructed complete paths and calculate the total fitness of each path. Select the path with the highest pheromone concentration as the final data format and communication protocol combination. Apply the selected data format and communication protocol to simulate the data transmission from CFD output to one-dimensional mechanism model input, and evaluate the data transmission delay, reliability and error.

[0014] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the boiler model dynamic coupling method based on the Internet of Things.

[0015] Compared with the prior art, the present invention has at least the following beneficial technical effects: The boiler model dynamic coupling method and system based on the Internet of Things provided by the present invention collects the data type output by the CDF model and the data type required by the one-dimensional mechanism model input, generates candidate sets of all available data formats and communication protocols, and constructs a decision graph based on the data format candidate set and the communication protocol candidate set, randomly generates multiple groups of search bodies, and initializes the initial node position of each search body on the decision graph, sets the initial pheromone concentration of each node, and each search body calculates the transition probability according to the pheromone concentration and combined matching degree of the node where each search body is currently located and all unvisited nodes, and each search body randomly selects a node according to the transition probability to select the next step path, and the search body sequentially visits the nodes in the graph until a complete path is constructed to generate a complete candidate solution, and calculates the total number of each path. Fitness, and update the pheromone concentration of the corresponding node according to the total fitness of each path, repeat the path construction according to the pheromone update, until the maximum number of iterations or the pheromone change value reaches the preset threshold, collect the constructed groups of complete paths, and calculate the total fitness of each group of paths, select the path with the highest pheromone concentration as the final data format and communication protocol combination, apply the selected data format and communication protocol, simulate the data transmission from CFD output to one-dimensional mechanism model input, evaluate data transmission delay, reliability and error, avoid error accumulation caused by mismatch between model input and output, improve the support capability for high-frequency transmission and multi-dimensional data, improve the operation efficiency of multi-level coupling models, respond to changes in boiler operating status in real time, and reduce the workload of developers in interface analysis and protocol design.

[0016] The present invention evaluates the deviation between the model output and the actual data through the mean square error. If the prediction error is higher than the preset threshold, the model parameters need to be updated, a population composed of multiple model parameter vectors is randomly initialized, multiple groups of individuals are set, and the current position of each group of individuals represents a candidate model parameter vector in the population. A multi-level coupling model is used to predict each group of model parameter vectors, and the fitness value is calculated. The model parameter vector with the lowest fitness value is taken as the optimal individual, and a group of coefficient vectors is set. In each round of iteration, a random number between 0 and 1 is generated. If the generated random number is greater than the preset probability parameter, the coefficient and step length for controlling the position update are calculated based on the coefficient vector, and the individual is shrunk and surrounded based on the generated coefficient and step length. After each iteration, the coefficient vector is updated based on the linear decreasing rule. If the generated random number is less than or equal to the preset probability parameter, the remaining individuals in the population are calculated. The distance between the individual position and the optimal individual position is calculated, and the spiral motion formula is used to simulate the motion law of each individual along the spiral trajectory close to the optimal individual, update the individual position, repeatedly select the optimal individual and iteratively update the position until the fitness value change value converges to the preset threshold range, and then compare the fitness values ​​of each group of individuals, and output the model parameter vector with the lowest fitness value as the optimal parameter configuration of the multi-level coupling model, and replace the original parameter configuration. At the same time, the subsequent operating status of the multi-level coupling model is monitored in real time to perform real-time parameter updates to ensure that the dynamic coupling model maintains accuracy and stability during long-term operation, avoids prediction inaccuracies caused by environmental disturbances or boiler aging, and avoids coupling imbalance caused by over-optimization of single-level model performance, improves the economy and safety of boiler operation, reduces the workload of manual parameter adjustment, and improves the automation level of industrial boiler management. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 This is a flow chart of the boiler model dynamic coupling method based on the Internet of Things proposed in the present invention.

[0019] Figure 2 This is a structural block diagram of the boiler model dynamic coupling system based on the Internet of Things proposed in the present invention. DETAILED DESCRIPTION

[0020] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0021] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0022] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an" and "the" are intended to include plural forms unless the context clearly indicates otherwise.

