Optimization method, system, equipment and medium of desulfurization efficiency enhancement device

Through real-time data processing and intelligent algorithms of the DCS system, the design of desulfurization pallet tower is optimized, which solves the problem of desulfurization efficiency calculation accuracy in the existing technology, and realizes a more efficient design of desulfurization and efficiency enhancement device, which is suitable for ultra-low emission transformation coal-fired power station.

CN114911169BActive Publication Date: 2024-12-03DATANG ENVIRONMENT IND GRP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210661547.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-13
Publication Date
2024-12-03
Estimated Expiration
2042-06-13

AI Technical Summary

Technical Problem

The existing desulfurization pallet design relies on outdated empirical formulas, resulting in a decrease in the calculation accuracy of desulfurization efficiency in coal-fired power stations after ultra-low emission transformation, which cannot meet the design needs in the new environment.

Method used

Key parameter data is obtained in real time through the DCS system, preprocessing and cleaning is performed, and combined with gray correlation analysis method and self-learning algorithm, the relative impact coefficients are fitted and corrected, the pallet tower design scheme is optimized, and the design accuracy is improved.

Benefits of technology

It realizes the design of desulfurization and efficiency enhancement device more suitable for ultra-low emission transformation units, improves the design accuracy and applicability of the pallet tower device, and meets the desulfurization efficiency needs in the new environment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114911169B_ABST
    Figure CN114911169B_ABST
Patent Text Reader

Abstract

The present invention provides an optimization method, system, equipment and medium for a desulfurization efficiency enhancement device, the method comprising: obtaining the key parameter data required for correction in the desulfurization system in real time according to the DCS system, and combining each key parameter into a data tuple after preprocessing the data; cleaning and screening each data tuple to obtain each target tuple; fitting and calculating each target tuple based on a set benchmark value and a completion model to obtain a training set, and fitting the empirical formula of the relative influence coefficient of each key parameter after correction through a self-learning algorithm to determine the influence coefficient of each key parameter on the desulfurization efficiency; and determining an optimized tray device design scheme. The present invention is based on an analysis method based on big data, based on a large amount of system data and a self-learning algorithm, to correct the empirical formula of the operating data, so that the design parameters will be more accurate and more suitable for the desulfurization system of the unit after ultra-low emission modification. Improving the design accuracy is conducive to the design and development of the tray tower device.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This document relates to the field of desulfurization technology, and in particular to an optimization method, system, equipment and medium for a desulfurization efficiency enhancement device. Background Art

[0002] Existing SO 2 It is the main atmospheric pollutant emitted by coal-fired power plants and one of the main causes of air pollution. Most coal-fired power plants use wet desulfurization technology to remove SO in tail flue gas. 2 Coal-fired power plants need to introduce desulfurization efficiency-enhancing technologies to achieve the goal of “ultra-low emissions”.

[0003] At present, the main desulfurization efficiency enhancement technologies include desulfurization enhancers, turbulent flow equalizers, swirl coupling technology, ALDR technology, desulfurization trays, etc. Coal-fired power plants mainly use desulfurization trays as an efficiency enhancement method, such as Figure 1 As shown. Based on the empirical formula, the factors affecting the desulfurization efficiency of each key parameter are fitted, as follows:

[0004] The pallet tower parameter design is based on the completion (C) model for pallet design calculation. The main principles are as follows:

[0005] First, set the target desulfurization efficiency η SO2 , calculate the corresponding total influence coefficient (C t ). For each key parameter, set the relative influence coefficient R. To complete the verification calculation, the total influence coefficient C must be achieved. t The product of each relative coefficient is equal to ∏R. When a certain parameter design calculation is required, it is only necessary to adjust the relative influence coefficient R corresponding to the parameter until C t By making it equal to ∏R, the recommended value of the parameter can be output.

[0006] The relative coefficient R is calculated based on the historical operation data or pilot test data of the site to obtain the effect of each key parameter on the desulfurization efficiency η SO2 The degree of influence is the empirical formula for calculating the relative coefficient.

[0007] To complete the verification calculation, the following conditions must be met: C t =∏R.

[0008] The empirical formula used in the tray tower design is based on the power plant operation data at the beginning of this century. For example, for the relative coefficient of liquid-gas ratio, the empirical formula obtained by fitting is: R L / G =0.3636+0.0159(L / G), but with the implementation of environmental protection measures such as ultra-low emissions, zero wastewater discharge, and enhanced desulfurization measures, the key parameter levels of wet desulfurization systems in coal-fired power plants have changed significantly. Therefore, using empirical formulas based on outdated data for tray efficiency design calculations will produce certain deviations.

