Quality soft measurement method for thermal power generation fuel and related device

Abstract

The invention belongs to the technical field of thermal power generation, and discloses a thermal power generation fuel quality soft measurement method and a related device, and the method comprises the steps: obtaining boiler real-time operation index data related to the fuel quality; based on a mutual information method, performing feature screening on the boiler real-time operation index data related to the fuel quality to obtain key real-time operation index data; inputting the key real-time operation index data into a fuel quality soft measurement model, and outputting to obtain a quality soft measurement result of the thermal power generation fuel; wherein the fuel quality soft measurement model is a model based on a long short-term memory neural network; the fuel quality soft measurement model based on the long-short-term memory neural network is used for predicting the fuel quality, real-time online detection of the fuel quality is achieved, boiler operation can be rapidly responded according to changes of the fuel quality, and then the combustion efficiency and emission control of the boiler are ensured; and the stability and the flexibility of the production process are enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of thermal power generation, and particularly relates to a method for soft measurement of the quality of thermal power generation fuel and related devices. Background Art

[0002] During the operation of a thermal power plant, the quality of the fuel has an important impact on the combustion efficiency and emission control of the boiler; in particular, the calorific value of the fuel directly determines the combustion efficiency and combustion load of the boiler, affecting the economy and environmental protection emission level of the power plant; currently, thermal power plants generally monitor the fuel quality through the regular sampling and testing method; this method usually involves manual sampling, sending the samples to the laboratory for chemical analysis, and measuring key indicators such as the calorific value, ash content, and volatile matter of the fuel.

[0003] However, the above traditional fuel quality detection method has serious lag, making it impossible for the boiler operation to respond quickly according to the change of fuel quality, thus affecting the boiler combustion efficiency and emission control; specifically, due to the long detection cycle, it is impossible to achieve real-time monitoring of fuel quality, resulting in the inability to adjust the combustion process and operation parameters in a timely manner; secondly, there may be certain errors in the sampling process, and it is impossible to fully reflect the change of fuel quality; finally, the cost of laboratory testing is relatively high, and the test results are difficult to be fed back to the actual operation in a short time. Summary of the Invention

[0004] Aiming at the technical problems existing in the prior art, the present invention provides a method for soft measurement of the quality of thermal power generation fuel and related devices to solve the technical problems that the traditional fuel quality detection method has serious lag, making it impossible for the boiler operation to respond quickly according to the change of fuel quality, thus affecting the boiler combustion efficiency and emission control.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows: The present invention provides a method for soft measurement of the quality of thermal power generation fuel, including: Obtaining real-time operation index data of the boiler related to the fuel quality; Based on the mutual information method, performing feature screening on the real-time operation index data of the boiler related to the fuel quality to obtain key real-time operation index data; Inputting the key real-time operation index data into the fuel quality soft measurement model, and outputting the soft measurement result of the quality of thermal power generation fuel; wherein, the fuel quality soft measurement model is a model based on a long short-term memory neural network.

[0006] Further, the real-time operation index data of the boiler related to the fuel quality includes boiler fuel supply data, characteristic fuel flow rate, furnace temperature, oxygen concentration, flue gas temperature, and fuel air flow rate.

[0007] Furthermore, the process of feature screening for the real-time operation index data of the boiler related to fuel quality based on the mutual information method to obtain the key real-time operation index data includes: Calculate the marginal information entropy of the real-time operation index data of the boiler related to fuel quality and the fuel quality index data respectively; wherein, the fuel quality index data includes fuel calorific value, volatile matter and ash content; Calculate the joint entropy of the real-time operation index data of the boiler and the fuel index data according to the marginal information entropy of the real-time operation index data of the boiler related to fuel quality and the marginal information entropy of the fuel quality index data; Based on the mutual information theory, obtain the mutual information value between the real-time operation index data of the boiler and the fuel quality index data according to the marginal information entropy of the real-time operation index data of the boiler related to fuel quality, the marginal information entropy of the fuel quality index data and the joint entropy of the real-time operation index data of the boiler and the fuel index data; Obtain the key real-time operation index data according to the mutual information value between the real-time operation index data of the boiler and the fuel quality index data.