[0023] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0024] Various structural schematic diagrams of the embodiments disclosed in the present invention are shown in the accompanying drawings. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0025] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0026] Example 1 like Figure 1 As shown, the boiler model dynamic coupling method based on the Internet of Things provided by the present invention includes: 1) Analyze the interface requirements for CFD model output and one-dimensional mechanism model input, and establish a unified data format standard and communication protocol; 2) Based on the unified data format standard and communication protocol, collect the matching data of CFD model output and one-dimensional mechanism model input, and train the error analysis model based on the matching data; 3) The output data of the CFD model is transferred to the error analysis model, and the CFD output data is dynamically adjusted according to the deviation information provided by the error analysis model, and the corrected data is input into the one-dimensional mechanism model to obtain the output data of the one-dimensional mechanism model; 4) Based on the output data of the two one-dimensional mechanism models, the fine-grained combustion simulation tasks handled by the CFD model and the macroscopic heat transfer and flow characteristics tasks handled by the one-dimensional mechanism model are set, and a multi-level coupling model of CFD and one-dimensional mechanism model is constructed through high- and low-level functional division of labor; 5) Collect historical data generated during the operation of the multi-level coupling model and optimize the parameters of the multi-level coupling model. At the same time, regularly evaluate the prediction accuracy of the multi-level coupling model and update the parameters; 6) Based on the prediction output of the multi-level coupling model after parameter update, dynamically adjust the boiler operating parameters, analyze the operating data, detect potential fault signs, generate early warning information and recommend optimization strategies when abnormalities occur.

[0027] Example 2 like Figure 1 As shown, the boiler model dynamic coupling method based on the Internet of Things provided by the present invention has the following specific steps: Analyze the interface requirements for CFD model output and one-dimensional mechanism model input, and establish unified data format standards and communication protocols.

[0028] Specifically, the data types output by the CDF model and the data types required for the input of the one-dimensional mechanism model are collected, and the feature set of the CFD model output data is constructed based on the collected data sets, and the one-dimensional mechanism model input parameter set, and the associated data between the two sets are detected to generate a corresponding key data set, and a candidate set of all available data formats and communication protocols is generated according to the key data set, and a decision graph is constructed based on the data format candidate set and the communication protocol candidate set, wherein the nodes in the decision graph represent candidate combinations of formats and protocols, and multiple groups of search bodies are randomly generated, and the initial node position of each search body on the decision graph is initialized, and the initial pheromone concentration of each node is set, and each search body is based on the pheromone concentration of the node where each search body is currently located and all unvisited nodes. The transfer probability is calculated based on the combination matching degree. Each search body randomly selects a node to select the next path according to the transfer probability. The search body visits the nodes in the graph in turn until a complete path is constructed to generate a complete candidate solution. The total fitness of each path is calculated, and the pheromone concentration of the corresponding node is updated according to the total fitness of each path. The path construction is repeated according to the pheromone update until the maximum number of iterations or the pheromone change value reaches the preset threshold. The constructed groups of complete paths are collected, and the total fitness of each group of paths is calculated. The path with the highest pheromone concentration is selected as the final data format and communication protocol combination. The selected data format and communication protocol are applied to simulate the data transmission from CFD output to one-dimensional mechanism model input, and the data transmission delay, reliability and error are evaluated.

[0029] Based on the unified data format standard and communication protocol, the matching data of CFD model output and one-dimensional mechanism model input are collected, and the error analysis model is trained based on the matching data.