[0009] For example, the empirical formula of the relative coefficient of liquid-gas ratio mentioned above has a too narrow range and is not suitable for units after ultra-low emission transformation. Moreover, the formula fits the relative coefficient of liquid-gas ratio into a linear model, which leads to a decrease in fitting accuracy, which will cause a certain deviation in the calculation of tray efficiency enhancement design.

[0010] In view of this, a method for calculating a desulfurization efficiency enhancement device of a unit that has been modified for ultra-low emission is provided. Summary of the invention

[0011] One or more embodiments of this specification provide a method for optimizing a desulfurization efficiency enhancement device, comprising the steps of:

[0012] According to the DCS system, the key parameter data required for correction in the desulfurization system is obtained in real time, and after preprocessing the data, each key parameter is combined into a data tuple;

[0013] Perform data cleaning on each data tuple, set the benchmark values ​​of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, and use the grey correlation analysis method and the set correlation sequence threshold, with the set benchmark values ​​of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, to screen and calculate each data tuple to obtain each target tuple;

[0014] The training set is obtained by fitting and calculating each target tuple based on the set benchmark value and the completion model, and the empirical formula of the relative influence coefficient of each key parameter after correction is determined by fitting through the self-learning algorithm to determine the influence coefficient of each key parameter on the desulfurization efficiency;

[0015] According to the influence coefficient of each key parameter on the desulfurization efficiency, the system parameter requirements are obtained by combining the material balance calculation and the desulfurization system differential pressure calculation, and finally the optimized tray device design scheme is obtained.

[0016] One or more embodiments of this specification provide an optimization system for a desulfurization efficiency enhancement device, including:

[0017] Data acquisition and preprocessing module: acquires the key parameter data required for correction in the desulfurization system in real time according to the DCS system, and constructs the initial data set of each parameter after preprocessing the data;

[0018] Data cleaning and screening module: used to clean each data tuple based on the data sequence obtained by preprocessing the key parameter data obtained by the data acquisition and preprocessing module, set the reference value of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, and use the grey correlation analysis method and the set correlation sequence threshold, with the set reference value of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, to screen and calculate each data tuple to obtain each target tuple;

[0019] Fitting and correction module: used to obtain the training set by fitting each target tuple based on the set benchmark value and the completion model, and to determine the empirical formula of the relative influence coefficient of each key parameter after correction through self-learning algorithm fitting, and to determine the influence coefficient of each key parameter on desulfurization efficiency;

[0020] Pallet device design scheme confirmation module: Based on the determined influence coefficients of each key parameter on the desulfurization efficiency, combined with the material balance calculation and the desulfurization system differential pressure calculation, the system parameter requirements are obtained, and finally the optimized pallet device design scheme is obtained.

[0021] One or more embodiments of the present specification provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the optimization method of the desulfurization efficiency enhancement device as described above when executing the computer program.

[0022] One or more embodiments of the present specification provide a computer-readable storage medium, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the optimization method of the desulfurization efficiency enhancement device as described above when executing the computer program.

[0023] The present invention is based on the analysis method of big data, based on a large amount of system data, and uses intelligent algorithms to clean and screen the original DCS data, and calculates similar parameter sequences with only a single parameter difference according to the set reference value to achieve parameter variable control. Based on the self-learning algorithm, the relationship between each variable and the desulfurization efficiency is fitted, and the empirical formula is corrected for the operating data. The design parameters obtained by the new formula will be more accurate and more suitable for the desulfurization system of the unit after ultra-low emission transformation. Improving the design accuracy is conducive to the design and development of the tray tower device. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate one or more embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0025] Figure 1 The design flow chart of the existing common pallet device provided for this manual;

[0026] Figure 2 A schematic flow chart of a method for optimizing a desulfurization efficiency enhancement device provided in one or more embodiments of this specification;

[0027] Figure 3 A schematic flow chart of a tray design after data fitting for an optimization method of a desulfurization efficiency enhancement device provided in one or more embodiments of this specification;

[0028] Figure 4 A schematic diagram of a system for optimizing a desulfurization efficiency enhancement device provided in one or more embodiments of this specification;

[0029] Figure 5 A schematic diagram of the structure of a computer device provided for one or more embodiments of this specification. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the following will be combined with the drawings in one or more embodiments of this specification to clearly and completely describe the technical solutions in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.

[0031] The present invention is described in detail below in conjunction with specific implementation methods and the accompanying drawings.