[0008] Furthermore, the construction process of the fuel quality soft measurement model includes: Obtain the historical operation index data of the boiler related to fuel quality; Based on the mutual information method, perform feature screening on the historical operation index data of the boiler related to fuel quality to obtain the key historical operation index data; Use the key historical operation index data as training data to train and verify a pre-constructed long short-term memory neural network to obtain a trained long short-term memory neural network; use the trained long short-term memory neural network as the fuel quality soft measurement model.

[0009] Furthermore, in the process of using the key historical operation index data as training data to train and verify a pre-constructed long short-term memory neural network to obtain a trained long short-term memory neural network, the seagull optimization algorithm is used to train and optimize the preset hyperparameters of the pre-constructed long short-term memory neural network.

[0010] Furthermore, the preset hyperparameters of the pre-constructed long short-term memory neural network include learning rate, number of hidden layer nodes and number of training times.

[0011] The present invention also provides a fuel quality soft measurement system for thermal power generation, including: A real-time data acquisition module for acquiring the real-time operation index data of the boiler related to fuel quality; A real-time data screening module, configured to perform feature screening on the real-time operation index data of the boiler related to fuel quality based on the mutual information method to obtain key real-time operation index data; A quality prediction module, configured to input the key real-time operation index data into a fuel quality soft measurement model, and output a soft measurement result of the quality of thermal power generation fuel; wherein, the fuel quality soft measurement model is a model based on a long short-term memory neural network.

[0012] The present invention also provides a device for soft measurement of the quality of thermal power generation fuel, including: A processor, suitable for executing a computer program; A computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by the processor, the method for soft measurement of the quality of thermal power generation fuel is executed.

[0013] The present invention also provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the method for soft measurement of the quality of thermal power generation fuel is implemented.

[0014] The present invention also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the method for soft measurement of the quality of thermal power generation fuel is implemented.

[0015] Compared with the prior art, the beneficial effects of the present invention are: The method for soft measurement of the quality of thermal power generation fuel provided by the present invention performs feature screening on the real-time operation index data of the boiler related to fuel quality through the mutual information method, and uses the screened key real-time operation index data as input, and uses a fuel quality soft measurement model based on a long short-term memory neural network to predict the fuel quality, realizing real-time online detection of the fuel quality, being able to quickly respond to boiler operations according to changes in fuel quality, thereby ensuring boiler combustion efficiency and emission control, enhancing the stability and flexibility of the production process, and thus improving the overall operation efficiency and environmental protection compliance of thermal power plants; wherein, through real-time online measurement of fuel quality, real-time monitoring of fuel quality is realized, avoiding the sampling delay and high cost problems brought by traditional methods; on the other hand, through accurate prediction and adjustment, it helps to optimize the combustion process, improve combustion efficiency, reduce emissions, and reduce energy waste and environmental pollution.

[0016] The soft measurement system for the quality of thermal power generation fuel, the device for soft measurement of the quality of thermal power generation fuel, the computer-readable storage medium and the computer program product provided by the present invention have all the advantages of the above-mentioned method for soft measurement of the quality of thermal power generation fuel. Description of the Drawings

[0017] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 Flow chart of the quality soft measurement method for the fuel of thermal power generation provided in Embodiment 1; Figure 2 Block diagram of the structure of the quality soft measurement system for the fuel of thermal power generation provided in Embodiment 2; Figure 3 Block diagram of the structure of the quality soft measurement device for the fuel of thermal power generation provided in Embodiment 3. Specific embodiments

[0019] In order to make the technical problems, technical solutions and beneficial effects solved by the present application more clear and understandable, the following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application; obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope protected by the present application.

[0020] The present invention provides a quality soft measurement method for the fuel of thermal power generation, including the following steps: Step 100: Obtain the real-time operation index data of the boiler related to the fuel quality.

[0021] Step 200: Based on the mutual information method, perform feature screening on the real-time operation index data of the boiler related to the fuel quality to obtain the key real-time operation index data.

[0022] Step 300: Input the key real-time operation index data into the fuel quality soft measurement model, and output the quality soft measurement result of the fuel for thermal power generation; wherein, the fuel quality soft measurement model is a model based on a long short-term memory neural network.