[0030] Specifically, the collected CFD model output data and one-dimensional mechanism model input data are preprocessed, and the key features of the output data and input data are extracted through feature dimensionality reduction. At the same time, the actual measurement or experimental data is recorded as the reference truth value. Based on the extracted key features of the output data, the key features of the input data and the reference truth value, the deviation between the CFD output and the one-dimensional model input is obtained, and a model data set is constructed. The model data set is divided into a training set and a validation set, and the error analysis model input layer and output layer are set, and a hidden layer is added. The ReLU nonlinear function is used as the hidden layer activation function, and then the weights and bias of the error analysis model are initialized, and the training set data is input into the error analysis model, and the model is used to verify the error analysis model. The linear transformation and activation function of each hidden layer in the model are calculated, and the final output feature transmission value is transferred to the output layer, and the error prediction value is calculated. The loss value of the deviation between the error prediction value and the true error is calculated by the mean square error loss function. The gradient of the loss value to each layer parameter in the error analysis model is calculated according to the chain rule. The gradient is propagated from the output layer to the input layer layer by layer, and the weights and biases are updated by the gradient descent algorithm. Multiple rounds of training are repeated until the loss function value converges to the preset range, and then the prediction error of the trained model is calculated by the validation set, and the performance indicators are evaluated. If the performance indicators do not meet the preset standards, the learning rate and batch size hyperparameters are adjusted, and the final error analysis model is deployed to the actual detection platform.

[0031] The output data of the CFD model is passed into the error analysis model. According to the deviation information provided by the error analysis model, the CFD output data is dynamically adjusted, and the corrected data is input into the one-dimensional mechanism model to obtain the output data of the one-dimensional mechanism model.

[0032] Example 3 like Figure 1 As shown, the boiler model dynamic coupling method based on the Internet of Things provided by the present invention has the following specific steps: Based on the two sets of output data, the fine-grained combustion simulation tasks handled by the CFD model and the macroscopic heat transfer and flow characteristics tasks handled by the one-dimensional mechanism model are set up, and a multi-level coupling model of CFD and one-dimensional mechanism model is constructed through high- and low-level functional division of labor.

[0033] Specifically, according to the functional information of the CFD model and the one-dimensional mechanism model, the functional positioning and hierarchical division of the CFD model and the one-dimensional mechanism model in the overall coupling model are determined, and the formats of various types of data of the two groups of models are unified based on a unified data format, and the communication interface between the two groups of models is set according to the selected communication protocol, the CFD model is divided into a high-level level to provide detailed local information, and the one-dimensional mechanism model is divided into a low-level level to handle the overall thermodynamic dynamics, and then the CFD local output is converted into a one-dimensional mechanism model input through an aggregation function, and the global result of the one-dimensional mechanism model is fed back to the CFD model as a boundary condition or initial value correction of the CFD model to form a multi-level coupling model for data interaction between levels, and the synchronization time step of the multi-level coupling model is set according to manual settings or automatic methods to make the data update frequency of the CFD and one-dimensional models consistent, and then the communication between the error analysis model and the multi-level coupling model is constructed, and the error analysis model is used to predict the deviation between the CFD output and the one-dimensional model input, and the CFD output is corrected in real time.

[0034] Collect historical data generated during the operation of the multi-level coupling model, optimize the parameters of the multi-level coupling model, and regularly evaluate the prediction accuracy of the multi-level coupling model and update the parameters.

[0035] Specifically, the output results of the multi-level coupling model are collected in real time, and the deviation between the model output and the actual data is evaluated by the mean square error. If the prediction error is higher than the preset threshold, the model parameters need to be updated, and a population composed of multiple model parameter vectors is randomly initialized. Multiple groups of individuals are set, and the current position of each group of individuals represents a candidate model parameter vector in the population. The multi-level coupling model is used to predict each group of model parameter vectors, and the fitness value is calculated. The model parameter vector with the lowest fitness value is taken as the optimal individual. A set of coefficient vectors is set. In each round of iteration, a random number between 0 and 1 is generated. If the generated random number is greater than the preset probability parameter, the coefficient and step size for controlling the position update are calculated based on the coefficient vector. According to the size of the direction, it is determined whether the remaining individuals are shrinking around the optimal solution. If the coefficient is less than 1, it means that the remaining individuals are surrounding the optimal individual, and the positions of the remaining individuals are updated according to the optimal individual. New, that is, adjust the parameter value of each individual. If the coefficient is ≥1, it means that the remaining individuals are far away from the optimal individual. Through random search, select random positions in the population space, update the positions of the remaining individuals, and update the coefficient vector based on the linear decreasing rule after each iteration. If the generated random number is less than or equal to the preset probability parameter, calculate the distance between the positions of the remaining individuals in the population and the position of the optimal individual, and simulate the movement law of each individual approaching the optimal individual along the spiral trajectory through the spiral motion formula, update the individual position, repeatedly select the optimal individual and iteratively update the position until the fitness value change value converges to the preset threshold range, then compare the fitness values ​​of each group of individuals, and output the model parameter vector with the lowest fitness value as the optimal parameter configuration of the multi-level coupling model, and replace the original parameter configuration, and monitor the subsequent operation status of the multi-level coupling model in real time to perform real-time parameter updates.

