Online soft measurement method, system and equipment for fineness of pulverized coal in coal mill and storage medium

Abstract

The invention discloses an online soft measurement method, system and device for the fineness of pulverized coal in a coal mill and a storage medium. The method comprises the steps that a mechanism approximation model of the fineness of the pulverized coal in the coal mill and an error compensation model based on a T-S fuzzy neural network are established; and estimating the pulverized coal fineness of the coal mill under the current working condition by using the mechanism approximation model, and compensating the mechanism approximation model by using the T-S fuzzy neural network error compensation model so as to obtain the pulverized coal fineness of the coal mill under the current working condition. According to the method, the initial parameters of the mechanism approximation model and the error compensation model are obtained through the historical operation data of the coal mill, and the established prediction model is corrected according to the operation condition of the coal mill, so that the pulverized coal fineness of the current coal mill is obtained according to the operation parameters and the coal quality parameters of the coal mill; and accurate and visual operation parameters are provided for the operation of the thermal power plant.

Description

Technical Field

[0001] The invention relates to the technical field of thermal power generation equipment, and in particular to an online soft measurement method, system, equipment and storage medium for the fineness of coal powder in a coal mill. Background Art

[0002] Coal powder fineness is an important parameter in the operation of thermal power plants. Reducing coal powder fineness means increasing the surface area of ​​coal powder. Increasing the surface area can speed up the chemical reaction rate. In addition, increasing the surface area can also accelerate the precipitation process of volatiles in coal powder. These two factors are conducive to the ignition and combustion of coal powder. Therefore, reducing coal powder fineness can improve boiler efficiency. Although from the perspective of improving boiler efficiency, coal powder fineness should be reduced as much as possible, reducing coal powder fineness requires increasing the power consumption of the pulverizer. The economic efficiency of the overall operation of a thermal power plant needs to comprehensively consider the boiler efficiency and the power consumption of the pulverizer, so there is an optimal value for coal powder fineness.

[0003] Since there is an optimal value for coal powder fineness, it is necessary to measure the coal powder fineness during the operation of thermal power plants. Traditional methods for measuring coal powder fineness include microwave measurement, laser pulsation, electrostatic measurement, etc. These methods test the fineness of coal powder through corresponding sensors, but the scouring of the sensor by coal powder and the deposition of coal powder on the sensor surface will have a great impact on the test results. At present, the cost of measuring equipment for directly testing coal powder fineness is also high. The high cost and uncertainty of test results make it difficult for traditional coal powder fineness measurement methods to be used in online measurements in thermal power plants.

[0004] Soft sensing establishes mathematical models and uses readily available parameters to predict variables that are difficult to measure. Therefore, its testing cost is low. At the same time, it can realize dynamic testing of variables, so it has been widely used in industrial applications.

[0005] There are many factors that affect the fineness of coal powder, mainly including: coal quality characteristics, coal mill structure characteristics and its operating characteristics, separator structure and its operating characteristics and other pulverizing system characteristic parameters. There is a strong correlation between the influencing factors, and there is a certain degree of coupling between the factors. The online parameter identification model established by relying on curve fitting or numerical simulation of a single parameter, the measurement accuracy cannot meet the actual operation and control needs of the pulverizing system. Therefore, the soft measurement of coal powder fineness needs to adopt a combination of mechanism modeling and online identification, comprehensively considering the influence of various factors on coal powder fineness, uniformity index, and coal powder concentration, so as to establish an online parameter identification model for the pulverizing system. Summary of the invention

[0006] The purpose of the present invention is to provide an online soft measurement method, system, equipment and storage medium for coal powder fineness in a coal mill, so as to solve the problem that coal powder fineness in a thermal power plant is difficult to measure online and provide a reference for the operation of the thermal power plant.

[0007] To achieve the above purpose, the technical solution provided by the present invention is:

[0008] The first aspect of the present application provides an online soft measurement method for the coal powder fineness in a coal mill, establishes a mechanism approximation model of the coal powder fineness of the coal mill and an error compensation model based on a TS fuzzy neural network; uses the mechanism approximation model to estimate the coal powder fineness of the coal mill under the current operating conditions, and uses the TS fuzzy neural network error compensation model to compensate the mechanism approximation model, thereby obtaining the coal powder fineness of the coal mill under the current operating conditions.