[0032] Method Embodiment

[0033] According to an embodiment of the present invention, a method for optimizing a desulfurization efficiency enhancement device is provided, such as Figure 2 As shown, it is a flow chart of the optimization method of the desulfurization efficiency enhancement device of this embodiment. According to the optimization method of the desulfurization efficiency enhancement device of this embodiment of the present invention, the steps include:

[0034] Step S1, obtaining key parameter data required for correction in the desulfurization system in real time according to the DCS system, and after preprocessing the data, combining each key parameter into a data tuple;

[0035] The key parameters of this embodiment include the entry SO 2 Concentration, inlet O 2 Concentration, slurry circulation pump flow, desulfurization efficiency measurement value, flue gas inlet pressure and temperature, slurry pH value, etc.

[0036] Step S2, performing data cleaning on each data tuple, setting each key parameter and the reference value of the relative influence coefficient corresponding to each key parameter as a reference sequence, using the grey correlation analysis method and the set correlation sequence threshold, taking the set key parameters and the reference value of the relative influence coefficient corresponding to each key parameter as the reference sequence, screening and calculating each data tuple to obtain each target tuple;

[0037] Step S3, fitting and calculating each target tuple based on the set benchmark value and the completion model to obtain a training set, and fitting and determining the revised empirical formula of the relative influence coefficient of each key parameter through a self-learning algorithm to determine the influence coefficient of each key parameter on the desulfurization efficiency;

[0038] Step S4: according to the influence coefficient of each key parameter on the desulfurization efficiency determined in step S3, the system parameter requirements are obtained by combining the material balance calculation and the desulfurization system differential pressure calculation, and finally an optimized tray device design scheme is obtained.

[0039] Preferably, in this embodiment, the self-learning algorithm is built into the DCS system via PLC.

[0040] This embodiment is based on the analysis method of big data. Based on a large amount of system data, the original DCS data is cleaned and screened through an intelligent algorithm, and similar parameter sequences with only a single parameter difference are calculated according to the set reference value to achieve parameter variable control. Based on the self-learning algorithm, the relationship between each variable and the desulfurization efficiency is fitted, and the empirical formula is corrected for the operating data. The design parameters obtained by the new formula will be more accurate and more suitable for the desulfurization system of the unit after the ultra-low emission transformation. Improving the design accuracy is conducive to the design and development of the tray tower device.

[0041] In this embodiment, in step S1, preprocessing the key parameter data includes importing, arranging, and normalizing the parameter data.

[0042] Import the key parameters of the correction system from the DCS system, including the import SO 2 Concentration, inlet O 2Concentration, slurry circulation pump flow, desulfurization efficiency measurement value, flue gas inlet pressure and temperature, slurry pH value, etc. In order to enhance the calculation efficiency, each parameter is associated with its own time node, and different key parameters at the same time are combined into a data tuple. Subsequent cleaning operations are based on these tuple data, without repeating the data search work. The preprocessing steps are as follows:

[0043] First, in order to unify the dimensions of each parameter and simplify the calculation, the normalization method is used to preprocess the data of each parameter. Normalization calculation mainly includes the following:

[0044] 1. Initialization.

[0045]

[0046] 2. Averaging process.

[0047]

[0048] 3. Interpolation method.

[0049]

[0050] All the above methods are based on positive data, that is, the larger the value, the more beneficial it is to increase the system benefits. For negative data, after dimensionless transformation, the opposite number or reciprocal can be used instead.

[0051] Through the above data arrangement, normalization and other data preprocessing steps, the entry SO 2 Concentration, inlet O 2 The key parameters in the desulfurization system, such as concentration, slurry circulation pump flow, desulfurization efficiency measurement, flue gas inlet pressure and temperature, slurry pH value, etc., are converted into a set of numbers formed by unified dimensionless values, and the physical objects are initially digitized. The standardized parameter set helps to reduce the amount of calculation of the cleaning and screening module and the fitting module, and adapt to the big data environment.

[0052] In this embodiment, in step S2, data cleaning and data screening are performed on each data tuple in step S1, which specifically includes the following steps:

[0053] S21. Set upper and lower limits for each key parameter according to the operation mechanism of the desulfurization system, and eliminate obviously erroneous and meaningless data in each data tuple through the set parameter range;

[0054] S22. By using the grey correlation analysis method and the set correlation sequence threshold, the key parameters and the reference values ​​of the numerical relative influence coefficients corresponding to the key parameters are set as reference sequences, and each data tuple is screened and calculated to obtain each target tuple.