[0036] Based on the prediction output of the multi-level coupling model after parameter update, the boiler operating parameters are dynamically adjusted, the operating data is analyzed, potential fault signs are detected, and early warning information is generated and optimization strategies are recommended when abnormalities occur.

[0037] Specifically, a multi-level coupling model is run to predict the output parameters of the boiler, and the deviation between the predicted output and the target value is calculated. At the same time, the boiler operating parameters are adjusted according to the deviation. The real-time operating parameters of the boiler are collected from the sensor, and the error between the model prediction and the actual operation is calculated. The operating states of the boiler and the operating adjustment actions are collected to construct the state space and action space of the boiler operation. The current boiler state is updated according to the real-time operating data, and based on the state space and action space of the boiler, the state transition probability of the current state transferring to the next state under different adjustment actions is calculated, and the corresponding state transition probability matrix is ​​constructed based on the calculated probability. According to the historical state and state transition probability, the probability of potential fault state is accumulated. If the fault probability is higher than the threshold, a fault warning is triggered, and the fault type is analyzed at the same time. If the fault is a medium or high risk fault, the boiler operation is terminated.

[0038] Example 4 like Figure 2 As shown, the boiler model dynamic coupling system based on the Internet of Things provided by the present invention includes: Model analysis module, which analyzes the interface requirements of CFD model output and one-dimensional mechanism model input, and establishes a unified data format standard and communication protocol; The error analysis model training module collects the matching data of CFD model output and one-dimensional mechanism model input based on the unified data format standard and communication protocol, and trains the error analysis model based on the matching data; The data processing module transfers the output data of the CFD model to the error analysis model, dynamically adjusts the CFD output data according to the deviation information provided by the error analysis model, and inputs the corrected data into the one-dimensional mechanism model to obtain the output data of the one-dimensional mechanism model; The multi-level coupling model construction module sets the fine-grained combustion simulation tasks that the CFD model is responsible for and the macroscopic heat transfer and flow characteristics tasks that the one-dimensional mechanism model handles based on the output data of two sets of one-dimensional mechanism models, and constructs a multi-level coupling model of CFD and one-dimensional mechanism models through high- and low-level functional division of labor; Optimize and update the module, collect historical data generated during the operation of the multi-level coupling model, optimize the parameters of the multi-level coupling model, and regularly evaluate the prediction accuracy of the multi-level coupling model and update the parameters; The detection and early warning module dynamically adjusts boiler operating parameters based on the predicted output of the multi-level coupling model after parameter update, analyzes operating data, detects potential fault signs, generates early warning information when anomalies occur, and recommends optimization strategies.

[0039] Example 5 The present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the boiler model dynamic coupling method based on the Internet of Things are implemented.

[0040] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0041] The present application is described with reference to the flowcharts and / or block diagrams of the methods, systems and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A system that specifies the functions of a box or multiple boxes.