[0023] The quality soft measurement method of thermal power generation fuel described in the present invention uses the real-time operation index data of the boiler to predict the fuel quality in real time, realizes the rapid online measurement of the fuel quality, and can significantly improve the response speed of the fuel combustion process control; secondly, the mutual information method is used to quantify the non-linear correlation between the real-time operation index data of the boiler and the fuel quality to propose redundant parameters; the variable screening based on mutual information not only optimizes the feature space, but also can significantly improve the generalization ability and prediction accuracy of the soft measurement model, provides an efficient and reliable technical support for the real-time monitoring and optimal control of complex industrial processes, and has high engineering application value; the model based on the long short-term memory neural network is used to predict the fuel quality, and the memory gate mechanism in the long short-term memory neural network can model the long-period time series dependence, accurately associate the historical operation state with the current fuel quality change, and ensure the accuracy of the fuel quality soft measurement result.

[0024] The following uses some specific embodiments to further explain the quality soft measurement method of thermal power generation fuel provided by the present invention: Embodiment 1 As shown in the attached Figure 1 figure, Embodiment 1 provides a quality soft measurement method of thermal power generation fuel, including the following steps: Step 1. Establish a fuel soft measurement model. The construction process of the fuel soft measurement model is as follows: Step 11. Obtain the historical operation index data of the boiler related to the fuel quality; among them, the historical operation index data of the boiler related to the fuel quality includes boiler fuel supply data, characteristic fuel flow rate, furnace temperature, oxygen concentration, flue gas temperature, and fuel air flow rate; specifically, based on the time series theory, historical data is collected from the distributed control system of the thermal power plant to obtain the historical operation index data of the boiler related to the fuel quality.

[0025] It should be noted that during the historical data collection process, due to various reasons such as sensor failures and link factors, some data may be missing or there may be outliers in the collected original historical data; therefore, after obtaining the original historical data, data preprocessing is performed, including missing value filling, outlier removal, data smoothing, and normalization processing, to ensure data quality and consistency, and then obtain the historical operation index data of the boiler related to the fuel quality; among them, the preprocessing process of the original historical data is the same as the general data preprocessing technology, which will not be elaborated here.

[0026] Step 12. Based on the mutual information method, perform feature screening on the historical operation index data of the boiler related to the fuel quality to obtain key historical operation index data; among them, the key historical operation index data includes the historical operation index data of the boiler related to the fuel quality, including characteristic fuel flow rate, furnace temperature, oxygen concentration, flue gas temperature, and fuel air flow rate.

[0027] Since not all of the collected historical data is closely related to the quality of the fuel, if all the collected historical data variables are used as input variables for soft sensor modeling, it will lead to redundancy in the soft sensor model, a large amount of computation, and is prone to overfitting, resulting in poor results of the soft sensor model. Therefore, it is necessary to perform feature selection on the historical data collected that is closely related to the fuel quality, and establish a soft sensor model based on the selected feature variables, which can reduce the modeling time, improve the model performance, reduce the selection of hardware sensors, save costs, and increase the usability of the system.

[0028] Specifically, the process of screening features from the boiler historical operation index data related to the fuel quality to obtain the key historical operation index data includes: Step 121: Calculate the marginal information entropy of the boiler historical operation index data related to the fuel quality and the fuel quality index data respectively; where the fuel quality index data is the laboratory fuel quality measurement data for a preset time period; preferably, the fuel quality index data includes fuel calorific value, volatile matter, and ash content.

[0029] Specifically, the process of calculating the marginal information entropy of the boiler historical operation index data related to the fuel quality and the fuel quality index data is as follows:

[0030]

[0031] Among them, is the marginal information entropy of the boiler historical operation index data related to the fuel quality; is the data set of the boiler historical operation index data related to the fuel quality; is the boiler historical operation index data related to the fuel quality; is the marginal probability density function of the boiler historical operation index data related to the fuel quality; is the marginal information entropy of the fuel quality index data; is the data set of the fuel quality index data; is the fuel quality index data; is the marginal probability density function of the fuel quality index data.

