Method, apparatus, electronic device and storage medium for obtaining slurry density

Abstract

The present disclosure relates to a method, device, electronic device and storage medium for obtaining slurry density, and relates to the field of thermal power generation. First, a density measurement model is trained by a whale optimization algorithm according to preset training samples, then a target control parameter value of a target control parameter for generating limestone slurry is obtained, and finally the target control parameter value is input into the density measurement model to obtain the slurry density of the limestone slurry output by the density measurement model. This avoids the problem in the related art that when a densitometer is directly used to obtain the slurry density, the densitometer is prone to inaccurate measurement results due to wear. Instead, the slurry density is obtained by using the target control parameter value obtained from a stable measurement related to the slurry density, reducing the cost of regular replacement and maintenance of the densitometer, improving the accuracy of the slurry density result, and being beneficial to improving the effect of flue gas desulfurization.

Description

Technical Field

[0001] The present disclosure relates to the field of thermal power generation, and particularly to a method, apparatus, electronic device, and storage medium for obtaining slurry density. Background Art

[0002] In the field of thermal power generation, limestone-gypsum flue gas desulfurization is currently the most mature and widely used flue gas desulfurization technology. The quality of the limestone slurry obtained by the slurry preparation system has a great influence on the desulfurization effect. Therefore, it is very important to accurately measure the slurry density of the limestone slurry.

[0003] Currently, most thermal power plants use a densitometer to measure the slurry density. The densitometer mostly adopts a probe pipeline installation structure. Due to the high viscosity of the slurry, the probe of the densitometer is easily eroded and worn by the slurry, resulting in errors in the measurement results of the slurry density. Summary of the Invention

[0004] To overcome the problems existing in the related art, the present disclosure provides a method, apparatus, electronic device, and storage medium for obtaining slurry density.

[0005] According to a first aspect of an embodiment of the present disclosure, a method for obtaining slurry density is provided. The method includes:

[0006] Training a density measurement model through a whale optimization algorithm according to a preset training sample. The elements of the population members of the whale optimization algorithm are the model parameters of the density measurement model. The initial element values of the elements of each population member in the whale optimization algorithm are obtained through Logistic chaotic mapping. The preset training sample includes the training control parameter values of multiple training control parameter groups and the training output parameter values of the corresponding training output parameters. Each training control parameter group includes multiple training control parameters. The training output parameter includes the slurry density;

[0007] Obtaining the target control parameter value of the target control parameter for generating the limestone slurry;

[0008] Inputting the target control parameter value into the density measurement model to obtain the slurry density of the limestone slurry output by the density measurement model.

[0009] Optionally, the density measurement model is a neural network model. The training the density measurement model through a whale optimization algorithm according to the preset training sample includes:

[0010] Determining the whale optimization algorithm parameters of the whale optimization algorithm according to the model parameters of the preset neural network model. The whale optimization algorithm parameters include the elements of the population members, the preset element value threshold of each element value, and the whale population size;

[0011] Obtain the current element value of each population member of the whale optimization algorithm through the Logistic chaotic mapping according to the whale optimization algorithm parameters and the preset chaotic mapping parameters, where the chaotic mapping parameters include a chaotic mapping coefficient;

[0012] Obtain the target model parameter value of the preset neural network model by using the whale optimization algorithm according to the whale optimization algorithm parameters, the current element value of each population member, and the preset training samples, and use the obtained target neural network model as the density measurement model.

[0013] Optionally, determining the whale optimization algorithm parameters of the whale optimization algorithm according to the model parameters of the preset neural network model includes:

[0014] Use the training control parameter as the input node of the preset neural network model, use the slurry density as the output node of the preset neural network model, determine the model parameters of the preset neural network model according to the input node, the output node, and the hidden neurons of the preset neural network model, and use the model parameters as the elements of the population members. The model parameters include the first connection weight, the first threshold of the hidden neuron, the second connection weight, and the second threshold of the output node. The first connection weight represents the weight between the input node and the hidden neuron, and the second connection weight represents the weight between the hidden neuron and the output node;

[0015] Determine the element value range of each element according to the model parameter value range of each preset model parameter;

[0016] Determine the whale population size according to the number of model parameter values and the expected time of model training.

[0017] Optionally, obtaining the target model parameter value of the preset neural network model by using the whale optimization algorithm according to the whale optimization algorithm parameters, the current element value of each population member, and the preset training samples, and using the obtained target neural network model as the density measurement model includes:

[0018] Obtain the fitness value of each population member according to the preset training samples. The fitness value is the mean square error between the predicted output parameter value and the corresponding training output parameter value. The predicted output parameter value is obtained through the preset neural network model according to the training control parameter value of each training control parameter group and the current element value of the population member;

[0019] Obtain the target element value of the target population member of the population according to the fitness value of each population member, where the target population member is the population member with the smallest fitness value in the population;

[0020] Perform the whale optimization algorithm to surround, prey on, and search for the target element value according to the whale optimization algorithm parameters, so as to update the current element value of each population member. Repeat the steps of obtaining the fitness value of each population member according to the preset training samples and obtaining the target element value of the target population member of the population according to the fitness value of each population member, so as to obtain the optimal element value corresponding to the optimal population member with the smallest fitness value in the population when the preset end condition is satisfied. Take the optimal element value as the target model parameter value of the preset neural network model, and take the obtained target neural network model as the density measurement model.

[0021] Optionally, the preset end condition includes that the fitness value corresponding to the optimal population member is less than or equal to a preset fitness value threshold, and / or the number of training times is greater than or equal to a preset training times threshold.