[0009] To optimize the above technical solutions, the specific measures taken also include:

[0010] The mechanism approximate model of coal powder fineness in coal mill is modeled by the following polynomial:

[0011]

[0012] The mechanism approximate model of coal powder fineness in coal mill includes the following seven variables:

[0013] HGI represents the coal grindability index, M represents the total moisture content of coal, m represents the coal feed rate of the coal mill, F represents the hydraulic loading force, T and Q represent the inlet air temperature and inlet air volume, respectively, and w represents the separator speed;

[0014] These seven variables are used as coefficient matrices, and a mechanism approximation model of the mechanism approximation model is constructed through 7*7 matrix A, 7*1 matrix B and constant C. The predicted value of coal mill coal powder fineness calculated by the mechanism approximation model of coal mill coal powder fineness is represented by R.

[0015] In the mechanism approximation model of coal powder fineness of the coal mill, the coal grindability coefficient HGI, the total moisture content of the coal M, the coal feed amount m, the hydraulic loading force F, the inlet air temperature T, the inlet air volume Q and the separator speed w are used as independent variables, and the coal powder fineness R of the coal mill calculated by the mechanism approximation model of coal powder fineness of the coal mill is used as the dependent variable.

[0016] Each independent variable in the mechanism approximation model of coal powder fineness of coal mill is used as the input of TS fuzzy neural network, and the difference between the actual test value of coal powder fineness of coal mill and the predicted value R of coal powder fineness of coal mill calculated by the mechanism approximation model of coal powder fineness of coal mill is used as the output of TS fuzzy neural network to form an error compensation model.

[0017] The soft measurement value of the coal powder fineness of the coal mill is the sum of the predicted value of the coal powder fineness of the coal mill calculated by the mechanism approximate model of the coal powder fineness of the coal mill and the calculation result of the error compensation model.

[0018] Furthermore, the parameters of the mechanism approximation model and the error compensation model of the coal powder fineness of the coal mill are obtained through a genetic algorithm, and the initial values ​​of the parameters are calculated through the historical operation data of the coal mill. At the same time, the parameters of the mechanism approximation model and the error compensation model of the coal powder fineness of the coal mill are updated using real-time data during the operation of the coal mill.

[0019] Preferably, the online soft measurement method for coal powder fineness in a coal mill specifically comprises the following steps:

[0020] Step S1: Collect several groups of data samples, each group of data samples includes parameters: coal feed rate of coal mill, hydraulic loading force, inlet air temperature, inlet air volume and separator speed; obtain coal powder fineness and parameters of coal mill through sampling test: coal grindability coefficient, coal total moisture; divide the data samples to form training set samples and test set samples;

[0021] Step S2: Based on the genetic algorithm, the initial values ​​of the coefficient matrix of the mechanism approximate model of coal powder fineness of the coal mill are obtained by using the training set samples;

[0022] Step S3: establishing an error compensation model using a TS fuzzy neural network, wherein the input layer of the neural network includes a plurality of nodes corresponding to each parameter respectively; the output layer of the neural network includes only one node corresponding to the difference between the actual test value of the coal powder fineness of the coal mill and the predicted value of the coal powder fineness of the coal mill calculated by the mechanism approximation model of the coal powder fineness of the coal mill;

[0023] Step S4: Determine the maximum number of training times and training accuracy of the neural network in step S3;

[0024] Step S5: Use the training set samples and the genetic algorithm to train the neural network in step S3. When the neural network meets the training accuracy or reaches the maximum number of training steps, stop the training.

[0025] Step S6: Use the test set samples to test the error compensation model. If the accuracy meets the requirements, proceed to the next step; otherwise, return to step S4;

[0026] Step S7: the coal mill starts running, and the sum of the coal mill coal powder fineness prediction value calculated by the coal mill coal powder fineness mechanism approximation model and the result calculated by the error compensation model is used as the soft measurement value of the coal mill coal powder fineness; at the same time, various parameters continue to be collected; the coal powder is sampled at intervals during the operation and the coal powder fineness is recorded; when the data volume is sufficient, the data is divided into a training set and a test set, and the process returns to step S2.

[0027] The second aspect of the present application provides an online soft measurement system for coal powder fineness in a coal mill, comprising:

[0028] Data acquisition module, used to collect data samples during the operation of the coal mill;

[0029] The mechanism approximate model estimation module is used to estimate the coal powder fineness of the coal mill under the current working conditions;

[0030] Error compensation model calculation module, used to compensate the mechanism approximation model;

[0031] The online soft measurement module is used to obtain the coal powder fineness of the coal mill under the current working conditions according to the compensation results.

[0032] The online soft measurement system for coal powder fineness in a coal mill is used to execute the online soft measurement method for coal powder fineness in a coal mill as described above.

[0033] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the online soft measurement method for coal powder fineness in the coal mill as described above is implemented.