[0032] Step 122: Calculate the joint entropy of the boiler historical operation index data and the fuel index data based on the marginal information entropy of the boiler historical operation index data related to the fuel quality and the marginal information entropy of the fuel quality index data; where the calculation process of the joint entropy of the boiler historical operation index data and the fuel index data is as follows:

[0033] Among them, is the joint entropy of the boiler historical operation index data and the fuel index data; The joint probability density function of the boiler historical operation index data and the fuel index data.

[0034] Step 123: Based on the mutual information theory, obtain the mutual information value between the boiler historical operation index data and the fuel quality index data according to the marginal information entropy of the boiler historical operation index data related to the fuel quality, the marginal information entropy of the fuel quality index data, and the joint entropy of the boiler historical operation index data and the fuel index data; among them, the calculation process of the mutual information value between the boiler historical operation index data and the fuel quality index data is as follows:

[0035] Among them, is the mutual information value between the boiler historical operation index data and the fuel quality index data.

[0036] Step 124: Obtain the key historical operation index data according to the mutual information value between the boiler historical operation index data and the fuel quality index data; specifically, compare the mutual information value between the boiler historical operation index data and the fuel quality index data with a preset mutual information value threshold; select the boiler historical operation index data corresponding to the mutual information value greater than the preset mutual information threshold as the key index to obtain the key historical operation index data; preferably, the preset mutual information threshold is 0.5.

[0037] It should be noted that compared with the traditional linear correlation analysis method, the mutual information method can effectively capture the non - linear correlation characteristics between variables, overcome the non - linear and strong coupling problems commonly existing in industrial process data, thereby improving the comprehensiveness and accuracy of variable selection; in the mutual information method, by quantifying the information sharing degree between the boiler operation index data and the fuel quality index data, redundant or low - correlation variables can be removed, reducing the model complexity and enhancing the robustness; in addition, mutual information has non - parametric characteristics, without presupposing the data distribution form, showing strong adaptability to noise data, especially suitable for industrial scenarios with multi - modality and non - Gaussian distribution; secondly, variable screening based on the mutual information method not only optimizes the feature space, but also can significantly improve the generalization ability and prediction accuracy of the soft - sensing model, providing an efficient and reliable technical support for the real - time monitoring and optimal control of complex industrial processes, and having high engineering application value.

[0038] It should also be noted that Embodiment 1 of the present invention also provides another implementation method for obtaining key historical operation index data, specifically: for all candidate variables in the boiler historical operation index data, sort them in descending order according to the mutual information value between the boiler historical operation index data and the fuel quality index data; from the sorting result, select the top several variables with a relatively high mutual information value with the target variable as the potential feature set, that is, obtain the key historical operation index data; among them, the number of selected features can be determined based on a set threshold or according to specific requirements.

[0039] Step 13: Use the key historical operation index data as training data to train and validate a pre-constructed Long Short-Term Memory (LSTM) neural network to obtain a trained LSTM neural network; use the trained LSTM neural network as the fuel quality soft sensor model.

[0040] Specifically, the process of training and validating a pre-constructed LSTM neural network to obtain a trained LSTM neural network is as follows: Step 131: Divide the key historical operation index data into a data set according to a ratio of 7:3 to obtain a training set and a validation set; among them, the training set is used for model training, and the validation set is used for model validation.

[0041] Step 132: Establish an LSTM neural network to construct a soft sensor model to obtain a pre-constructed LSTM neural network; among them, considering that the fuel quality has a strong correlation before and after in the time series, the LSTM neural network is used to construct the soft sensor model to utilize the gating mechanism of the LSTM neural network, so that the long-term dependence relationship can be effectively captured when processing time series data; secondly, through the dynamic update mechanism of the cell state, high-precision modeling of complex non-linear time series features in the data can be realized.

[0042] Step 133: Use the training set and the validation set to train and validate the pre-constructed LSTM neural network to obtain a trained LSTM neural network; among them, the seagull optimization algorithm is used to train and optimize the preset hyperparameters of the pre-constructed LSTM neural network; preferably, the preset hyperparameters of the pre-constructed LSTM neural network include the learning rate, the number of hidden layer nodes, and the number of training times.