[0022] Optionally, the method further includes:

[0023] Obtain multiple groups of historical data of the limestone slurry, where the historical data includes the alternative control parameter values of multiple alternative control parameter groups and the output parameter values of the corresponding output parameters. Each alternative control parameter group includes multiple alternative control parameters of the limestone slurry, and the output parameter includes the slurry density;

[0024] Determine the target control parameter from the multiple alternative control parameters by using the grey relational analysis method according to the multiple groups of historical data;

[0025] Obtain the preset training samples from the multiple groups of historical data according to the target control parameter.

[0026] Optionally, the determining the target control parameter from the multiple alternative control parameters by using the grey relational analysis method according to the multiple groups of historical data includes:

[0027] Obtain the grey relational degree between the alternative control parameter value of each alternative control parameter and the corresponding output parameter value according to the historical data;

[0028] Determine the target control parameter from the multiple alternative control parameters according to the grey relational degree and a preset grey relational degree threshold, where the grey relational degree of the target control parameter is greater than or equal to the preset grey relational degree threshold.

[0029] Optionally, the method further includes:

[0030] After obtaining multiple groups of historical data of the limestone slurry, abnormal historical data is deleted from the multiple groups of historical data according to a plurality of preset control parameter value ranges, where the abnormal historical data indicates that the alternative control parameter value of any one of the alternative control parameters in the alternative control parameter group exceeds the corresponding control parameter value range.

[0031] Optionally, the target control parameter and the training control parameter include the cyclone inlet pressure, the instantaneous feed rate, the instantaneous dilution water flow rate, and the instantaneous grinding water flow rate.

[0032] According to a second aspect of the embodiments of the present disclosure, there is provided an apparatus for obtaining the slurry density, the apparatus including:

[0033] A model training module configured to train a density measurement model through a whale optimization algorithm according to a preset training sample, where the elements of the population members of the whale optimization algorithm are the model parameters of the density measurement model, and the initial element values of the elements of each population member in the whale optimization algorithm are obtained through a Logistic chaotic mapping. The preset training sample includes the training control parameter values of multiple training control parameter groups and the training output parameter values of the corresponding training output parameters. Each training control parameter group includes multiple training control parameters, and the training output parameter includes the slurry density;

[0034] A parameter acquisition module configured to acquire the target control parameter values of the target control parameters for generating the limestone slurry;

[0035] A result acquisition module configured to input the target control parameter values into a pre-trained density measurement model to obtain the slurry density of the limestone slurry output by the density measurement model.

[0036] Optionally, the density measurement model is a neural network model, and the model training module is further configured to:

[0037] Determine the whale optimization algorithm parameters of the whale optimization algorithm according to the model parameters of the preset neural network model. The whale optimization algorithm parameters include the elements of the population members, the preset element value thresholds of each element value, and the whale population size;

[0038] Obtain the current element values of each population member of the whale optimization algorithm through the Logistic chaotic mapping according to the whale optimization algorithm parameters and the preset chaotic mapping parameters. The chaotic mapping parameters include the chaotic mapping coefficient;

[0039] Using the whale optimization algorithm, obtain the target model parameter values of the preset neural network model according to the whale optimization algorithm parameters, the current element values of each population member, and the preset training samples, and use the obtained target neural network model as the density measurement model.

[0040] Optionally, the model training module is further configured to:

[0041] Use the training control parameters as the input nodes of the preset neural network model, use the slurry density as the output node of the preset neural network model, determine the model parameters of the preset neural network model according to the input nodes, the output nodes, and the hidden neurons of the preset neural network model, and use the model parameters as the elements of the population members. The model parameters include the first connection weight, the first threshold of the hidden neuron, the second connection weight, and the second threshold of the output node. The first connection weight represents the weight between the input node and the hidden neuron, and the second connection weight represents the weight between the hidden neuron and the output node;

[0042] Determine the element value range of each element according to the model parameter value range of each preset model parameter;

[0043] Determine the whale population size according to the number of model parameter values and the expected time of model training.

[0044] Optionally, the model training module is further configured to:

[0045] Obtain the fitness value of each population member according to the preset training samples. The fitness value is the mean square error between the predicted output parameter value and the corresponding training output parameter value. The predicted output parameter value is obtained through the preset neural network model according to the training control parameter values of each training control parameter group and the current element values of the population members;

[0046] Obtain the target element value of the target population member of the population according to the fitness value of each population member. The target population member is the population member with the smallest fitness value in the population;

[0047] According to the parameters of the whale optimization algorithm, the whale optimization algorithm is adopted to complete the enclosure, predation and search of the target element value, so as to update the current element value of each population member. The steps of obtaining the fitness value of each population member according to the preset training samples and obtaining the target element value of the target population member of the population according to the fitness value of each population member are repeated, so that when the preset end condition is met, the optimal element value corresponding to the optimal population member with the smallest fitness value in the population is obtained. The optimal element value is used as the target model parameter value of the preset neural network model, and the obtained target neural network model is used as the density measurement model.

[0048] Optionally, the device further includes a training sample acquisition module, which is configured to:

[0049] Obtain multiple groups of historical data of the limestone slurry, where the historical data includes the alternative control parameter values of multiple alternative control parameter groups and the output parameter values of the corresponding output parameters. Each alternative control parameter group includes multiple alternative control parameters of the limestone slurry, and the output parameter includes the slurry density;

[0050] Determine the target control parameter from the multiple alternative control parameters by using the grey relational analysis method according to the multiple groups of historical data;

[0051] Obtain the preset training samples from the multiple groups of historical data according to the target control parameter.

[0052] Optionally, the training sample acquisition module is further configured to:

[0053] Obtain the grey relational degree between the alternative control parameter value of each alternative control parameter and the corresponding output parameter value according to the historical data;

[0054] Determine the target control parameter from the multiple alternative control parameters according to the grey relational degree and a preset grey relational degree threshold, and the grey relational degree of the target control parameter is greater than or equal to the preset grey relational degree threshold.