[0034] A fourth aspect of the present application provides a computer-readable storage medium storing a computer program, wherein the computer program enables a computer to execute the online soft measurement method for coal powder fineness in a coal mill as described above.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The present invention establishes a mechanism approximate model of coal powder fineness of a coal mill, and simultaneously establishes an error compensation model of the mechanism approximate model by using a TS fuzzy neural network; the initial parameters of the mechanism approximate model and the error compensation model are obtained through the historical operation data of the coal mill, and the established prediction model is corrected according to the operating conditions of the coal mill, so as to obtain the coal powder fineness of the current coal mill according to the operating parameters of the coal mill and the coal quality parameters, thereby providing accurate and intuitive operating parameters for the operation of a thermal power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 : Schematic diagram of the control method of the online soft measurement method of coal powder fineness in the coal mill of the present invention. DETAILED DESCRIPTION

[0038] The above contents of the present invention are further described in detail below in the form of embodiments, but this should not be understood as the scope of the above subject matter of the present invention being limited to the following embodiments, and all technologies realized based on the above contents of the present invention belong to the scope of the present invention.

[0039] In one embodiment, the present invention proposes an online soft measurement method for coal powder fineness in a coal mill, establishes a mechanism approximate model of coal powder fineness in a coal mill, and establishes an error compensation model of the mechanism approximate model using a TS fuzzy neural network. The initial parameters of the mechanism approximate model and the error compensation model of coal powder fineness in the coal mill are obtained through the historical operation data of the coal mill, and the established prediction model is corrected according to the operating conditions of the coal mill, so as to obtain the coal powder fineness of the current coal mill according to the operating parameters of the coal mill and the coal quality parameters, and provide accurate and intuitive operating parameters for the operation of the thermal power plant.

[0040] The specific technical solutions of the present invention are as follows:

[0041] The mechanism approximate model of coal powder fineness in coal mill is established as shown in the following formula:

[0042]

[0043] Where R is the coal powder fineness, HGI is the coal grindability coefficient, M is the total moisture content of the coal, m is the coal feed rate of the coal mill, F is the hydraulic loading force, T is the inlet air temperature, Q is the inlet air volume, w is the separator speed, A, B, and C are the model coefficients of the mechanism approximate model of the coal powder fineness of the coal mill, A is a 7*7 matrix, B is a 7*1 matrix, and C is a constant.

[0044] The historical operation data of coal mill are used as training samples, and the genetic algorithm is used to obtain the initial values ​​of model coefficients A, B and C in the mechanism approximate model of coal powder fineness of coal mill.

[0045] The error compensation model of the mechanism approximate model is established by using the TS fuzzy neural network model. The input layer of the TS fuzzy neural network model is the coal grindability coefficient, coal total moisture, coal feed of the coal mill, hydraulic loading force, inlet air temperature, inlet air volume and separator speed, which is a 7-dimensional vector; the output layer of the error compensation model based on the TS fuzzy neural network is the difference between the test value of the coal mill coal fineness and the predicted value of the coal mill coal fineness calculated by the mechanism approximate model of the coal mill coal fineness. The middle layer of the TS fuzzy neural network error compensation model contains a certain number of neuron nodes.

[0046] In some implementations, historical operation data of the coal mill and initial parameters of the mechanism approximation model are used as training sets, and a genetic algorithm is used to train a compensation model based on the TS fuzzy neural network error.

[0047] In some embodiments, during the operation of the coal mill, the operating conditions of the coal mill (coal feed rate of the coal mill, hydraulic loading force, inlet air temperature and inlet air volume) and coal quality characteristic parameters (coal grindability coefficient and total moisture content of the coal) are first collected, and the coal powder fineness of the coal mill is calculated using the mechanism approximation model of the coal powder fineness of the coal mill, and then the error of the mechanical technology model is calculated using the compensation model based on the TS fuzzy neural network error; the sum of the predicted value of the coal powder fineness of the coal mill calculated by the mechanism approximation model of the coal powder fineness of the coal mill and the calculation result of the error compensation model is the soft measurement value of the coal powder fineness of the coal mill.

[0048] In some embodiments, during the operation of the coal mill, the coal powder is sampled and its fineness is tested, and the operating conditions of the coal mill and the coal quality parameters when the coal powder is sampled are recorded, and the sampling values ​​are used to update the parameters of the mechanism technology model and the error compensation model.

[0049] The invention establishes a mechanism approximate model of coal powder fineness of coal mill and an error compensation model based on TS fuzzy neural network, thereby realizing online soft measurement of coal powder fineness of coal mill and providing reference data for the operation of thermal power plant.