[0043] It should be noted that in the process of configuring the hyperparameter combination of the LSTM neural network, there are significant non-linear coupling effects on the model prediction accuracy, convergence efficiency, and computational resource consumption. The traditional empirical parameter tuning method is difficult to balance the collaborative optimization of model prediction accuracy and computational efficiency. To achieve the global optimum of the model performance, in this Embodiment 1, the seagull optimization algorithm is adopted as the core optimization strategy, and a dynamic adaptive parameter optimization framework is constructed by simulating the bionic mechanism of seagull population migration and predation behavior. Based on the global search mechanism of swarm intelligence, the seagull optimization algorithm can efficiently traverse the hyperparameter solution space, accurately locate the parameter combination that makes the objective function reach the optimum, and avoid falling into the local optimum solution. Thus, on the premise of ensuring the prediction accuracy, the training time cost is significantly reduced, providing a feasibility guarantee for the deployment of the LSTM neural network.

[0044] Specifically, the process of using the seagull optimization algorithm to train and optimize the preset hyperparameters of the pre-constructed long short-term memory neural network includes: Step 1331: Determine the number of input layers and output layers of the pre-constructed LSTM neural network; set the preset parameters of the seagull optimization algorithm. Among them, the preset parameters of the seagull optimization algorithm include inertia weight, acceleration factor, seagull population size, and maximum number of iterations.

[0045] Step 1332: Set the position of each seagull individual ; among them, is the number of training times of the pre-constructed LSTM neural network, and the range is set in [10, 200]; is the learning rate of the pre-constructed LSTM neural network, and the range is set in [0.0001, 0.1]; is the number of hidden layer units of the pre-constructed LSTM neural network, and the range is set in [16, 512].

[0046] Step 1333: Evaluate the fitness function and update the position and velocity of the seagull individuals. Among them, the fitness function uses the mean square error method, and the calculation formula of the fitness function is:

[0047] Among them, is the fitness function value of the th seagull individual, is the total number of seagull individuals; is the true value; is the predicted value.

[0048] For each seagull individual, use the current position to train the pre-constructed LSTM neural network, and calculate the loss on the validation set to obtain the fitness value. Among them, the velocity update expression is:

[0049] wherein, is the speed of the th gull individual; at time is the inertia factor, controlling the influence of the old speed; is the speed of the th gull individual at time , are acceleration factors; , are random numbers between [0, 1]; is the optimal position of the th gull individual; is the th gull individual at time is the global optimal position of the population.

[0050] The position update expression is:

[0051]

[0052] wherein, is; is; , are dynamically adjusted coefficients; is the difference between the global optimal position of the population and the position of the th gull individual at time

[0053] Step 1334. After each update, compare the fitness value of each gull individual with the historical optimal solution; if the fitness of the current solution is better, update the individual optimal solution; if the fitness of the current solution is better, update the global optimal solution of the population.

[0054] Step 1335. Repeat the operations of Steps 1333 and 1334 until the set maximum number of iterations is reached; When the gull optimization algorithm reaches the stop condition, the global optimal solution of the pre-constructed LSTM neural network is the optimal number of training times, learning rate, and number of hidden layer units, and substituting them into the pre-constructed LSTM neural network to obtain the trained LSTM neural network.

[0055] Step 2: Obtain the real-time operation index data of the boiler related to the fuel quality. Among them, the real-time operation index data of the boiler related to the fuel quality includes the boiler fuel supply data, the characteristic fuel flow rate, the furnace temperature, the oxygen concentration, the flue gas temperature, and the fuel air flow rate; it should be noted that the process of obtaining the real-time operation index data of the boiler related to the fuel quality in Step 2 is similar to the process of obtaining the historical operation index data of the boiler related to the fuel quality in the above Step 11, and will not be elaborated here.

[0056] Step 3: Based on the mutual information method, perform feature screening on the real-time operation index data of the boiler related to the fuel quality to obtain the key real-time operation index data. Specifically, the process is as follows: The process of performing feature screening on the real-time operation index data of the boiler related to the fuel quality based on the mutual information method to obtain the key real-time operation index data includes: calculating the marginal information entropy of the real-time operation index data of the boiler related to the fuel quality and the fuel quality index data respectively; among them, the fuel quality index data includes the fuel calorific value, the volatile matter, and the ash content; according to the marginal information entropy of the real-time operation index data of the boiler related to the fuel quality and the marginal information entropy of the fuel quality index data, calculate the joint entropy of the real-time operation index data of the boiler and the fuel index data; based on the mutual information theory, according to the marginal information entropy of the real-time operation index data of the boiler related to the fuel quality, the marginal information entropy of the fuel quality index data, and the joint entropy of the real-time operation index data of the boiler and the fuel index data, obtain the mutual information value between the real-time operation index data of the boiler and the fuel quality index data; according to the mutual information value between the real-time operation index data of the boiler and the fuel quality index data, obtain the key real-time operation index data.