[0042] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0043] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0044] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

[0045] In addition, it should be understood that although this specification is described in accordance with the implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation modes that can be understood by those skilled in the art. The above content is only to illustrate the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A boiler model dynamic coupling method based on the Internet of Things, characterized in that: include: 1) Analyze the interface requirements for CFD model output and one-dimensional mechanism model input, and establish a unified data format standard and communication protocol; 2) Based on the unified data format standard and communication protocol, collect the matching data of CFD model output and one-dimensional mechanism model input, and train the error analysis model based on the matching data; 3) The output data of the CFD model is transferred to the error analysis model, and the CFD output data is dynamically adjusted according to the deviation information provided by the error analysis model, and the corrected data is input into the one-dimensional mechanism model to obtain the output data of the one-dimensional mechanism model; 4) Based on the output data of the two one-dimensional mechanism models, the fine-grained combustion simulation tasks handled by the CFD model and the macroscopic heat transfer and flow characteristics tasks handled by the one-dimensional mechanism model are set, and a multi-level coupling model of CFD and one-dimensional mechanism model is constructed through high- and low-level functional division of labor; 5) Collect historical data generated during the operation of the multi-level coupling model and optimize the parameters of the multi-level coupling model. At the same time, regularly evaluate the prediction accuracy of the multi-level coupling model and update the parameters; 6) Based on the prediction output of the multi-level coupling model after parameter update, dynamically adjust the boiler operating parameters, analyze the operating data, detect potential fault signs, generate early warning information and recommend optimization strategies when abnormalities occur.

2. The method for dynamic coupling of boiler model based on Internet of Things according to claim 1 is characterized in that: The specific steps for establishing a unified data format standard and communication protocol described in step 1) are as follows: Q1.1: Collect the data types of CDF model output and the data types required for one-dimensional mechanism model input, and build the feature set of CFD model output data and the one-dimensional mechanism model input parameter set based on the collected data sets, and detect the correlation data between the two sets to generate the corresponding key data set; Q1.2: Generate candidate sets of all available data formats and communication protocols based on the key data set, and construct a decision graph based on the candidate sets of data formats and communication protocols, where the nodes in the decision graph represent candidate combinations of formats and protocols. Randomly generate multiple groups of search entities, initialize the initial node position of each search entity on the decision graph, set the initial pheromone concentration of each node, and calculate the transfer probability of each search entity based on the pheromone concentration and combination matching degree of the node where each search entity is currently located and all unvisited nodes; Q1.3: Each search agent randomly selects a node according to the transition probability to select the next path. The search agent visits the nodes in the graph in turn until a complete path is constructed to generate a complete candidate solution. The total fitness of each path is calculated, and the pheromone concentration of the corresponding node is updated according to the total fitness of each path. The path construction is repeated according to the pheromone update until the maximum number of iterations or the pheromone change value reaches the preset threshold; Q1.4: Collect the constructed complete paths and calculate the total fitness of each path. Select the path with the highest pheromone concentration as the final data format and communication protocol combination. Apply the selected data format and communication protocol to simulate the data transmission from CFD output to one-dimensional mechanism model input, and evaluate the data transmission delay, reliability and error.

3. The method for dynamic coupling of boiler model based on Internet of Things according to claim 2 is characterized in that: The specific steps of training the error analysis model based on matching data in step 2) are as follows: Q2.1: Preprocess the collected CFD model output data and one-dimensional mechanism model input data, and extract the key features of the output data and input data through feature dimensionality reduction. At the same time, record the actual measurement or experimental data as the reference truth value. Based on the extracted key features of the output data, key features of the input data and the reference truth value, obtain the deviation between the CFD output and the one-dimensional model input, and construct the model data set; Q2.2: Divide the model data set into a training set and a validation set, set the input layer and output layer of the error analysis model, add a hidden layer, use the ReLU nonlinear function as the hidden layer activation function, and then initialize the weights and biases of the error analysis model; Q2.3: Input the training set data into the error analysis model, and calculate through the linear transformation and activation function of each hidden layer in the model, and finally output the feature transmission value to the output layer, and calculate the error prediction value, and calculate the loss value of the deviation between the error prediction value and the true error through the mean square error loss function; Q2.4: Calculate the gradient of the loss value with respect to each layer parameter in the error analysis model according to the chain rule, propagate the gradient from the output layer to the input layer layer by layer, and update the weights and biases through the gradient descent algorithm. Repeat multiple rounds of training until the loss function value converges to the preset range. Calculate the prediction error of the trained model through the validation set and evaluate the performance indicators. If the performance indicators do not meet the preset standards, adjust the learning rate and batch size hyperparameters, and deploy the final error analysis model to the actual detection platform.