[0055] Optionally, the training sample acquisition module is further configured to:

[0056] After obtaining multiple groups of historical data of the limestone slurry, delete abnormal historical data from the multiple groups of historical data according to a preset range of multiple control parameter values. The abnormal historical data indicates that the alternative control parameter value of any alternative control parameter in the alternative control parameter group exceeds the corresponding control parameter value range.

[0057] According to the third aspect of the embodiments of the present disclosure, an electronic device is provided. The electronic device includes:

[0058] A memory storing a computer program thereon;

[0059] A processor for executing the computer program in the memory to implement the steps of the method according to any one of the embodiments in the first aspect above.

[0060] According to a fourth aspect of the embodiments of the present disclosure, a non - transitory computer - readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the steps of the method described in the first aspect above are implemented.

[0061] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects:

[0062] The present disclosure first trains a density measurement model through a whale optimization algorithm according to preset training samples, then obtains the target control parameter value of the target control parameter for generating limestone slurry, and finally inputs the target control parameter value into the density measurement model to obtain the slurry density of the limestone slurry output by the density measurement model. This avoids the problem in the related art that the density meter is prone to inaccurate measurement results due to wear when directly obtaining the slurry density, and instead obtains the slurry density by using the target control parameter value obtained from a stable measurement related to the slurry density, reducing the cost of regular replacement and maintenance of the density meter, improving the accuracy of the slurry density result, and being beneficial to improving the effect of flue gas desulfurization.

[0063] It should be understood that the above general description and subsequent detailed description are only exemplary and explanatory, and cannot limit the present disclosure.

[0064] Other features and advantages of the present disclosure will be described in detail in the subsequent specific implementation part. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure, but do not constitute a limitation to the present disclosure.

[0066] Figure 1 is a flowchart of a method for obtaining slurry density shown according to an exemplary embodiment.

[0067] Figure 2 is a flowchart of another method for obtaining slurry density shown according to an exemplary embodiment.

[0068] Figure 3 is a flowchart of yet another method for obtaining slurry density shown according to an exemplary embodiment.

[0069] Figure 4It is a flowchart of yet another method for obtaining slurry density shown according to an exemplary embodiment.

[0070] Figure 5 It is a block diagram of a device for obtaining slurry density shown according to an exemplary embodiment.

[0071] Figure 6 It is a block diagram of another device for obtaining slurry density shown according to an exemplary embodiment.

[0072] Figure 7 It is a block diagram of an electronic device shown according to an exemplary embodiment. Detailed implementation manners

[0073] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims. It should be understood that the specific implementation manners described herein are only used to illustrate and explain the present disclosure and are not used to limit the present disclosure.

[0074] It should be noted that all actions of obtaining signals, information, or data in this application are carried out on the premise of complying with the corresponding data protection regulations and policies of the country where the location is located and obtaining the authorization given by the owner of the corresponding device.

[0075] First, the application scenario of the present disclosure will be described. The present disclosure can be applied to the field of thermal power generation. Limestone-gypsum flue gas desulfurization is the mainstream flue gas desulfurization technology in current thermal power plants. The quality of the limestone slurry obtained through the slurry preparation system has a great impact on the desulfurization effect. Therefore, it is very important to accurately measure the slurry density of the limestone slurry.

[0076] Currently, most thermal power plants use density meters to measure the slurry density. The density meters mostly have a probe pipeline installation structure. Due to the high viscosity of the slurry, the probes of the density meters are easily eroded and worn by the slurry, resulting in errors in the measurement results of the slurry density.

[0077] The inventors noticed that many control parameters of the slurry preparation system, such as the inlet pressure of the cyclone station, the instantaneous feeding amount, etc., are closely related to the final slurry density. Therefore, accurate slurry density can be obtained based on the control parameter values of these control parameters through deep learning methods. Therefore, the present disclosure provides a method for obtaining slurry density, which uses the control parameter values obtained from stable measurements related to the slurry density to obtain the slurry density, reduces the cost of regular replacement and maintenance of the density meter, improves the accuracy of the slurry density result, and is beneficial to improving the effect of flue gas desulfurization.

[0078] The following describes the present disclosure in conjunction with specific embodiments.

[0079] Figure 1 is a flowchart of a method for obtaining slurry density shown according to an exemplary embodiment. As Figure 1 shown, the method may include the following steps:

[0080] In step S101, a density measurement model is trained through the whale optimization algorithm according to a preset training sample.

[0081] Among them, the preset training sample includes the training control parameter values of multiple training control parameter groups and the training output parameter values of the corresponding training output parameters. Each training control parameter group includes multiple training control parameters. The training output parameter includes the slurry density. The density measurement model can be a supervised deep learning model. For example, it can be a neural network model. The density measurement model is obtained by training with the preset training sample. The selection of the specific model in the present disclosure is not limited.

[0082] This whale optimization algorithm is also known as WOA (Whale Optimization Algorithm). It is an algorithm that simulates the spiral bubble net feeding strategy of whales to optimize preset parameters, and can effectively overcome the problems of slow convergence speed and easy to fall into local extrema of traditional neural network models.

[0083] In the whale optimization algorithm, each whale population can be considered as the value of a model parameter vector of a density measurement model. In order to improve the quality of the initial value, the whale population randomly traverses the value space of the model parameter vector, and the initial element values of the elements of each population member in each whale optimization algorithm are obtained through Logistic chaotic mapping.

[0084] Exemplarily, the density measurement model can be a neural network model. Figure 2 is a flowchart of another method for obtaining slurry density shown according to an exemplary embodiment. As Figure 2 shown, the method may include the following steps:

[0085] In step S1011, whale optimization algorithm parameters are determined according to the model parameters of a preset neural network model.