[0057] It should be noted that the process of performing feature screening on the real-time operation index data of the boiler related to the fuel quality in Step 3 to obtain the key real-time operation index data is similar to the process of performing feature screening on the historical operation index data of the boiler related to the fuel quality in the above Step 12, and will not be elaborated here.

[0058] Step 4: Input the key real-time operation index data into the fuel quality soft sensor model, and output the soft measurement result of the fuel quality of the thermal power generation fuel; among them, the fuel quality soft sensor model is a model based on the long short-term memory neural network.

[0059] The quality soft measurement method of thermal power generation fuel described in Embodiment 1 can establish a mathematical model between fuel quality index data and operation index data that is easy to measure online during boiler operation, enabling real-time prediction of fuel quality. It can not only effectively make up for the deficiencies of traditional detection methods but also achieve real-time monitoring and adjustment of fuel quality parameters during the boiler combustion process, ensuring the efficiency and environmental protection of the combustion process, thereby optimizing combustion efficiency, reducing energy consumption and emissions, and improving the overall operation level of thermal power plants.

[0060] Embodiment 2 As shown in the Figure 2 accompanying figure, Embodiment 2 provides a quality soft measurement system for thermal power generation fuel, including: a real-time data acquisition module, a real-time data screening module, and a quality prediction module.

[0061] The real-time data acquisition module is used to acquire real-time operation index data of the boiler related to fuel quality; the real-time data screening module is used to perform feature screening on the real-time operation index data of the boiler related to fuel quality based on the mutual information method to obtain key real-time operation index data; the quality prediction module is used to input the key real-time operation index data into the fuel quality soft measurement model and output the quality soft measurement result of thermal power generation fuel; among them, the fuel quality soft measurement model is a model based on a long short-term memory neural network.

[0062] Furthermore, the quality soft measurement system for thermal power generation fuel further includes a modeling module; the modeling module is used to establish a fuel soft measurement model.

[0063] Embodiment 3 As shown in the Figure 3 accompanying figure, Embodiment 3 provides a quality soft measurement device for thermal power generation fuel, including: a memory for storing a computer program; a processor for implementing the steps of the quality soft measurement method of thermal power generation fuel when executing the computer program.

[0064] When the processor executes the computer program, it implements the steps of the above-mentioned quality soft measurement method of thermal power generation fuel, for example: Establish a fuel soft measurement model; acquire real-time operation index data of the boiler related to fuel quality; perform feature screening on the real-time operation index data of the boiler related to fuel quality based on the mutual information method to obtain key real-time operation index data; input the key real-time operation index data into the fuel quality soft measurement model and output the quality soft measurement result of thermal power generation fuel; among them, the fuel quality soft measurement model is a model based on a long short-term memory neural network.

[0065] Or, when the processor executes the computer program, it implements the functions of each module in the above-mentioned quality soft measurement system for thermal power generation fuel, for example: A modeling module for establishing a fuel soft-sensing model; a real-time data acquisition module for acquiring real-time operation index data of a boiler related to fuel quality; a real-time data screening module for performing feature screening on the real-time operation index data of the boiler related to fuel quality based on the mutual information method to obtain key real-time operation index data; a quality prediction module for inputting the key real-time operation index data into the fuel quality soft-sensing model and outputting a soft-sensing result of the quality of thermal power generation fuel; wherein, the fuel quality soft-sensing model is a model based on a long short-term memory neural network.

[0066] Exemplarily, the computer program can be divided into one or more modules / units, and the one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of completing preset functions, and the instruction segments are used to describe the execution process of the computer program in the thermal power generation fuel quality soft-sensing device.