4. The method for dynamic coupling of boiler model based on Internet of Things according to claim 3 is characterized in that: The specific steps of constructing a multi-level coupling model of CFD and one-dimensional mechanism model through high-level and low-level functional division as described in step 4) are as follows: Q3.1: Based on the functional information of the CFD model and the one-dimensional mechanism model, determine the functional positioning and hierarchical division of the CFD model and the one-dimensional mechanism model in the overall coupling model, unify the formats of various types of data of the two groups of models based on a unified data format, and set up the communication interface between the two groups of models according to the selected communication protocol; Q3.2: Divide the CFD model into high-level levels to provide detailed local information, and the one-dimensional mechanism model into low-level levels to handle the overall thermodynamic dynamics. Then, convert the CFD local output into the one-dimensional mechanism model input through an aggregation function, and feed the global results of the one-dimensional mechanism model back to the CFD model as the boundary conditions or initial value corrections of the CFD model to form a multi-level coupling model with data interaction between levels; Q3.3: Set the synchronization time step of the multi-level coupling model manually or automatically to make the data update frequency of the CFD and one-dimensional models consistent. Then build the communication between the error analysis model and the multi-level coupling model, and use the error analysis model to predict the deviation between the CFD output and the one-dimensional model input, and correct the CFD output in real time.

5. The method for dynamic coupling of boiler model based on Internet of Things according to claim 1 is characterized in that: The specific steps for regularly evaluating the prediction accuracy of the multi-level coupling model and updating the parameters described in step 5) are as follows: Q4.1: Collect the output results of the multi-level coupling model in real time, and evaluate the deviation between the model output and the actual data through the mean square error. If the prediction error is higher than the preset threshold, the model parameters need to be updated and a population consisting of multiple model parameter vectors is randomly initialized; Q4.2: Set up multiple groups of individuals. The current position of each group of individuals represents a candidate model parameter vector in the population. Use a multi-level coupling model to predict each group of model parameter vectors and calculate the fitness value. The model parameter vector with the lowest fitness value is the optimal individual. Set up a set of coefficient vectors and generate a random number between 0 and 1 in each round of iteration. Q4.3: If the generated random number is greater than the preset probability parameter, the coefficient and step size for controlling the position update are calculated based on the coefficient vector, and the remaining individuals are judged according to the size of the direction to determine whether they are shrinking around the optimal solution. If the coefficient is less than 1, it means that the remaining individuals are surrounding the optimal individual, and the positions of the remaining individuals are updated according to the optimal individual, that is, the parameter values ​​of each individual are adjusted. If the coefficient is ≥1, it means that the remaining individuals are far away from the optimal individual. Through random search, random positions are selected in the population space, and the positions of the remaining individuals are updated. After each iteration, the coefficient vector is updated based on the linear decreasing rule. Q4.4: If the generated random number is less than or equal to the preset probability parameter, the distance between the positions of the remaining individuals in the population and the position of the optimal individual is calculated, and the spiral motion formula is used to simulate the movement law of each individual along the spiral trajectory to approach the optimal individual, and the individual position is updated; Q4.5: Repeatedly select the best individual and iteratively update the position until the fitness value change converges to the preset threshold range. Then compare the fitness values ​​of each group of individuals and output the model parameter vector with the lowest fitness value as the optimal parameter configuration of the multi-level coupling model, and replace the original parameter configuration. At the same time, monitor the subsequent operation status of the multi-level coupling model in real time to perform real-time parameter updates.

