Chloride ion potentiometric titration error prediction method based on neural network

Abstract

The invention provides a chloride ion potentiometric titration error prediction method based on a neural network. The method comprises the following specific steps: (1) selecting factors which can influence a test result in a thermal power plant desulfurization wastewater chloride ion determination process as research objects; (2) measuring corresponding actual errors of the desulfurization wastewater sample as basic data; (3) analyzing reliability factors influencing errors of the chloride ion potentiometric titration method; and (4) taking the reliability factor as an input variable, establishing a BP neural network model for prediction, and obtaining a prediction result. According to the method, errors of chloride ion monitoring caused by various influence factors in the desulfurization wastewater monitoring process in practical application can be predicted, and the accuracy of thermal power plant wastewater quality analysis is improved.

Description

Technical Field

[0001] This invention belongs to the field of water quality analysis technology for wastewater from thermal power plants, specifically relating to a method for predicting chloride ion potentiometric titration errors based on neural networks. Background Technology

[0002] In the treatment of desulfurization wastewater from thermal power plants, chloride ion concentrations often reach extremely high levels, ranging from thousands to tens of thousands of milligrams per liter, exhibiting strong corrosive characteristics. This high-concentration chloride ion environment promotes the corrosion process of metal materials, posing a threat to facilities such as pipes, valves, and pumps. This not only leads to increased maintenance costs but also creates potential safety risks. Furthermore, it hinders the effectiveness of subsequent wastewater treatment processes. Therefore, continuous monitoring of chloride ion content in desulfurization wastewater is extremely crucial.

[0003] Current online titration systems utilize the peak threshold of the first derivative curve and the positive / negative inflection point of the second derivative to automatically determine the endpoint. The ion concentration is output based on the volume of standard titration solution consumed at the abrupt change. However, these systems do not consider the interfering ions present in desulfurization wastewater in practical applications, nor the specific operating methods of the instrument. This may lead to a certain deviation between the measured data and the actual data, affecting the reliability and repeatability of the ion concentration results obtained in practical applications.

[0004] Therefore, designing an operational method that can predict the accuracy of online chloride titration in practical applications and improve the accuracy of potentiometric titration in measuring chloride concentration in power plant desulfurization wastewater is an urgent problem to be solved in the field of wastewater quality analysis technology for thermal power plants. Summary of the Invention

[0005] The present invention aims to at least partially solve one of the technical problems in the related art.

[0006] Therefore, the first objective of this invention is to propose a method for predicting chloride ion potentiometric titration errors based on neural networks.

[0007] The second objective of this invention is to propose a device for predicting chloride ion potentiometric titration errors based on a neural network.

[0008] To achieve the above objectives, a first aspect of the present invention proposes a method for predicting chloride ion potentiometric titration errors based on a neural network, comprising: S1, obtain data on nine influencing factors and corresponding actual error data in the chloride ion determination process in desulfurization wastewater of thermal power plants; S2, Grey relational analysis was used to determine the reliability of the nine influencing factors, and key influencing factors with a correlation degree greater than 0.3 were selected as input variables; S3. Based on the key influencing factors, a BP neural network prediction model is constructed. The input data is normalized by a two-stage sigmoid function to map the data to the [-1,1] interval. The dataset is divided into training set and test set in a 5:1 ratio for model training. S4. The trained BP neural network model is used to predict the error of the real-time measured chloride ion concentration, and the titration operation parameters are dynamically adjusted according to the prediction results to compensate for the measurement error.

[0009] In one embodiment of the present invention, S1 includes: S11, influencing factors include chloride ion concentration gradient, presence of heavy metal ions, raw water dilution factor, wastewater pH value, concentration of titrant, stirrer speed, electrode maintenance, potential acquisition method and titration speed. S12, the chloride ion concentration error is calculated according to the formula... Perform calculations, where This represents the volume of silver nitrate standard titration solution consumed by the sample. To transfer the sample volume, This represents the actual concentration of the silver nitrate standard titration solution. The molar mass of chlorine, The chloride ion concentration of the prepared sample solution; S13, the summation concentration of the heavy metal ions is calculated according to the formula... Perform calculations, where The valence state of heavy metal ions. This represents the concentration of the corresponding heavy metal ions.