[0067] For example, the computer program can be divided into a modeling module, a real-time data acquisition module, a real-time data screening module, and a quality prediction module. The specific functions of each module are as follows: The modeling module is used to establish a fuel soft-sensing model; the real-time data acquisition module is used to acquire real-time operation index data of a boiler related to fuel quality; the real-time data screening module is used to perform feature screening on the real-time operation index data of the boiler related to fuel quality based on the mutual information method to obtain key real-time operation index data; the quality prediction module is used to input the key real-time operation index data into the fuel quality soft-sensing model and output a soft-sensing result of the quality of thermal power generation fuel; wherein, the fuel quality soft-sensing model is a model based on a long short-term memory neural network.

[0068] The thermal power generation fuel quality soft-sensing device can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The thermal power generation fuel quality soft-sensing device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above are examples of the thermal power generation fuel quality soft-sensing device, and do not constitute a limitation on the thermal power generation fuel quality soft-sensing device. It may include more components than the above, or combine some components, or different components. For example, the thermal power generation fuel quality soft-sensing device may further include input / output devices, network access devices, a bus, etc.

[0069] The so-called processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor, etc. The processor is the control center of the quality soft measurement device for thermal power generation fuel, and connects all parts of the quality soft measurement device for thermal power generation fuel through various interfaces and lines.

[0070] The memory can be used to store the computer programs and / or modules. The processor realizes various functions of the quality soft measurement device for thermal power generation fuel by running or executing the computer programs and / or modules stored in the memory, and by calling the data stored in the memory.

[0071] The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0072] Embodiment 4 Embodiment 4 of the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of the quality soft measurement method for thermal power generation fuel, such as: Establish a fuel soft measurement model; obtain real-time boiler operation index data related to fuel quality; based on the mutual information method, perform feature screening on the real-time boiler operation index data related to fuel quality to obtain key real-time operation index data; input the key real-time operation index data into the fuel quality soft measurement model, and output the quality soft measurement result of the thermal power generation fuel; among them, the fuel quality soft measurement model is a model based on a long short-term memory neural network.

[0073] If the modules / units integrated in the quality soft measurement system of the thermal power generation fuel are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0074] Based on such understanding, all or part of the processes in the above-mentioned quality soft measurement method of thermal power generation fuel can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned quality soft measurement method of thermal power generation fuel can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or preset intermediate form, etc.

[0075] The computer-readable storage medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0076] It should be noted that the content included in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice within the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0077] Embodiment 5 This Embodiment 5 provides a computer product. The computer program product includes a computer program, and the computer program is stored in a computer-readable storage medium; the processor of the quality soft measurement device for thermal power generation fuel reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the quality soft measurement device for thermal power generation fuel can execute the quality soft measurement method described in Embodiment 1, which will not be elaborated here.

[0078] It should be noted that those of ordinary skill in the art can understand that all or part of the processes in implementing the above-mentioned method embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the above-mentioned method embodiments.

[0079] The soft measurement method for the quality of thermal power generation fuel obtains real-time operation index data of the boiler related to fuel quality during the power generation process from the distributed control system, finds eigenvalues with a relatively high degree of correlation with the fuel quality index data through the mutual information theory, and uses the eigenvalues as the input of the trained long short-term memory neural network to predict the fuel quality; this method can realize real-time online detection of fuel quality and then realize online measurement of fuel quality; on the one hand, the real-time monitoring of fuel quality avoids the sampling delay and high cost problems brought by traditional methods; on the other hand, through accurate prediction and adjustment, it helps to optimize the combustion process, improve combustion efficiency, reduce emissions, reduce energy waste and environmental pollution; secondly, it enhances the stability and flexibility of the production process, thereby improving the overall operation efficiency and environmental protection compliance.

[0080] In the present invention, by utilizing the relationship between the measurable parameters of online monitoring and the fuel quality parameters, the real-time estimation of fuel quality is realized by using mathematical modeling and data-driven methods. There is no need to directly sample and analyze the fuel. The calorific value, ash content, volatile matter and other quality parameters of the fuel can be predicted through real-time data, so as to realize the precise control of the boiler combustion process, improve combustion efficiency and reduce emissions.