6. The method for dynamic coupling of boiler model based on Internet of Things according to claim 5 is characterized in that: The specific steps for detecting potential fault signs in step 6) are as follows: Q5.1: Run the multi-level coupling model to predict the output parameters of the boiler and calculate the deviation between the predicted output and the target value. At the same time, adjust the boiler operating parameters according to the deviation. Collect the real-time operating parameters of the boiler from the sensor and calculate the error between the model prediction and the actual operation. Collect the operating states of the boiler and the operating adjustment actions to construct the state space and action space of the boiler operation. S5.2: Update the current boiler state according to the real-time operation data, and calculate the state transition probability of the current state transferring to the next state under different adjustment actions based on the state space and action space of the boiler, and construct the corresponding state transition probability matrix based on the calculated probability; S5.3: Based on the historical state and state transition probability, the probability of potential fault state is accumulated. If the fault probability is higher than the threshold, a fault warning is triggered and the fault type is analyzed. If the fault is a medium or high risk fault, the boiler operation is terminated.

7. The method for dynamic coupling of boiler model based on Internet of Things according to claim 1 is characterized in that: Also includes: 7) Design multi-scenario experiments to collect boiler operation data, compare the experimental results with the prediction results of the multi-level coupling model, calculate the error index, and optimize the model parameters based on the error analysis results.

8. The boiler model dynamic coupling system based on the Internet of Things is characterized by: include: Model analysis module, which analyzes the interface requirements of CFD model output and one-dimensional mechanism model input, and establishes a unified data format standard and communication protocol; The error analysis model training module collects the matching data of CFD model output and one-dimensional mechanism model input based on the unified data format standard and communication protocol, and trains the error analysis model based on the matching data; The data processing module transfers the output data of the CFD model to the error analysis model, dynamically adjusts the CFD output data according to the deviation information provided by the error analysis model, and inputs the corrected data into the one-dimensional mechanism model to obtain the output data of the one-dimensional mechanism model; The multi-level coupling model construction module sets the fine-grained combustion simulation tasks that the CFD model is responsible for and the macroscopic heat transfer and flow characteristics tasks that the one-dimensional mechanism model handles based on the output data of two sets of one-dimensional mechanism models, and constructs a multi-level coupling model of CFD and one-dimensional mechanism models through high- and low-level functional division of labor; Optimize and update the module, collect historical data generated during the operation of the multi-level coupling model, optimize the parameters of the multi-level coupling model, and regularly evaluate the prediction accuracy of the multi-level coupling model and update the parameters; The detection and early warning module dynamically adjusts boiler operating parameters based on the predicted output of the multi-level coupling model after parameter update, analyzes operating data, detects potential fault signs, generates early warning information when anomalies occur, and recommends optimization strategies.

9. The boiler model dynamic coupling system based on the Internet of Things according to claim 8 is characterized in that: In the model analysis module, the specific steps of establishing a unified data format standard and communication protocol are as follows: Q1.1: Collect the data types of CDF model output and the data types required for one-dimensional mechanism model input, and build the feature set of CFD model output data and the one-dimensional mechanism model input parameter set based on the collected data sets, and detect the correlation data between the two sets to generate the corresponding key data set; Q1.2: Generate candidate sets of all available data formats and communication protocols based on the key data set, and construct a decision graph based on the candidate sets of data formats and communication protocols, where the nodes in the decision graph represent candidate combinations of formats and protocols. Randomly generate multiple groups of search entities, initialize the initial node position of each search entity on the decision graph, set the initial pheromone concentration of each node, and calculate the transfer probability of each search entity based on the pheromone concentration and combination matching degree of the node where each search entity is currently located and all unvisited nodes; Q1.3: Each search agent randomly selects a node according to the transition probability to select the next path. The search agent visits the nodes in the graph in turn until a complete path is constructed to generate a complete candidate solution. The total fitness of each path is calculated, and the pheromone concentration of the corresponding node is updated according to the total fitness of each path. The path construction is repeated according to the pheromone update until the maximum number of iterations or the pheromone change value reaches the preset threshold; Q1.4: Collect the constructed complete paths and calculate the total fitness of each path. Select the path with the highest pheromone concentration as the final data format and communication protocol combination. Apply the selected data format and communication protocol to simulate the data transmission from CFD output to one-dimensional mechanism model input, and evaluate the data transmission delay, reliability and error.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the boiler model dynamic coupling method based on the Internet of Things described in any one of claims 1 to 7 are implemented.

Images