[0055] In this embodiment, data cleaning and screening mainly include data cleaning function and data screening function, and ultimately achieve the goals of outlier removal, range screening, etc.; the goal of data cleaning and screening is transformed from a single parameter to different time tuples.

[0056] According to the requirements, the benchmark value sequence should be set according to the operation level of the desulfurization system, which may include the inlet SO 2 The standard values ​​of key parameters in the desulfurization system, such as concentration, slurry circulation pump flow, flue gas inlet flow, slurry pH value, and the benchmark values ​​of the relative coefficients of each key parameter, are all processed and calculated based on the benchmark value sequence during the cleaning, screening and fitting process.

[0057] Data cleaning is to eliminate data outliers and improve the quality of calculation samples. According to the operation mechanism of the desulfurization system, upper and lower limits are set for each key parameter, and obviously wrong data and meaningless data (such as shutdown section system parameters) are eliminated through the set parameter range. When over-limit data appears in a time tuple, the tuple is judged as a "poor quality" tuple and is eliminated.

[0058] Data screening is to select target tuples according to the above-set benchmark value sequence to achieve the purpose of controlling variables. 2 When fitting and calculating the concentration influencing factors, filter out the inlet SO 2 The sample data with the concentration parameter closest to the reference value can be considered to have only one difference between the different tuples at the entrance SO 2 The concentration changes allow independent study of the inlet SO 2 The influence of concentration on desulfurization efficiency. This embodiment compares the similarity of different tuples using the grey correlation analysis method. The grey correlation analysis method is also a multi-factor, multi-target statistical analysis method. This method is based on the sample data of each factor. The corresponding indicators are obtained using the grey correlation to describe the strength, size and order of the relationship between each indicator. The correlation between indicators is measured according to the similarity or difference in the development trends between the indicators; the specific analysis steps of the grey correlation analysis method are as follows:

[0059] S31. Collect key parameter data and build the original matrix:

[0060]

[0061] Among them, m represents the number of parameters, X′ i =(x′ i (1),x′ i (2),…,x′ i (m)) T ,i=1,2,…,n。

[0062] S32. Determine the reference sequence. Note:

[0063] X′ 0 =(x′ 0 (1),x′ 0 (2),…,x′ 0 (m)) T

[0064] The reference sequence can be the optimal value or the worst value of each parameter, or other reference values ​​can be selected according to needs. In the process of parameter cleaning and screening of this model, the benchmark value sequence is used as the reference sequence.

[0065] S33. The initial parameter data is dimensionless, and the data sequence obtained after dimensionless transformation forms the following matrix:

[0066]

[0067] In this model, the initial parameter matrix has been dimensionally non-dimensionalized in step S1. It is only necessary to set a reference sequence and perform corresponding dimensionless operations on the reference sequence to complete steps S31 to S33.

[0068] S34, calculate the difference between each evaluation object and the reference index one by one, that is, |x 0 (k)-x i (k)|;

[0069] S35. Confirm and

[0070] S36. Calculate the correlation coefficient between each sequence and the benchmark value sequence:

[0071]

[0072] In the formula, ρ is the resolution coefficient, which takes values ​​within (0,1) and is used to increase the significance of differences between sequences. It is usually set to 0.5.

[0073] S37. Calculate the mean of the correlation coefficients between the comparison sequence and the reference value sequence respectively, and record the mean as the correlation sequence:

[0074]

[0075] Through the above steps, the similarity between different targets and the reference value sequence is quantified as a correlation order. The sequence with a larger correlation order has a higher similarity to the reference value sequence. The variable control is realized through the grey correlation analysis method, and the target tuple (or target sequence) of the fitting algorithm is obtained.

[0076] The data cleaning and screening process of this embodiment reflects the advantages of array tuples. During the cleaning and screening process, it is only necessary to record the time nodes of the target sequence and extract the corresponding parameter subsets from the initially acquired data, without the need to perform denormalization calculations in the cleaning and screening module. At the same time, there is no need to store a large amount of historical data, saving storage space.

[0077] Data cleaning and screening is a further preprocessing of the original data. Different from conventional data screening, data processing not only includes the removal of abnormal values, but also adds similarity sequence screening based on benchmark values. Through similarity screening calculation, not only can the purpose of controlling variables be achieved, but also the degree of influence of key parameters on desulfurization efficiency can be independently studied. At the same time, the current power plant operation data can be used in real time, which is flexible.

[0078] In this embodiment, step S3 obtains a training set by fitting each target tuple based on the set benchmark value and the completion model, and determines the empirical formula of the relative influence coefficient of each key parameter after correction through self-learning algorithm fitting, and determines the influence coefficient of each key parameter on the desulfurization efficiency, which specifically includes the following implementation process.