[0086] Among them, the whale optimization algorithm parameters include elements of population members, a preset element value threshold for each element value, and the size of the whale population.

[0087] In some possible implementation manners, the whale optimization algorithm parameters are determined according to the model parameters of the preset neural network model through the following steps:

[0088] Step 1: Use the training control parameter as the input node of the preset neural network model, use the slurry density as the output node of the preset neural network model, and determine the model parameters of the preset neural network model according to the input node, the output node, and the hidden neurons of the preset neural network model, and use the model parameters as the elements of the population members.

[0089] Among them, the model parameters include the first connection weight, the first threshold of the hidden neuron, the second connection weight, and the second threshold of the output node. The first connection weight represents the weight between the input node and the hidden neuron, and the second connection weight represents the weight between the hidden neuron and the output node.

[0090] In some possible implementation manners, the upper limit of the number of hidden neurons can be determined through the following formula (1), and then, with the upper limit of the hidden neurons as the number of loops, multiple neural networks are constructed. Using preset training samples, the number of hidden neurons with the smallest mean square error is selected from the multiple neural networks. For specific reference to the technical description related to neural networks, this disclosure will not elaborate further. It is also possible to use a preset number of hidden layer nodes, and this disclosure does not limit this.

[0091] (Formula 1)

[0092] Among them, is the upper limit of the number of hidden neurons, m and n are the numbers of input nodes and output nodes respectively, and a is generally a preset integer with a value range of 1 to 10. For example, the value can be 5.

[0093] After determining the number of hidden neurons of the preset neural network model through the above method, the number of model parameters of the preset neural network model can be determined through the following formula (2) according to the number of input nodes, the number of output nodes, and the number of hidden neurons.

[0094] (Formula 2)

[0095] Among them, K is the number of model parameters of the preset neural network model, that is, the number of elements of the population members in the whale optimization algorithm, is the number of input nodes, is the number of hidden neurons, and is the number of output nodes. The model parameters of the preset neural network model are used as the elements of the population members.

[0096] Exemplarily, when there are 4 input nodes, 1 output node, and 5 hidden layer neurons, the model parameters of the preset neural network model include 20 first connection weights, 5 first thresholds, 5 second connection weights, and 1 second threshold. The number of the model parameters is 31, that is, the number of elements in the whale population members is 31.

[0097] Step 2: Determine the value range of each element according to the value range of each model parameter of the preset.

[0098] Exemplarily, according to the value range of each model parameter of the preset (such as the upper and lower limits of the first connection weight, the upper and lower limits of the first threshold, the upper and lower limits of the second connection weight, and the upper and lower limits of the second threshold), determine the value range of each corresponding element in the population members. This is convenient for determining the search space of each element of the population members in the whale optimization algorithm.

[0099] It should be noted that the upper and lower limits of the model parameter values of the above model parameters can be the same or different. The same upper and lower limits indicate that the value range of the model parameter is a preset fixed value, and the present disclosure does not limit this.

[0100] Step 3: Determine the whale population size according to the number of model parameter values and the expected time of model training.

[0101] The larger the population size, the more dispersed the distribution in the entire global range, the larger the search space range, and the easier it is to find the global optimal solution. However, the larger the population size, the longer the model training time. In some possible implementation manners, the size of the whale population can be determined according to the number of model parameters of the preset neural network model and the expected time of model training. For specific details, please refer to the relevant descriptions in the related technologies regarding the determination of the whale population size.

[0102] In another possible implementation manner, the preset whale population size can also be used as the whale population size. For example, the preset whale population size can be 30, and the present disclosure does not limit this.

[0103] In step S1012, according to the whale optimization algorithm parameters and the preset chaotic mapping parameters, obtain the current element value of each population member of the whale optimization algorithm through the Logistic chaotic mapping.

[0104] Among them, the chaotic mapping parameters include the chaotic mapping coefficient.

[0105] In some possible implementation manners, the current element value of each population member of the whale optimization algorithm can be determined through the following steps.

[0106] First, obtain the chaotic mapping coefficient of each element of each population member according to the following formula three.

[0107] (Formula Three)

[0108] Wherein, is the chaotic mapping coefficient of the i-th element in the j-th population member, is a random number randomly generated within the range of (0, 1), is the chaos coefficient, which is a preset constant. For example, it can be 4. The value range of j is [1, N - 1], where N is the population size, that is, the number of population members. The value range of i is [1, K], where K is the number of elements in the population member.

[0109] Then, obtain the current element value of each element in each population member according to the chaotic mapping coefficient and the following formula four.

[0110] (Formula Four)

[0111] Wherein, is the chaotic mapping coefficient of the i-th element in the j-th population member, is the current element value of the i-th element in the j-th population member, and are respectively the upper limit and the lower limit of the element value range of the i-th element in the population member, is greater than or equal to , the value range of j is [1, N], where N is the population size, that is, the number of population members. The value range of i is [1, K], where K is the number of elements in the population member.

[0112] In step S1013, according to the whale optimization algorithm parameters, the current element value of each population member, and the preset training samples, use the whale optimization algorithm to obtain the target model parameter value of the preset neural network model, and use the obtained target neural network model as the density measurement model.

[0113] Exemplarily, in step S1011, the model structure of the preset neural network model has been determined. In this step, the target model parameter value of the preset neural network model can be obtained by using the whale optimization algorithm according to the whale optimization algorithm parameters, the current element value of each population member, and the preset training samples through the following steps, so as to obtain the target neural network model, and use this target neural network model as the density measurement model.

[0114] Step 1: Obtain the fitness value of each population member according to the preset training samples.