[0050] The technical solution of the present invention is further described in detail below in conjunction with specific embodiments:

[0051] Taking a coal mill in a power plant as an example, the implementation steps are as follows:

[0052] Step 1: Use the DCS system to collect 15 days of operation data of the power plant. Each set of data includes coal feed rate of coal mill, hydraulic loading force, inlet air temperature, inlet air volume and separator speed. Coal powder fineness, coal grindability coefficient and coal total moisture are obtained through sampling test. The sample data are divided into the first 8 days of data as training set samples and the last 7 days of data as test set samples.

[0053] Step 2: Based on the genetic algorithm, the initial values ​​of the coefficient matrix of the mechanism approximate model of coal powder fineness of the coal mill are obtained using the training set samples;

[0054] Step 3: Use TS fuzzy neural network to establish error compensation model; the input layer of the neural network is 7 nodes: coal grindability coefficient, coal total moisture, coal feed of coal mill, hydraulic loading force, inlet air temperature, inlet air volume and separator speed; the output layer of the neural network contains only one node: the difference between the test value of coal powder fineness of coal mill and the predicted value of mechanism approximation model;

[0055] Step 4: Determine the maximum number of training times and training accuracy of the neural network in step 3

[0056] Step 5: Use the training set samples and genetic algorithm to train the neural network in step 3. When the neural network meets the training accuracy or reaches the maximum number of training steps, stop training;

[0057] Step 6: Use the test set samples to test the error compensation model. If the accuracy meets the requirements, proceed to the next step; otherwise, return to step 4;

[0058] Step 7: The coal mill starts running, and the sum of the calculation results of the mechanism approximation model and the calculation results of the error compensation model is used as the soft measurement value of the coal powder fineness of the coal mill; at the same time, the DCS system continues to collect coal feed, hydraulic loading force, inlet air temperature, inlet air volume, separator speed, coal grindability coefficient and coal total moisture and other parameters; during the operation, the coal powder is sampled after a period of time, and the coal powder fineness is recorded. When the amount of data is sufficient, the data is divided into a training set and a test set, and return to step 2.

[0059] In another embodiment, the present invention provides an online soft measurement system for coal powder fineness in a coal mill, comprising:

[0060] Data acquisition module, used to collect data samples during the operation of the coal mill;

[0061] The mechanism approximate model estimation module is used to estimate the coal powder fineness of the coal mill under the current working conditions;

[0062] Error compensation model calculation module, used to compensate the mechanism approximation model;

[0063] The online soft measurement module is used to obtain the coal powder fineness of the coal mill under the current working conditions according to the compensation results.

[0064] The online soft measurement system for coal powder fineness in a coal mill is used to implement the online soft measurement method for coal powder fineness in a coal mill as described above.

[0065] In another embodiment, the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the online soft measurement method for coal powder fineness in the coal mill as described above is implemented.

[0066] A fourth aspect of the present application provides a computer-readable storage medium storing a computer program, wherein the computer program enables a computer to execute the above-mentioned online soft measurement method for the fineness of coal powder in a coal mill.

[0067] In the embodiments disclosed in the present application, the computer storage medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment. The computer storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the above. More specific examples of computer storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0068] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any technician familiar with the profession, without departing from the scope of the technical solution of the present invention, according to the technical essence of the present invention, any simple modification, equivalent replacement and improvement made to the above embodiment still falls within the protection scope of the technical solution of the present invention.

Claims

1. An online soft measurement method for coal powder fineness in a coal mill, characterized in that: A mechanism approximate model of coal powder fineness of coal mill and an error compensation model based on TS fuzzy neural network are established; the mechanism approximate model is used to estimate the coal powder fineness of coal mill under current operating conditions, and the TS fuzzy neural network error compensation model is used to compensate the mechanism approximate model, so as to obtain the coal powder fineness of coal mill under current operating conditions.

2. The on-line soft measurement method for coal powder fineness in a coal mill according to claim 1, characterized in that: The mechanism approximate model of coal powder fineness in coal mill is modeled by the following polynomial: The mechanism approximate model of coal powder fineness in the coal mill includes the following seven variables: HGI represents the coal grindability index, M represents the total moisture content of coal, m represents the coal feed rate of the coal mill, F represents the hydraulic loading force, T and Q represent the inlet air temperature and inlet air volume, respectively, and w represents the separator speed; These seven variables are used as coefficient matrices, and a mechanism approximation model of the mechanism approximation model is constructed through 7*7 matrix A, 7*1 matrix B and constant C. The predicted value of coal mill coal powder fineness calculated by the mechanism approximation model of coal mill coal powder fineness is represented by R.