[0079] 1) First, the training set is obtained by fitting each target tuple based on the set benchmark value and the completion model, including:

[0080] The independent variables (key parameters) and relative influencing factors are calculated, and finally a sequence of independent variables (key parameters) and a sequence of relative influencing coefficients are obtained. 2 When fitting the concentration influencing factor, SO is calculated based on the data of the filtered tuple and the completion (C) model. 2 Relative influence coefficient of concentration And determine the corresponding entry SO 2 The corresponding independent variable sequence and influence coefficient sequence are calculated and used as the training set for the fitting process.

[0081] 2) Then, the empirical formula of the relative influence coefficient of each key parameter after correction is determined by fitting through the self-learning algorithm, and the influence coefficient of each key parameter on the desulfurization efficiency is determined, including the steps of:

[0082] Based on the training set determined above, the fitting process is implemented based on the least squares regression algorithm, which specifically includes the following steps:

[0083] For the input training set

[0084] D={(x 1 ,y 1 ),(x 2 ,y 2 ),…,(x m ,y m )}

[0085] Linear regression model builds a model that satisfies f(x i )=wX(x i )+b, and f(x i ) tends to y i , i = 1, 2...m, that is, the error is minimized, where w is the weight matrix to be solved and b is the bias coefficient.

[0086] Among them, the Euclidean distance is used as the error standard, and the solution model based on minimizing the mean square error is the least square method.

[0087] Then the regression problem can be simplified to a function group The constructed objective function S(x):

[0088]

[0089] For m groups of data, the error δ must be satisfied i =S * (x i )-y i Minimize the sum of squares:

[0090] Function Group Linearly independent, determined according to the characteristics of data changes, usually using polynomials

[0091] Based on the self-learning algorithm and on-site operation data, the empirical formula for pallet design was finally corrected, making it more suitable for ultra-low emissions and power production processes in a big data environment. The self-learning algorithm is used to fit the empirical formula, and a relatively accurate empirical formula curve can be obtained without calculating the relative coefficient value under each working condition.

[0092] The following is a specific example to illustrate the optimization design process of the desulfurization efficiency enhancement device implemented in the above embodiment.

[0093] Based on the above process, the entry SO 2 The relative influence coefficients of key parameters such as concentration, liquid-gas ratio, and slurry pH value were calculated and fitted.

[0094] For example, for the entry SO 2 Concentration influence coefficient Fitting calculation. The initially acquired data is preprocessed and normalized to obtain a tuple. First, set the entry SO 2 The standard values ​​of key parameters in the desulfurization system, such as concentration, slurry circulation pump flow, flue gas inlet flow, slurry pH value, and the standard values ​​of the corresponding relative influence coefficients, are then calculated using the grey correlation analysis method to remove the inlet SO 2The similarity between each parameter tuple and the reference value sequence after concentration is called the correlation sequence. The correlation sequence threshold is set artificially, and the sequence above this threshold is The target sequence of the fitting calculation. It can be understood that only the entry SO 2 Different concentrations can be independently imported for SO 2 Concentration influence coefficient Fitting calculations do not need to be disturbed by other parameters.

[0095] Finish After the target sequence of the fitting calculation is screened, the training set required for the fitting calculation is calculated according to the set benchmark value and the completion model, that is, the entry SO 2 Concentration series and relative influence coefficients Sequence, and based on the self-learning algorithm fitting to obtain the corrected entry SO 2 Concentration relative influence coefficient The empirical formula truly reflects SO 2 The intrinsic relationship between concentration and desulfurization rate.

[0096] Based on the above method and the data of long-term operation of the existing desulfurization system, the influence of various important parameters on the desulfurization efficiency is obtained, and the influence of various parameters (for example, inlet SO 2 The fitting empirical formula of the influence of key parameters such as concentration, liquid-gas ratio, slurry pH value on desulfurization efficiency is obtained. On this basis, the optimization design of desulfurization efficiency enhancement device is continued, such as Figure 3 shown.

[0097] After fitting various key parameters according to the self-learning algorithm, the system parameter requirements are obtained by combining the material balance calculation and the desulfurization system differential pressure calculation, and finally the pallet device design scheme is obtained.

[0098] The beneficial effects of this implementation method include:

[0099] a) Standardized and dimensionless processing of key parameters effectively improves processing and calculation speed.

[0100] b) Based on the similar sequence screening step of standard values, the number of variables is controlled according to the unit operation mechanism to reduce the dimension of the fitting formula.