[0115] Among them, the fitness value is the mean square error of the predicted output parameter value and the corresponding training output parameter value. The predicted output parameter value is obtained through a preset neural network model according to the training control parameter value of each training control parameter group and the current element value of the population member.

[0116] In some possible implementation manners, the current element value is used as the model parameter of the preset neural network model, and the training control parameter value is substituted into the preset neural network model to obtain the predicted output parameter value corresponding to the training control parameter value.

[0117] Exemplarily, the fitness value can be determined according to the predicted output parameter value, the training output parameter value, and the following formula five.

[0118] (Formula Five)

[0119] Among them, is the fitness value of the i-th population member, is the predicted output parameter value obtained by substituting the k-th training control parameter value into the preset neural network model formed by the i-th population member, is the training output parameter value corresponding to the k-th training control parameter value, and T is the number of training samples in the preset training samples.

[0120] Step 2: Obtain the target element value of the target population member of the population according to the fitness value of each population member.

[0121] Among them, the target population member is the population member with the smallest fitness value in the population.

[0122] Exemplarily, after obtaining the fitness value of each population member, the population member with the smallest fitness value can be used as the target population member, and the element value corresponding to the target population member is the target element value.

[0123] Step 3: Complete the encirclement, predation, and search of the target element value by adopting the whale optimization algorithm according to the whale optimization algorithm parameters to update the current element value of each population member, and repeat the steps of obtaining the fitness value of each population member according to the preset training samples and obtaining the target element value of the target population member of the population according to the fitness value of each population member, so as to obtain the optimal element value corresponding to the optimal population member with the smallest fitness value in the population when the preset end condition is satisfied, use the optimal element value as the target model parameter value of the preset neural network model, and use the obtained target neural network model as the density measurement model.

[0124] The encircling and hunting in the whale optimization algorithm mimics the way of using a spiral bubble net during the whale hunting process to approach the optimal solution. The value of this target element is the current optimal solution in this round of hunting. Each population member can perform a shrinking encirclement on this current optimal solution, or can swim around the prey in a continuously shrinking circle in a spiral update manner while swimming along a spiral path. The specific mathematical model of the encircling and hunting can refer to the description of the whale optimization algorithm in the related technology, which will not be elaborated in this disclosure. In addition to the above encircling and hunting process, the population members can also search for prey randomly, that is, use a scheme of randomly selecting individuals for position update, thereby improving the global optimization ability of the WOA algorithm and enabling the whale algorithm to have the ability to jump out of the local optimal solution. The specific mathematical model of the random search can refer to the description of the whale optimization algorithm in the related technology.

[0125] In the above encircling and hunting and searching processes of the population members, by training and updating the current element value of each population member, the steps of repeatedly obtaining the fitness value of each population member according to the preset training samples and obtaining the target element value of the target population member of the population according to the fitness value of each population member are repeated. During this process, the target population member and the corresponding target element value will be continuously optimized and updated, and the fitness value will continuously decrease accordingly. When the preset end condition is satisfied, the optimal element value corresponding to the optimal population member with the smallest fitness value in the population is obtained as the target model parameter value of the preset neural network model, and the obtained target neural network model is used as the density measurement model.

[0126] In some embodiments, the preset end condition may include that the fitness value corresponding to the optimal population member is less than or equal to the preset fitness value threshold, and / or the number of training times is greater than or equal to the preset number of training times threshold.

[0127] In another embodiment, the value range of j in Formula III can also be [1, k*N - 1]. Correspondingly, the value range of j in Formula IV can be [1, k*N], where k can be a natural number greater than 1. For example, k can be 2. In this way, through the above Formula III and Formula IV, the current element values of k*N population members can be obtained, and the N population members with the smallest fitness can be selected from the k*N population members. The density measurement model can be obtained by using the whale optimization algorithm for the screened population members, which can further improve the efficiency of the whale optimization algorithm.

[0128] In step S102, the target control parameter value of the target control parameter for generating the limestone slurry is obtained.

[0129] Exemplarily, the target control parameter value of the target control parameter for generating the limestone slurry can be collected by corresponding multiple sensors.

[0130] In step S103, the target control parameter value is input into the density measurement model to obtain the slurry density of the limestone slurry output by the density measurement model.

[0131] Exemplarily, the target control parameter value is input into the density measurement model trained in step S101, so as to obtain the slurry density of the limestone slurry output by the density measurement model.

[0132] Through the above solution, it is possible to avoid the problem in the related art that when directly obtaining the slurry density by using a densitometer, the densitometer is prone to inaccurate measurement results due to wear. Instead, the slurry density is obtained by using the target control parameter value obtained from a stable measurement related to the slurry density, which reduces the cost of regularly replacing and maintaining the densitometer, improves the accuracy of the slurry density result, and is beneficial to improving the effect of flue gas desulfurization.

[0133] Multiple control parameters of the slurry preparation system will affect the final slurry density (for example, the instantaneous feed rate of the weighing feeder, the instantaneous flow rate of the grinding water of the wet ball mill, the density of the finished slurry tank, the instantaneous flow rate of the dilution water, the level of the recirculation tank, the phase A current of the ball mill, the outlet pressure of the recirculation pump, the level of the finished slurry tank, the opening of the dilution water regulating valve, the opening of the grinding water regulating valve, and the inlet pressure of the cyclone station). If multiple control parameters are used as the input parameters of the density measurement model, it will lead to a large amount of calculation for training the density measurement model and affect the efficiency of obtaining the density measurement model.

[0134] In another embodiment, the target control parameter that has the greatest influence on the slurry density can be selected from multiple alternative control parameters, so as to improve the efficiency of obtaining the density measurement model.

[0135] Figure 3 is a flowchart of a method for obtaining slurry density shown according to an exemplary embodiment, as Figure 3 shown, the method may further include the following steps:

[0136] In step S104, multiple groups of historical data of the limestone slurry are obtained.