3. The on-line soft measurement method for coal powder fineness in a coal mill according to claim 2, characterized in that: In the mechanism approximation model of coal powder fineness of the coal mill, the coal grindability coefficient HGI, the total moisture content of the coal M, the coal feed amount m, the hydraulic loading force F, the inlet air temperature T, the inlet air volume Q and the separator speed w are used as independent variables, and the coal powder fineness R of the coal mill calculated by the mechanism approximation model of coal powder fineness of the coal mill is used as the dependent variable.

4. The on-line soft measurement method for coal powder fineness in a coal mill according to claim 3 is characterized in that: Each independent variable in the mechanism approximation model of coal powder fineness of coal mill is used as the input of TS fuzzy neural network, and the difference between the actual test value of coal powder fineness of coal mill and the predicted value R of coal powder fineness of coal mill calculated by the mechanism approximation model of coal powder fineness of coal mill is used as the output of TS fuzzy neural network to form an error compensation model.

5. The on-line soft measurement method for coal powder fineness in a coal mill according to claim 1, characterized in that: The soft measurement value of the coal powder fineness of the coal mill is the sum of the predicted value of the coal powder fineness of the coal mill calculated by the mechanism approximate model of the coal powder fineness of the coal mill and the calculation result of the error compensation model.

6. The on-line soft measurement method for coal powder fineness in a coal mill according to claim 1, characterized in that: The parameters of the mechanism approximation model and the error compensation model of the coal powder fineness of the coal mill are obtained through a genetic algorithm, and the initial values ​​of the parameters are calculated through the historical operation data of the coal mill. At the same time, the parameters of the mechanism approximation model and the error compensation model of the coal powder fineness of the coal mill are updated using real-time data during the operation of the coal mill.

7. The on-line soft measurement method for coal powder fineness in a coal mill according to claim 1, characterized in that: The specific steps include: Step S1: Collect several groups of data samples, each group of data samples includes parameters: coal feed rate of coal mill, hydraulic loading force, inlet air temperature, inlet air volume and separator speed; obtain coal powder fineness and parameters of coal mill through sampling test: coal grindability coefficient, coal total moisture; divide the data samples to form training set samples and test set samples; Step S2: Based on the genetic algorithm, the initial values ​​of the coefficient matrix of the mechanism approximate model of coal powder fineness of the coal mill are obtained by using the training set samples; Step S3: establishing an error compensation model using a TS fuzzy neural network, wherein the input layer of the neural network includes a plurality of nodes corresponding to each parameter respectively; the output layer of the neural network includes only one node corresponding to the difference between the actual test value of the coal powder fineness of the coal mill and the predicted value of the coal powder fineness of the coal mill calculated by the mechanism approximation model of the coal powder fineness of the coal mill; Step S4: Determine the maximum number of training times and training accuracy of the neural network in step S3; Step S5: Use the training set samples and the genetic algorithm to train the neural network in step S3. When the neural network meets the training accuracy or reaches the maximum number of training steps, stop the training. Step S6: Use the test set samples to test the error compensation model. If the accuracy meets the requirements, proceed to the next step; Otherwise, return to step S4; Step S7: the coal mill starts running, and the sum of the coal mill coal powder fineness prediction value calculated by the coal mill coal powder fineness mechanism approximation model and the result calculated by the error compensation model is used as the soft measurement value of the coal mill coal powder fineness; at the same time, various parameters continue to be collected; the coal powder is sampled at intervals during the operation and the coal powder fineness is recorded; when the data volume is sufficient, the data is divided into a training set and a test set, and the process returns to step S2.

8. An online soft measurement system for coal powder fineness in a coal mill, characterized in that: include: Data acquisition module, used to collect data samples during the operation of the coal mill; The mechanism approximate model estimation module is used to estimate the coal powder fineness of the coal mill under the current working conditions; Error compensation model calculation module, used to compensate the mechanism approximation model; An online soft measurement module is used to obtain the coal powder fineness of the coal mill under the current working conditions according to the compensation results; The online soft measurement system for coal powder fineness in a coal mill is used to execute the online soft measurement method for coal powder fineness in a coal mill according to any one of claims 1 to 7.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, an online soft measurement method for coal powder fineness in a coal mill as claimed in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, wherein the computer program enables a computer to execute the online soft measurement method for coal powder fineness in a coal mill according to any one of claims 1 to 7.

Images