[0010] In one embodiment of the present invention, S2 includes: S21, the nine influencing factors are divided into preset intervals, among which the chloride ion concentration interval is: The heavy metal concentration range is ; S22, the data of the nine influencing factors are dimensionless by using one of the following methods: maximum-minimum method, extreme value method, or standard deviation method.

[0011] In one embodiment of the present invention, S3 further includes: S31 uses a two-stage sigmoid function to normalize the input data; its transformation formula is as follows: ,in Here, Xmax represents the maximum value in the data, and Xmin represents the minimum value in the data. S32, set the number of hidden layer nodes of the BP neural network to... ,in This represents the number of key influencing factors selected.

[0012] In one embodiment of the present invention, S4 includes: S41, when the error compensation value of the prediction result exceeds a preset threshold, the BP neural network model is iteratively optimized based on the error compensation value; S42 stores the optimized model parameters in a preset database and prioritizes the use of these parameters for error prediction in subsequent titration operations.

[0013] To achieve the above objectives, a second aspect of the present invention provides a device for predicting chloride ion potentiometric titration errors based on a neural network, comprising: The data acquisition module acquires data on nine influencing factors in the chloride ion determination process in desulfurization wastewater from thermal power plants, along with corresponding actual error data. The reliability determination module uses grey relational analysis to determine the reliability of the nine influencing factors and selects key influencing factors with a correlation degree greater than 0.3 as input variables. The model building module constructs a BP neural network prediction model based on the key influencing factors, processes the input data through a two-stage sigmoid function to map the data to the [-1,1] interval, and divides the dataset into training and test sets in a 5:1 ratio for model training. The error prediction and adjustment module uses a trained BP neural network model to predict the error of the real-time measured chloride ion concentration, and dynamically adjusts the titration operation parameters based on the prediction results to compensate for the measurement error.

[0014] The present invention provides a method for predicting chloride ion potentiometric titration error based on neural networks, which can effectively predict and compensate for measurement errors in chloride ion potentiometric titration in desulfurization wastewater from thermal power plants, thereby improving the accuracy and reliability of online monitoring.

[0015] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0016] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a method for predicting chloride ion potentiometric titration error based on a neural network according to an embodiment of the present invention; Figure 2 This is an architecture diagram of a method for predicting chloride ion potentiometric titration error based on a neural network according to an embodiment of the present invention; Figure 3 This is a structural diagram of a chloride ion potentiometric titration error prediction device based on a neural network according to an embodiment of the present invention. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0018] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0019] The following description, with reference to the accompanying drawings, describes a method and apparatus for predicting chloride ion potentiometric titration errors based on a neural network, according to an embodiment of the present invention.

[0020] Figure 1 This is a flowchart of a method for predicting chloride ion potentiometric titration error based on a neural network according to an embodiment of the present invention, as shown below. Figure 1 As shown, it includes: S1, obtain data on nine influencing factors and corresponding actual error data in the chloride ion determination process in desulfurization wastewater of thermal power plants; S2, Grey relational analysis was used to determine the reliability of the nine influencing factors, and key influencing factors with a correlation degree greater than 0.3 were selected as input variables; S3. Based on the key influencing factors, a BP neural network prediction model is constructed. The input data is normalized by a two-stage sigmoid function to map the data to the [-1,1] interval. The dataset is divided into training set and test set in a 5:1 ratio for model training. S4. The trained BP neural network model is used to predict the error of the real-time measured chloride ion concentration, and the titration operation parameters are dynamically adjusted according to the prediction results to compensate for the measurement error.

[0021] The present invention provides a method for predicting chloride ion potentiometric titration error based on neural networks, which can effectively predict and compensate for measurement errors in chloride ion potentiometric titration in desulfurization wastewater from thermal power plants, thereby improving the accuracy and reliability of online monitoring.

[0022] The following describes in detail, with reference to the accompanying drawings, a method for predicting chloride ion potentiometric titration error based on a neural network according to an embodiment of the present invention.

[0023] The purpose of this invention is to further reduce the error in determining chloride ions in desulfurization wastewater using potentiometric titration in thermal power plants and improve accuracy. Specifically, it relates to a method for predicting chloride ion potentiometric titration errors using a neural network.