[0137] Among them, the historical data includes the alternative control parameter values of multiple alternative control parameter groups and the output parameter values of the corresponding output parameters. Each alternative control parameter group includes multiple alternative control parameters of the limestone slurry, and the output parameter includes the slurry density.

[0138] The manner of obtaining the multiple groups of historical data can refer to the implementation of sensors and information collectors in the related art, and the present disclosure will not elaborate.

[0139] In step S105, the target control parameter is determined from multiple alternative control parameters according to the multiple groups of historical data by using the grey relational analysis method.

[0140] Grey Relational Analysis is a method for measuring the degree of association between factors based on the similarity or dissimilarity of the development trends between factors.

[0141] Exemplarily, the grey relational degree of each alternative control parameter in historical data can be obtained in the following manner, and the target control parameter can be determined from multiple alternative control parameters.

[0142] First, obtain the grey relational degree between the alternative control parameter value of each alternative control parameter and the corresponding output parameter value according to the historical data.

[0143] In some possible implementation manners, the grey relational degree of each alternative control parameter can be obtained through the following formula six.

[0144] (Formula six)

[0145] Among them, s and are intermediate variables, represents the value of the j-th alternative control parameter in the i-th historical data, is the output parameter value in the i-th historical data, N is the number of historical data, is the grey relational degree between the j-th alternative control parameter and the output parameter. To avoid inaccurate calculation of the grey relational degree due to the difference in absolute value taking of the alternative control parameter value and the output parameter value, the above and both need to be dimensionless processed. The example given in formula six is to perform dimensionless processing by dividing the alternative control parameter value by the first term of the alternative control parameter value. The present disclosure does not limit the processing manner.

[0146] Then, determine the target control parameter from multiple alternative control parameters according to the grey relational degree and a preset grey relational degree threshold.

[0147] Among them, the grey relational degree of the target control parameter is greater than or equal to the preset grey relational degree threshold.

[0148] Exemplarily, an alternative control parameter with a grey relational degree greater than or equal to the preset grey relational degree threshold can be selected as the target control parameter. Exemplarily, the preset grey relational degree threshold can be 75%.

[0149] In another possible implementation manner, multiple alternative control parameters with the largest grey relational degrees can also be selected as the target control parameters. Exemplarily, 4 alternative control parameters with the largest grey relational degrees can be selected as the target control parameters.

[0150] In step S106, obtain preset training samples from multiple groups of historical data according to the target control parameter.

[0151] Through the above solution, the model can be simplified as much as possible while ensuring the accuracy of the density measurement model, reducing the computational complexity of model training, and improving the efficiency of model training.

[0152] Figure 4 is a flowchart of a method for obtaining the density of slurry shown according to an exemplary embodiment. As Figure 4 shown, the method may further include the following steps:

[0153] In step S107, after obtaining multiple groups of historical data of limestone slurry, abnormal historical data is deleted from the multiple groups of historical data according to a preset range of multiple control parameter values.

[0154] Among them, the abnormal historical data represents that the alternative control parameter value of any alternative control parameter in the alternative control parameter group exceeds the corresponding control parameter value range.

[0155] Through the above solution, it is possible to screen the historical data, delete the abnormal historical data in the historical data that exceeds the preset control parameter value range, further improve the reliability of the preset training samples, and improve the efficiency of model training.

[0156] In another embodiment, the target control parameter and the training control parameter include the inlet pressure of the cyclone station, the instantaneous feed rate, the instantaneous flow rate of dilution water, and the instantaneous flow rate of grinding water.

[0157] The method for obtaining the multiple groups of historical data may refer to the implementation of sensors and information collectors in related technologies, which will not be elaborated in this disclosure.

[0158] Through the above solution, the model can be simplified as much as possible while ensuring the accuracy of the density measurement model, reducing the computational complexity of model training, and improving the efficiency of model training.

[0159] Figure 5 is a block diagram of a device 500 for obtaining the density of slurry shown according to an exemplary embodiment. As Figure 5 shown, the device 500 for obtaining the density of slurry includes:

[0160] A model training module 501, configured to train a density measurement model through a whale optimization algorithm according to a preset training sample. The elements of the population members of the whale optimization algorithm are the model parameters of the density measurement model. The initial element values of the elements of each population member in the whale optimization algorithm are obtained through Logistic chaotic mapping. The preset training sample includes the training control parameter values of multiple training control parameter groups and the training output parameter values of the corresponding training output parameters. Each training control parameter group includes multiple training control parameters of limestone slurry, and the training output parameter includes the slurry density;

[0161] A parameter acquisition module 502, configured to acquire a target control parameter value of a target control parameter for generating limestone slurry;

[0162] A result acquisition module 503, configured to input the target control parameter value into a pre-trained density measurement model to obtain the slurry density of the limestone slurry output by the density measurement model.

[0163] Optionally, the density measurement model is a neural network model, and the model training module 501 is further configured to:

[0164] Determine whale optimization algorithm parameters of the whale optimization algorithm according to the model parameters of the preset neural network model, where the whale optimization algorithm parameters include elements of the population members, a preset element value threshold for each element value, and the whale population size;

[0165] Obtain the current element value of each population member of the whale optimization algorithm through Logistic chaotic mapping according to the whale optimization algorithm parameters and the preset chaotic mapping parameters, where the chaotic mapping parameters include a chaotic mapping coefficient;

[0166] Obtain a target model parameter value of the preset neural network model by using the whale optimization algorithm according to the whale optimization algorithm parameters, the current element value of each population member, and the preset training samples, and use the obtained target neural network model as the density measurement model.