[0101] c) The combination of intelligent algorithms and physical objects can be adapted to the operating environment of different units under different working conditions.

[0102] System Example

[0103] According to an embodiment of the present invention, a system for optimizing a desulfurization efficiency enhancement device is provided, such as Figure 4 As shown, it is a schematic diagram of the optimization system of the desulfurization efficiency-enhancing device of this embodiment. The optimization system of the desulfurization efficiency-enhancing device according to the embodiment of the present invention includes:

[0104] Data acquisition and preprocessing module: acquires the key parameter data required for correction in the desulfurization system in real time according to the DCS system, and constructs the initial data set of each parameter after preprocessing the data;

[0105] The key parameters of this embodiment include the entry SO 2 Concentration, inlet O 2 Concentration, slurry circulation pump flow, desulfurization efficiency measurement value, flue gas inlet pressure and temperature, slurry pH value, etc.

[0106] The module's preprocessing of key parameter data includes the following steps:

[0107] First, in order to unify the dimensions of each parameter and simplify the calculation, the normalization method is used to preprocess the data of each parameter. Normalization calculation mainly includes the following:

[0108] 1. Initialization.

[0109]

[0110] 2. Averaging process.

[0111]

[0112] 3. Interpolation method.

[0113]

[0114] All the above methods are based on positive data, that is, the larger the value, the more beneficial it is to increase the system benefits. For negative data, after dimensionless transformation, the opposite number or reciprocal can be used instead.

[0115] Through the above data arrangement, normalization and other data preprocessing steps, the entry SO 2 Concentration, inlet O 2 The key parameters in the desulfurization system, such as concentration, slurry circulation pump flow, desulfurization efficiency measurement, flue gas inlet pressure and temperature, slurry pH value, etc., are converted into a set of numbers formed by unified dimensionless values, and the physical objects are initially digitized. The standardized parameter set helps to reduce the amount of calculation of the cleaning and screening module and the fitting module, and adapt to the big data environment.

[0116] Data cleaning and screening module: It is used to clean each data tuple based on the data sequence obtained by preprocessing the key parameter data obtained by the data acquisition and preprocessing module, set the benchmark value of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, and use the grey correlation analysis method and the set correlation sequence threshold, with the set benchmark value of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, to screen and calculate each data tuple to obtain each target tuple.

[0117] Data cleaning and filtering mainly include data cleaning function and data filtering function, and the ultimate goal is to achieve the goals of outlier removal and range screening; the goal of data cleaning and filtering is transformed from a single parameter to different time tuples.

[0118] Data screening is to select target tuples according to the above-set benchmark value sequence to achieve the purpose of controlling variables. This embodiment uses the grey correlation analysis method to compare the similarity of different tuples. The specific analysis steps are as follows:

[0119] A1. Collect evaluation data and build the original matrix:

[0120]

[0121] Among them, m represents the number of indicators, X′ i =(x′ i (1),x′ i (2),…,x′ i (m)) T ,i=1,2,…,n。

[0122] A2. Determine the reference sequence, recorded as:

[0123] X′ 0 =(x′ 0 (1),x′ 0 (2),…,x′ 0 (m)) T

[0124] The reference sequence can be the optimal value or the worst value of each indicator, or other reference values ​​can be selected according to needs. In the process of parameter cleaning and screening of this model, the benchmark value sequence is used as the reference sequence.

[0125] A3. Non-dimensionalize the initial data. The data sequence obtained after non-dimensionalization forms the following matrix:

[0126]

[0127] A4. Calculate the difference between each evaluation object and the reference index one by one, that is, |x 0 (k)-x i (k)|.

[0128] A5. Confirm and

[0129] A6. Calculate the correlation coefficient between each sequence and the benchmark value sequence:

[0130]

[0131] In the formula, ρ is the resolution coefficient, which takes values ​​within (0,1) and is used to increase the significance of differences between sequences. It is usually set to 0.5.

[0132] A7. Calculate the mean of the correlation coefficients between the comparison sequence and the reference value sequence respectively, and record this mean as the correlation sequence:

[0133]

[0134] Through the above steps, the similarity between different targets and the benchmark value sequence is quantified as the correlation order. The sequence with a larger correlation order has a higher similarity with the benchmark value sequence. The variable control is realized through the grey correlation analysis method, and the target tuple of the fitting algorithm is obtained.

[0135] Fitting and correction module: It is used to obtain the training set by fitting each target tuple based on the set benchmark value and the completion model, and to determine the empirical formula of the relative influence coefficient of each key parameter after correction through self-learning algorithm fitting, so as to determine the influence coefficient of each key parameter on desulfurization efficiency.