[0024] like Figure 2 As shown, the method for predicting the error of chloride ion potentiometric titration using a neural network according to the present invention includes the following steps: Step 1: Select the factors that will affect the test results during the determination of chloride ions in desulfurization wastewater from thermal power plants as the research object.

[0025] Step 2: The actual error of the collected samples is used as the basic data.

[0026] Step 3: Analyze the reliability factors affecting the error of chloride ion potentiometric titration.

[0027] Step 4: Use reliability factors as input variables to build a BP neural network model.

[0028] Step 5: Make a prediction and obtain the prediction result.

[0029] Furthermore, the influencing factors mentioned in step one include chloride ion concentration gradient, presence of heavy metal ions, raw water dilution factor, wastewater pH value, concentration of titration reagent, stirrer speed, electrode maintenance, potential acquisition method, and titration speed.

[0030] Furthermore, in step one, the nine research subjects are designated as X1, X2, X3, X4, X5, X6, X7, X8, and X9, respectively.

[0031] Furthermore, the basic data in step two comes from manually measured values ​​and readings from an automatic titration instrument for desulfurization wastewater, and the error E is calculated using the following formula:

[0032] In the formula, V1 is the volume of silver nitrate standard titration solution consumed by the sample, in mL; V is the volume of sample transferred, in mL; c is the accurate value of the actual concentration of the silver nitrate standard titration solution, in mol / L; M is the molar mass of chlorine, taken as 35.45 g / mol; and ρ1 is the concentration of chloride ions in the prepared sample solution, in mg / L.

[0033] Furthermore, each research subject has n different levels.

[0034] Furthermore, since the chloride ion concentration in the desulfurization wastewater from the thermal power plant mentioned in step two is Y'1~Y1, the value of object X1 can be divided into different intervals, namely... , ...and .

[0035] Furthermore, the heavy metal concentration mentioned in step two is the sum of various heavy metal ions, calculated using the following formula:

[0036] in, Represents the valence state of heavy metal ions. This represents the concentration of the corresponding heavy metal ions.

[0037] Furthermore, since the heavy metal concentration in the desulfurization wastewater from the thermal power plant mentioned in step two is Y'2~Y2, the value of object X2 can be divided into different intervals, namely... , ...and .

[0038] Furthermore, the reliability factors mentioned in step three are determined by correlation coefficients and correlation degrees.

[0039] Furthermore, the correlation coefficient and correlation degree were obtained through grey relational analysis using SPSSAU.

[0040] Furthermore, the specific steps of grey relational analysis are as follows: 1) Determine the comparison sequence as Xs=(xs(1), xs(2), xs(3), xs(4), ..., xs(m)), and the reference sequence as Y0=(y0(1), y0(2), y0(3), y0(4), ..., y0(m)).

[0041] 2) Select one of the following methods—maximum-minimum method, extreme value method, and standard deviation method—to perform dimensionless processing on the data of the nine research subjects to eliminate the influence of dimensions.

[0042] 3) The system automatically calculates the correlation coefficient and correlation degree between each reference sequence and the comparison sequence.

[0043] 4) Determine the influence of the research object on the magnitude of the error of potentiometric titration based on the correlation degree, remove research objects with low correlation degree, and select influencing factors with high correlation degree to control them.

[0044] Furthermore, in step three, when the correlation degree is ≤0.3, it is considered to have low correlation degree and is judged as an unreliable factor; when the correlation degree is >0.3, it is considered to have high correlation degree and is judged as a reliable factor.

[0045] Furthermore, the specific steps for neural network prediction in step four are as follows: 1) Normalize the data.

[0046] 2) Determine the input layer, hidden layer, and output layer of the neural network.

[0047] 3) Make predictions.

[0048] Furthermore, in step four, the BP neural network uses a two-stage sigmoid function to normalize the data, fixing the output object to the range [-1, 1]. Its transformation function is shown below:

[0049] Among them, X i For the data to be processed, X max X represents the maximum value in the data. min It is the minimum value in the data.

[0050] Furthermore, in step four, the input layer of the BP neural network is k layers, which are k influencing factors with a correlation degree > 0.3.

[0051] Furthermore, in step four, the output layer of the BP neural network prediction model represents the chloride ion potentiometric titration error.

[0052] Furthermore, the number of hidden layer nodes in step four .