[0167] Optionally, the model training module 501 is further configured to:

[0168] Use the training control parameter as the input node of the preset neural network model, use the slurry density as the output node of the preset neural network model, determine the model parameters of the preset neural network model according to the input node, the output node, and the hidden neurons of the preset neural network model, and use the model parameters as the elements of the population members. The model parameters include a first connection weight, a first threshold of the hidden neuron, a second connection weight, and a second threshold of the output node. The first connection weight represents the weight between the input node and the hidden neuron, and the second connection weight represents the weight between the hidden neuron and the output node;

[0169] Determine the element value range of each element according to the model parameter value range of each model parameter;

[0170] Determine the whale population size according to the number of model parameter values and the expected time of model training.

[0171] Optionally, the model training module 501 is further configured to:

[0172] Obtain the fitness value of each population member according to the preset training samples. The fitness value is the mean square error between the predicted output parameter value and the corresponding training output parameter value. The predicted output parameter value is obtained through a preset neural network model according to the training control parameter value of each training control parameter group and the current element value of the population member.

[0173] Obtain the target element value of the target population member of the population according to the fitness value of each population member. The target population member is the population member with the smallest fitness value in the population.

[0174] Adopt the whale optimization algorithm according to the whale optimization algorithm parameters to complete the encirclement, predation and search of the target element value, so as to update the current element value of each population member. Repeat the steps of obtaining the fitness value of each population member according to the preset training samples and obtaining the target element value of the target population member of the population according to the fitness value of each population member, so as to obtain the optimal element value corresponding to the optimal population member with the smallest fitness value in the population when the preset end condition is satisfied. Take the optimal element value as the target model parameter value of the preset neural network model, and take the obtained target neural network model as the density measurement model.

[0175] Figure 6 It is a block diagram of a device 500 for obtaining slurry density shown according to an exemplary embodiment. As Figure 6 shown, the device 500 for obtaining slurry density further includes a training sample acquisition module 504, which is configured to:

[0176] Obtain multiple groups of historical data of limestone slurry. The historical data includes the alternative control parameter values of multiple alternative control parameter groups and the output parameter values of the corresponding output parameters. Each alternative control parameter group includes multiple alternative control parameters of limestone slurry, and the output parameter includes slurry density.

[0177] Determine the target control parameter from multiple alternative control parameters by using the grey relational analysis method according to multiple groups of historical data.

[0178] Obtain the preset training samples from multiple groups of historical data according to the target control parameter.

[0179] Optionally, the training sample acquisition module 504 is further configured to:

[0180] Obtain the grey relational degree between the alternative control parameter value of each alternative control parameter and the corresponding output parameter value according to the historical data.

[0181] Determine the target control parameter from multiple alternative control parameters according to the grey relational degree and the preset grey relational degree threshold. The grey relational degree of the target control parameter is greater than or equal to the preset grey relational degree threshold.

[0182] Optionally, the training sample acquisition module 504 is further configured to:

[0183] After obtaining multiple groups of historical data of the limestone slurry, abnormal historical data is deleted from the multiple groups of historical data according to a preset range of multiple control parameter values, where the abnormal historical data represents that the alternative control parameter value of any alternative control parameter in the alternative control parameter group exceeds the corresponding control parameter value range.

[0184] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment related to the method, and will not be elaborated here.

[0185] In the above technical solution, the problem that the density meter is prone to inaccurate measurement results due to wear when directly obtaining the slurry density in the related art is avoided. Instead, the slurry density is obtained by using the target control parameter value obtained from a stable measurement related to the slurry density, which reduces the cost of regularly replacing and maintaining the density meter, improves the accuracy of the slurry density result, and is beneficial to improving the effect of flue gas desulfurization.

[0186] Figure 7 It is a block diagram of an electronic device 700 shown according to an exemplary embodiment. As Figure 7 shown, the electronic device 700 may include: a processor 701, a memory 702. The electronic device 700 may further include one or more of a multimedia component 703, an input / output (I / O) interface 704, and a communication component 705.

[0187] Among them, the processor 701 is used to control the overall operation of the electronic device 700 to complete all or part of the steps in the above-mentioned method for obtaining slurry density. The memory 702 is used to store various types of data to support the operation of the electronic device 700. These data may include, for example, instructions for any application or method operating on the electronic device 700, as well as application-related data, such as contact data, sent and received messages, pictures, audio, video, and so on. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, magnetic disk or optical disc. The multimedia component 703 may include a screen and an audio component. Among them, the screen may be a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 702 or sent through the communication component 705. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 704 provides an interface between the processor 701 and other interface modules. The above-mentioned other interface modules may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IOT, eMTC, or other 5G, etc., or a combination of one or more of them, is not limited herein. Therefore, the corresponding communication component 705 may include: a Wi-Fi module, a Bluetooth module, an NFC module, and so on.

[0188] In another exemplary embodiment, a non-transitory computer-readable storage medium including program instructions is further provided. When the program instructions are executed by a processor, the steps of the above-described method for obtaining slurry density are implemented. For example, the computer-readable storage medium may be the above-described memory 702 including program instructions, and the above program instructions may be executed by the processor 701 of the electronic device 700 to complete the above-described method for obtaining slurry density.

[0189] Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the specification and practicing the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed by the present disclosure. The specification and embodiments are only to be considered as exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0190] It should be understood that the present disclosure is not limited to the exact structures already described and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A method for obtaining slurry density, characterized in that, The method includes: Training a density measurement model through a whale optimization algorithm according to a preset training sample, where the elements of the population members of the whale optimization algorithm are the model parameters of the density measurement model, the initial element values of the elements of each population member in the whale optimization algorithm are obtained through a Logistic chaotic map, the preset training sample includes the training control parameter values of multiple training control parameter groups and the training output parameter values of the corresponding training output parameters, each of the training control parameter groups includes multiple training control parameters, and the training output parameter includes the slurry density; Obtaining the target control parameter value of the target control parameter for generating limestone slurry; Inputting the target control parameter value into the density measurement model to obtain the slurry density of the limestone slurry output by the density measurement model.