[0136] Among them, the fitting correction module is built into the DCS system through PLC.

[0137] The fitting correction module includes a training set calculation submodule, which is used for key parameter calculation and relative influence factor calculation to obtain a sequence of key parameters and a sequence of relative influence coefficients as a training set.

[0138] Self-learning algorithm fitting submodule: Use the training set to determine the corrected empirical formula of the relative influence coefficient of each key parameter through self-learning algorithm fitting, and determine the influence coefficient of each key parameter on the desulfurization efficiency, including the following steps:

[0139] Based on the training set determined above, the fitting process is implemented based on the least squares regression algorithm, which specifically includes the following steps:

[0140] For the input training set

[0141] D={(x 1 ,y 1 ),(x 2 ,y 2 ),…,(x m ,y m )}

[0142] Linear regression model builds a model that satisfies f(x i )=wX(x i )+b, and f(x i ) tends to y i , i = 1, 2...m, that is, the error is minimized, where w is the weight matrix to be solved and b is the bias coefficient.

[0143] Among them, the Euclidean distance is used as the error standard, and the solution model based on minimizing the mean square error is the least square method.

[0144] Then the regression problem can be simplified to a function group The constructed objective function S(x):

[0145]

[0146] For m groups of data, the error δ must be satisfied i =S * (x i )-y i Minimize the sum of squares:

[0147] Function Group Linearly independent, determined according to the characteristics of data changes, usually using polynomials

[0148] Pallet device design scheme confirmation module: Based on the determined influence coefficients of each key parameter on the desulfurization efficiency, combined with the material balance calculation and the desulfurization system differential pressure calculation, the system parameter requirements are obtained, and finally the optimized pallet device design scheme is obtained.

[0149] The embodiment of the present invention is a system embodiment corresponding to the above-mentioned method embodiment. The specific operations of the processing steps of each module can be understood by referring to the description of the method embodiment, which will not be repeated here.

[0150] like Figure 5 As shown, the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the optimization method of the desulfurization efficiency enhancement device in the above embodiment is implemented, or when the computer program is executed by a processor, the optimization method of the desulfurization efficiency enhancement device in the above embodiment is implemented. When the computer program is executed by the processor, the following method steps are implemented:

[0151] Step S1, obtaining key parameter data required for correction in the desulfurization system in real time according to the DCS system, and after preprocessing the data, combining each key parameter into a data tuple;

[0152] The key parameters of this embodiment include the entry SO 2 Concentration, inlet O2 Concentration, slurry circulation pump flow, desulfurization efficiency measurement value, flue gas inlet pressure and temperature, slurry pH value, etc.

[0153] Step S2, performing data cleaning on each data tuple, setting each key parameter and the reference value of the relative influence coefficient corresponding to each key parameter as a reference sequence, using the grey correlation analysis method and the set correlation sequence threshold, taking the set key parameters and the reference value of the relative influence coefficient corresponding to each key parameter as the reference sequence, screening and calculating each data tuple to obtain each target tuple;

[0154] Step S3, fitting and calculating each target tuple based on the set benchmark value and the completion model to obtain a training set, and fitting and determining the revised empirical formula of the relative influence coefficient of each key parameter through a self-learning algorithm to determine the influence coefficient of each key parameter on the desulfurization efficiency;

[0155] Step S4: according to the influence coefficient of each key parameter on the desulfurization efficiency determined in step S3, the system parameter requirements are obtained by combining the material balance calculation and the desulfurization system differential pressure calculation, and finally an optimized tray device design scheme is obtained.

[0156] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0157] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0158] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing a desulfurization efficiency enhancement device, characterized in that: The following steps are involved: According to the DCS system, the key parameter data required for correction in the desulfurization system is obtained in real time, and after pre-processing the data, each key parameter is associated with its own time node, and different key parameters at the same time are combined into a data tuple; Key parameters include inlet SO2 concentration, inlet O2 concentration, slurry circulation pump flow, desulfurization efficiency measurement, flue gas inlet pressure and temperature, and slurry pH; Perform data cleaning on each data tuple, set the benchmark values ​​of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, and use the grey correlation analysis method and the set correlation sequence threshold, with the set benchmark values ​​of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, to screen and calculate each data tuple to obtain each target tuple; According to each key parameter in the target tuple, the relative influence coefficient corresponding to each key parameter is calculated by the completion model, so as to obtain each key parameter sequence and the relative influence coefficient sequence corresponding to each key parameter sequence, and use them as training sets. The corrected empirical formula of the relative influence coefficient of each key parameter is determined by fitting through the self-learning algorithm, and the influence coefficient of each key parameter on the desulfurization efficiency is determined; According to the influence coefficient of each key parameter on the desulfurization efficiency, the system parameter requirements are obtained by combining the material balance calculation and the desulfurization system differential pressure calculation, and finally the optimized tray device design scheme is obtained.