[0053] Furthermore, the establishment of the BP neural network in step four includes determining the training set and the test set, with the number of training sets being [number missing]. There are [number] test sets. indivual.

[0054] Furthermore, in step four, the prediction results are verified to ensure the accuracy and stability of the model. If the simulation results are not ideal, the amount of training data can be increased or the results can be reversed to improve the accuracy of the results.

[0055] To achieve the above embodiments, such as Figure 3 As shown, this embodiment also provides a chloride ion potentiometric titration error prediction device 10 based on neural networks. The device 10 includes a data acquisition module 100, a reliability determination module 200, a model construction module 300, and an error prediction and adjustment module 400.

[0056] The data acquisition module 100 acquires data on nine influencing factors in the chloride ion determination process in desulfurization wastewater from thermal power plants, along with corresponding actual error data. The reliability determination module 200 uses grey relational analysis to determine the reliability of the nine influencing factors and selects key influencing factors with a correlation degree greater than 0.3 as input variables. The model building module 300 constructs a BP neural network prediction model based on the key influencing factors, processes the input data through a two-stage sigmoid function normalization, maps the data to the [-1,1] interval, and divides the dataset into training and test sets in a 5:1 ratio for model training. The error prediction and adjustment module 400 uses a trained BP neural network model to predict the error of the real-time measured chloride ion concentration, and dynamically adjusts the titration operation parameters based on the prediction results to compensate for the measurement error.

[0057] Furthermore, the data acquisition module 100 is also used for: Influencing factors include chloride ion concentration gradient, presence of heavy metal ions, raw water dilution factor, wastewater pH value, concentration of titrant, stirrer speed, electrode maintenance, potential acquisition method, and titration speed. The chloride ion concentration error is calculated according to the formula. Perform calculations, where This represents the volume of silver nitrate standard titration solution consumed by the sample. To transfer the sample volume, This represents the actual concentration of the silver nitrate standard titration solution. The molar mass of chlorine, The chloride ion concentration of the prepared sample solution; The summation concentration of the heavy metal ions is calculated according to the formula. Perform calculations, where The valence state of heavy metal ions. This represents the concentration of the corresponding heavy metal ions.

[0058] Furthermore, the aforementioned reliability determination module 200 is also used for: The nine influencing factors were divided into preset intervals, with the chloride ion concentration interval being [missing information]. The heavy metal concentration range is ; The data of the nine influencing factors are dimensionless by using one of the following methods: maximum-minimum method, extreme value method, or standard deviation method.

[0059] Furthermore, the aforementioned model building module 300 is also used for: The input data is normalized using a two-stage sigmoid function, and its transformation formula is as follows: ,in For the data to be processed, X max X represents the maximum value in the data. min It is the minimum value in the data; Set the number of hidden layer nodes in the BP neural network to ,in This represents the number of key influencing factors selected.

[0060] Furthermore, the aforementioned error prediction and adjustment module 400 is also used for: When the error compensation value of the prediction result exceeds a preset threshold, the BP neural network model is iteratively optimized based on the error compensation value. The optimized model parameters are stored in a preset database and are used preferentially for error prediction in subsequent titration operations.

[0061] An embodiment of the present invention provides a chloride ion potentiometric titration error prediction device based on a neural network, which can effectively predict and compensate for the measurement error of chloride ion potentiometric titration in desulfurization wastewater of thermal power plants, thereby improving the accuracy and reliability of online monitoring.

[0062] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0063] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for predicting chloride ion potentiometric titration error based on neural networks, characterized in that, include: S1, obtain data on nine influencing factors and corresponding actual error data in the chloride ion determination process in desulfurization wastewater of thermal power plants; S2, Grey relational analysis was used to determine the reliability of the nine influencing factors, and key influencing factors with a correlation degree greater than 0.3 were selected as input variables; S3. Based on the key influencing factors, a BP neural network prediction model is constructed. The input data is normalized by a two-stage sigmoid function to map the data to the [-1,1] interval. The dataset is divided into training set and test set in a 5:1 ratio for model training. S4. The trained BP neural network model is used to predict the error of the real-time measured chloride ion concentration, and the titration operation parameters are dynamically adjusted according to the prediction results to compensate for the measurement error.