[0081] The above embodiments are only one of the implementation manners capable of realizing the technical solution of the present invention. The scope of protection required by the present invention is not limited only by this embodiment, but also includes any changes, substitutions and other implementation manners that are easily conceivable by those skilled in the art within the technical scope disclosed by the present invention.

Claims

1. A method for soft measurement of the quality of fuel for thermal power generation, characterized in that, Including: Obtain the real-time operation index data of the boiler related to the fuel quality; Based on the mutual information method, perform feature screening on the real-time operation index data of the boiler related to the fuel quality to obtain the key real-time operation index data; Input the key real-time operation index data into the fuel quality soft sensor model, and output the soft measurement result of the fuel quality of thermal power generation fuel; wherein, the fuel quality soft sensor model is a model based on a long short-term memory neural network.

2. The quality soft measurement method of a fuel for thermal power generation according to claim 1, characterized in that, The real-time operation index data of the boiler related to the fuel quality includes boiler fuel supply data, characteristic fuel flow rate, furnace temperature, oxygen concentration, flue gas temperature, and fuel air flow rate.

3. The quality soft measurement method of a fuel for thermal power generation according to claim 1, characterized in that The process of performing feature screening on the real-time operation index data of the boiler related to the fuel quality based on the mutual information method to obtain the key real-time operation index data includes: Calculate the marginal information entropy of the real-time operation index data of the boiler related to the fuel quality and the fuel quality index data respectively; wherein, the fuel quality index data includes fuel calorific value, volatile matter, and ash content; Calculate the joint entropy of the real-time operation index data of the boiler and the fuel index data according to the marginal information entropy of the real-time operation index data of the boiler related to the fuel quality and the marginal information entropy of the fuel quality index data; Based on the mutual information theory, obtain the mutual information value between the real-time operation index data of the boiler and the fuel quality index data according to the marginal information entropy of the real-time operation index data of the boiler related to the fuel quality, the marginal information entropy of the fuel quality index data, and the joint entropy of the real-time operation index data of the boiler and the fuel index data; Obtain the key real-time operation index data according to the mutual information value between the real-time operation index data of the boiler and the fuel quality index data.

4. A quality soft measurement method for a fuel in thermal power generation according to claim 1, characterized in that, The construction process of the fuel quality soft sensor model includes: Obtain the historical operation index data of the boiler related to the fuel quality; Based on the mutual information method, perform feature screening on the historical operation index data of the boiler related to the fuel quality to obtain the key historical operation index data; Use the key historical operation index data as training data to train and verify a pre-constructed long short-term memory neural network to obtain a trained long short-term memory neural network; use the trained long short-term memory neural network as the fuel quality soft sensor model.

5. A quality soft measurement method for a fuel in thermal power generation according to claim 4, characterized in that In the process of using the key historical operation index data as training data to train and verify a pre-constructed long short-term memory neural network to obtain a trained long short-term memory neural network, the seagull optimization algorithm is used to train and optimize the preset hyperparameters of the pre-constructed long short-term memory neural network.

6. A quality soft measurement method for a fuel in thermal power generation according to claim 5, characterized in that The preset hyperparameters of the pre-constructed long short-term memory neural network include learning rate, number of hidden layer nodes, and number of training times.

7. A quality soft measurement system for fuel in thermal power generation, characterized in that, Including: A real-time data acquisition module for obtaining the real-time operation index data of the boiler related to the fuel quality; A real-time data screening module for performing feature screening on the real-time operation index data of the boiler related to the fuel quality based on the mutual information method to obtain the key real-time operation index data; A quality prediction module for inputting the key real-time operation index data into a fuel quality soft sensor model to output a soft measurement result of the thermal power generation fuel quality; wherein, the fuel quality soft sensor model is a model based on a long short-term memory neural network.

8. A quality soft measurement device for thermal power generation fuel, characterized in that, Comprising: A processor adapted to execute a computer program; A computer-readable storage medium storing a computer program which, when executed by the processor, executes the method for soft measurement of the thermal power generation fuel quality according to any one of claims 1-6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for soft measurement of the thermal power generation fuel quality according to any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program which, when executed by the processor, implements the method for soft measurement of the thermal power generation fuel quality according to any one of claims 1-6.

Images