2. The optimization method of a desulfurization efficiency enhancement device according to claim 1, characterized in that: The data preprocessing includes the import, arrangement, normalization and dimensionless processing of each parameter data; and then each parameter is associated with its own time node, so that different key parameters at the same time are combined into a data tuple.

3. The optimization method of a desulfurization efficiency enhancement device according to claim 1, characterized in that: The data cleaning of each data tuple includes setting upper and lower limits for each key parameter according to the operation mechanism of the desulfurization system, and eliminating obviously erroneous and meaningless data of each data tuple through the set parameter range; Through the grey correlation analysis method and the set correlation sequence threshold, the key parameters and the benchmark values ​​of the relative influence coefficients corresponding to the key parameters are set as the reference sequence, and each data tuple is screened and calculated to obtain each target tuple.

4. The optimization method of a desulfurization efficiency enhancement device according to claim 1, characterized in that: The self-learning algorithm fitting uses a least squares regression algorithm.

5. An optimization system for a desulfurization efficiency enhancement device, characterized in that: include: Data acquisition and preprocessing module: According to the DCS system, the key parameter data required for correction in the desulfurization system are acquired in real time, and after preprocessing the data, the initial data set of each parameter is constructed; the key parameters include inlet SO2 concentration, inlet O2 concentration, slurry circulation pump flow, desulfurization efficiency measurement value, flue gas inlet pressure and temperature, and slurry pH value; Data cleaning and screening module: used for preprocessing the key parameter data acquired based on the data acquisition and preprocessing module, associating each key parameter with its own time node, combining different key parameters at the same time into a data tuple, performing data cleaning on each data tuple, setting the benchmark value of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, and using the grey correlation analysis method and the set correlation sequence threshold, taking the benchmark value of each key parameter and the relative influence coefficient corresponding to each key parameter as the reference sequence, screening and calculating each data tuple to obtain each target tuple; Fitting and correction module: used to calculate the relative influence coefficient corresponding to each key parameter through the completion model according to each key parameter in the target tuple, so as to obtain each key parameter sequence and the relative influence coefficient sequence corresponding to each key parameter sequence, and use it as a training set, and determine the empirical formula of the relative influence coefficient of each key parameter after correction through self-learning algorithm fitting, and determine the influence coefficient of each key parameter on desulfurization efficiency; Pallet device design scheme confirmation module: Based on the determined influence coefficients of each key parameter on the desulfurization efficiency, combined with the material balance calculation and the desulfurization system differential pressure calculation, the system parameter requirements are obtained, and finally the optimized pallet device design scheme is obtained.

6. The optimization system of a desulfurization efficiency enhancement device according to claim 5, characterized in that: The data acquisition and preprocessing module preprocesses the data including importing, arranging, normalizing and dimensionless processing of each parameter data; and then associating each parameter with its own time node, so that different key parameters at the same time are combined into a data tuple.

7. The optimization system of a desulfurization efficiency enhancement device according to claim 5, characterized in that: The data cleaning and screening module is specifically implemented as follows: Data cleaning of each data tuple includes setting upper and lower limits for each key parameter according to the operation mechanism of the desulfurization system, and eliminating obviously erroneous and meaningless data in each data tuple through the set parameter range; Through the grey correlation analysis method and the set correlation sequence threshold, the key parameters and the benchmark values ​​of the relative influence coefficients corresponding to the key parameters are set as the reference sequence, and each data tuple is screened and calculated to obtain each target tuple.

8. The optimization system of a desulfurization efficiency enhancement device according to claim 5, characterized in that: The fitting correction module includes The training set calculation submodule is used to calculate key parameters and relative influencing factors, and obtain a sequence of key parameters and a sequence of relative influencing coefficients as a training set; Self-learning algorithm fitting submodule: using the training set, the least squares regression algorithm is used to fit the empirical formula for the relative influence coefficient of each key parameter after correction, and the influence coefficient of each key parameter on the desulfurization efficiency is determined.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the optimization method of the desulfurization efficiency enhancement device according to any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the optimization method of the desulfurization efficiency enhancement device according to any one of claims 1 to 4 is implemented.