2. The method as described in claim 1, characterized in that, S1 includes: S11, influencing factors include chloride ion concentration gradient, presence of heavy metal ions, raw water dilution factor, wastewater pH value, concentration of titrant, stirrer speed, electrode maintenance, potential acquisition method, and titration speed. S12, the chloride ion concentration error is calculated according to the formula... Perform calculations, where This represents the volume of silver nitrate standard titration solution consumed by the sample. To transfer the sample volume, This represents the actual concentration of the silver nitrate standard titration solution. The molar mass of chlorine, The chloride ion concentration of the prepared sample solution; S13, the summation concentration of the heavy metal ions is calculated according to the formula... Perform calculations, where The valence state of heavy metal ions. This represents the concentration of the corresponding heavy metal ions.

3. The method as described in claim 1, characterized in that, The S2 includes: S21, the nine influencing factors are divided into preset intervals, among which the chloride ion concentration interval is: The heavy metal concentration range is ; S22, the data of the nine influencing factors are dimensionless by using one of the following methods: maximum-minimum method, extreme value method, or standard deviation method.

4. The method as described in claim 1, characterized in that, The S3 further includes: S31 uses a two-stage sigmoid function to normalize the input data; its transformation formula is as follows: ,in For the data to be processed, X max X represents the maximum value in the data. min It is the minimum value in the data; S32, set the number of hidden layer nodes of the BP neural network to... ,in This represents the number of key influencing factors selected.

5. The method as described in claim 1, characterized in that, The S4 includes: S41, when the error compensation value of the prediction result exceeds a preset threshold, the BP neural network model is iteratively optimized based on the error compensation value; S42 stores the optimized model parameters in a preset database and prioritizes the use of these parameters for error prediction in subsequent titration operations.

6. A device for predicting chloride ion potentiometric titration error based on a neural network, characterized in that, include: The data acquisition module acquires data on nine influencing factors in the chloride ion determination process in desulfurization wastewater from thermal power plants, along with corresponding actual error data. The reliability determination module uses grey relational analysis to determine the reliability of the nine influencing factors and selects key influencing factors with a correlation degree greater than 0.3 as input variables. The model building module constructs a BP neural network prediction model based on the key influencing factors, processes the input data through a two-stage sigmoid function to map the data to the [-1,1] interval, and divides the dataset into training and test sets in a 5:1 ratio for model training. The error prediction and adjustment module uses a trained BP neural network model to predict the error of the real-time measured chloride ion concentration, and dynamically adjusts the titration operation parameters based on the prediction results to compensate for the measurement error.

7. The apparatus as claimed in claim 6, characterized in that, The data acquisition module is also used for: Influencing factors include chloride ion concentration gradient, presence of heavy metal ions, raw water dilution factor, wastewater pH value, concentration of titrant, stirrer speed, electrode maintenance, potential acquisition method, and titration speed. The chloride ion concentration error is calculated according to the formula. Perform calculations, where This represents the volume of silver nitrate standard titration solution consumed by the sample. To transfer the sample volume, This represents the actual concentration of the silver nitrate standard titration solution. The molar mass of chlorine, The chloride ion concentration of the prepared sample solution; The summation concentration of the heavy metal ions is calculated according to the formula. Perform calculations, where The valence state of heavy metal ions. This represents the concentration of the corresponding heavy metal ions.

8. The apparatus as claimed in claim 6, characterized in that, The reliability determination module is also used for: The nine influencing factors were divided into preset intervals, with the chloride ion concentration interval being [missing information]. The heavy metal concentration range is ; The data of the nine influencing factors are dimensionless by using one of the following methods: maximum-minimum method, extreme value method, or standard deviation method.

9. The apparatus as claimed in claim 6, characterized in that, The model building module is also used for: The input data is normalized using a two-stage sigmoid function, and its transformation formula is as follows: ,in For the data to be processed, X max X represents the maximum value in the data. min It is the minimum value in the data; Set the number of hidden layer nodes in the BP neural network to ,in This represents the number of key influencing factors selected.

10. The apparatus as claimed in claim 6, characterized in that, The error prediction and adjustment module is also used for: When the error compensation value of the prediction result exceeds a preset threshold, the BP neural network model is iteratively optimized based on the error compensation value. The optimized model parameters are stored in a preset database and are used preferentially for error prediction in subsequent titration operations.

Images