2. The method according to claim 1, wherein The density measurement model is a neural network model, and the training of the density measurement model through the whale optimization algorithm according to the preset training sample includes: Determining the whale optimization algorithm parameters of the whale optimization algorithm according to the model parameters of the preset neural network model, where the whale optimization algorithm parameters include the elements of the population members, the preset element value threshold of each element value, and the whale population size; Obtaining the current element value of each population member of the whale optimization algorithm through the Logistic chaotic map according to the whale optimization algorithm parameters and the preset chaotic map parameters, where the chaotic map parameters include a chaotic map coefficient; Obtaining the target model parameter value of the preset neural network model by using the whale optimization algorithm according to the whale optimization algorithm parameters, the current element value of each population member, and the preset training sample, and taking the obtained target neural network model as the density measurement model.

3. The method according to claim 2, wherein The determining the whale optimization algorithm parameters of the whale optimization algorithm according to the model parameters of the preset neural network model includes: Taking the training control parameter as the input node of the preset neural network model, taking the slurry density as the output node of the preset neural network model, determining the model parameters of the preset neural network model according to the input node, the output node, and the hidden neurons of the preset neural network model, and taking the model parameters as the elements of the population members, where the model parameters include the first connection weight, the first threshold of the hidden neuron, the second connection weight, and the second threshold of the output node, the first connection weight represents the weight between the input node and the hidden neuron, and the second connection weight represents the weight between the hidden neuron and the output node; Determining the element value range of each element according to the preset model parameter value range of each model parameter; Determining the whale population size according to the number of model parameter values and the expected time of model training.

4. The method according to claim 2, wherein The obtaining the target model parameter value of the preset neural network model by using the whale optimization algorithm according to the whale optimization algorithm parameters, the current element value of each population member, and the preset training sample, and taking the obtained target neural network model as the density measurement model includes: Obtain the fitness value of each population member according to the preset training samples. The fitness value is the mean square error between the predicted output parameter value and the corresponding training output parameter value. The predicted output parameter value is obtained through the preset neural network model according to the training control parameter value of each training control parameter group and the current element value of the population member. Obtain the target element value of the target population member of the population according to the fitness value of each population member. The target population member is the population member with the smallest fitness value in the population. Adopt the whale optimization algorithm according to the whale optimization algorithm parameters to complete the encirclement, predation, and search of the target element value, so as to update the current element value of each population member. Repeat the steps of obtaining the fitness value of each population member according to the preset training samples and obtaining the target element value of the target population member of the population according to the fitness value of each population member, so as to obtain the optimal element value corresponding to the optimal population member with the smallest fitness value in the population when the preset end condition is satisfied. Take the optimal element value as the target model parameter value of the preset neural network model, and take the obtained target neural network model as the density measurement model.

5. The method according to claim 4, characterized in that, The preset end condition includes that the fitness value corresponding to the optimal population member is less than or equal to the preset fitness value threshold, and / or the number of training times is greater than or equal to the preset training times threshold.

6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: Obtain multiple groups of historical data of the limestone slurry. The historical data includes the alternative control parameter values of multiple alternative control parameter groups and the output parameter values of the corresponding output parameters. Each alternative control parameter group includes multiple alternative control parameters of the limestone slurry, and the output parameter includes the slurry density. Determine the target control parameter from the multiple alternative control parameters according to the multiple groups of historical data by using the grey relational analysis method. Obtain the preset training samples from the multiple groups of historical data according to the target control parameter.

7. The method according to claim 6, characterized in that, The determining the target control parameter from the multiple alternative control parameters according to the multiple groups of historical data by using the grey relational analysis method includes: Obtain the grey relational degree between the alternative control parameter value of each alternative control parameter and the corresponding output parameter value according to the historical data. Determine the target control parameter from the multiple alternative control parameters according to the grey relational degree and the preset grey relational degree threshold. The grey relational degree of the target control parameter is greater than or equal to the preset grey relational degree threshold.

8. The method according to claim 6, wherein The method further includes: After obtaining the multiple groups of historical data of the limestone slurry, delete the abnormal historical data from the multiple groups of historical data according to the preset ranges of multiple control parameter values. The abnormal historical data indicates that the alternative control parameter value of any alternative control parameter in the alternative control parameter group exceeds the corresponding control parameter value range.

9. The method according to any one of claims 1 to 5, characterized in that, The target control parameter and the training control parameter include the inlet pressure of the cyclone station, the instantaneous feed rate, the instantaneous dilution water flow rate, and the instantaneous grinding water flow rate.

10. An apparatus for obtaining the density of a slurry, characterized in that, The device includes: A model training module, configured to train a density measurement model according to preset training samples through a whale optimization algorithm, where the elements of the population members of the whale optimization algorithm are the model parameters of the density measurement model, and the initial element values of the elements of each population member in the whale optimization algorithm are obtained through a Logistic chaotic map. The preset training samples include the training control parameter values of multiple training control parameter groups and the training output parameter values of the corresponding training output parameters. Each training control parameter group includes multiple training control parameters, and the training output parameters include the slurry density; A parameter acquisition module, configured to acquire the target control parameter value of the target control parameter for generating limestone slurry; A result acquisition module, configured to input the target control parameter value into a pre-trained density measurement model to obtain the slurry density of the limestone slurry output by the density measurement model.

11. An electronic device, characterized in that, Comprising: A memory, on which a computer program is stored; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1-9.

12. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method according to any one of claims 1-9.

Images