Acid-base neutralization prediction method and device for wastewater reaction tank

Abstract

The present invention provides a method and device for predicting acid-base neutralization of a wastewater reaction tank. The method comprises: obtaining historical inflow, outflow, and pH values ​​of the reaction tank, and dividing the historical inflow, outflow, and pH values ​​into a training data set and a test data set; using the training data set to train a pre-established initial partial least squares model, and using the test data set to test the trained initial partial least squares model to obtain an acid-base neutralization prediction model; and inputting the obtained inflow and outflow of the reaction tank into the acid-base neutralization prediction model to obtain a base neutralization prediction result. By establishing a nonlinear partial least squares model of the wastewater reaction tank, the present invention exhibits high modeling accuracy when the number of modeling samples is small, thereby accurately and efficiently modeling the wastewater acid-base neutralization process, accurately predicting the pH value of the neutralized wastewater, and avoiding the consequences of inadequate wastewater neutralization.

Description

Technical Field

[0001] The invention relates to the technical field of acid-base neutralization, in particular to a method and device for predicting acid-base neutralization of a wastewater reaction tank. Background Art

[0002] Wastewater treatment systems typically treat domestic or industrial wastewater. Industrial production generates a wide variety of pollutants, and the types and concentrations of pollutants produced by different industries vary significantly. Domestic sewage refers to the wastewater discharged by people in their daily activities. This wastewater is primarily contaminated by domestic waste and human excrement. The amount, composition, and concentration of pollutants are related to people's living habits and water consumption. Domestic sewage generally does not contain toxic substances, but it does have conditions suitable for microbial growth and contains a large number of bacteria and pathogens, which poses certain hazards from a hygienic perspective.

[0003] Currently, wastewater treatment requires acid-base neutralization. This process can be modeled to predict the pH of the treated effluent, thereby achieving wastewater treatment.

[0004] With the continuous development of artificial intelligence technology, data-driven modeling methods are becoming more and more common. Generally speaking, in order to establish high-precision data models, large-scale modeling samples are often required. However, in actual industrial production, such as wastewater treatment, the amount of modeling data available is often limited. For example, for newly built units or units undergoing unit modification and commissioning, there will be a small amount of historical process data available for modeling, or a limited amount of representative test data obtained from hot tests. In these cases, especially for complex multivariable industrial processes, the small data sample size and the high degree of correlation between variables bring great difficulties to modeling. Therefore, how to model wastewater acid-base neutralization and achieve accurate prediction of pH values ​​is an urgent problem to be solved. Summary of the Invention

[0005] In view of the problems existing in the prior art, the main purpose of the embodiments of the present invention is to provide a method and device for predicting acid-base neutralization of a wastewater reaction tank, so as to achieve accurate prediction of the neutralized pH value of the wastewater.

[0006] To achieve the above objectives, an embodiment of the present invention provides a method for predicting acid-base neutralization in a wastewater reaction tank, the method comprising:

[0007] Obtain the historical inflow, outflow, and pH value of the reaction tank, and divide the historical inflow, outflow, and pH value into a training data set and a test data set;

[0008] The pre-established initial partial least squares model is trained using the training data set, and the trained initial partial least squares model is tested using the test data set to obtain an acid-base neutralization prediction model;

[0009] The obtained reaction tank inflow and outflow are input into the acid-base neutralization prediction model to obtain the alkali neutralization prediction result.

[0010] Optionally, in one embodiment of the present invention, the historical inflow includes historical strong acid inflow, historical buffer flow inflow and historical strong alkali inflow.

[0011] Optionally, in one embodiment of the present invention, the initial partial least squares model includes an internal extreme learning machine mapping function and an external partial least squares method.

[0012] Optionally, in one embodiment of the present invention, using a training data set to train a pre-established initial partial least squares model includes:

[0013] Using the external partial least squares method in the initial partial least squares model, feature information is extracted from the training data set to obtain feature information;

[0014] The internal extreme learning machine mapping function in the initial partial least squares model is used to update the weights of the initial partial least squares model according to the feature information to obtain the trained initial partial least squares model.

[0015] An embodiment of the present invention further provides a wastewater reaction tank acid-base neutralization prediction device, the device comprising:

[0016] A data acquisition module is used to obtain the historical inflow, historical outflow and historical pH value of the reaction tank, and divide the historical inflow, historical outflow and historical pH value into a training data set and a test data set;

[0017] The model training module is used to train the pre-established initial partial least squares model using the training data set, and to test the trained initial partial least squares model using the test data set to obtain an acid-base neutralization prediction model;

[0018] The prediction result module is used to input the obtained reaction tank inflow and reaction tank outflow into the acid-base neutralization prediction model to obtain the alkali neutralization prediction result.

[0019] Optionally, in one embodiment of the present invention, the historical inflow includes historical strong acid inflow, historical buffer flow inflow and historical strong alkali inflow.

[0020] Optionally, in one embodiment of the present invention, the initial partial least squares model includes an internal extreme learning machine mapping function and an external partial least squares method.

[0021] Optionally, in one embodiment of the present invention, the model training module includes:

[0022] A feature information unit is used to extract feature information from a training data set using an external partial least squares method in an initial partial least squares model to obtain feature information;

[0023] The weight updating unit is used to use the internal extreme learning machine mapping function in the initial partial least squares model to update the weights of the initial partial least squares model according to the feature information to obtain a trained initial partial least squares model.

[0024] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above method when executing the program.

[0025] The present invention also provides a computer-readable storage medium storing a computer program for executing the above method.

[0026] The present invention establishes a nonlinear partial least squares model of the wastewater reaction tank, which shows higher modeling accuracy when the modeling samples are small, thereby accurately and efficiently modeling the wastewater acid-base neutralization process, accurately predicting the pH value of the neutralized wastewater, and avoiding the consequences of inadequate neutralization of the wastewater. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0028] Figure 1 This is a flow chart of a method for predicting acid-base neutralization in a wastewater reaction tank according to an embodiment of the present invention;

[0029] Figure 2 This is a flow chart of model training in an embodiment of the present invention;

[0030] Figure 3 Schematic diagram of the acid-base neutralization prediction model in an embodiment of the present invention;

[0031] Figure 4 This is a schematic diagram of acid-base neutralization of a wastewater reaction tank in a specific embodiment of the present invention;

[0032] Figure 5 This is a schematic structural diagram of a device for predicting acid-base neutralization of a wastewater reaction tank according to an embodiment of the present invention;

[0033] Figure 6 This is a schematic diagram of the structure of the model training module in an embodiment of the present invention;

[0034] Figure 7 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The embodiments of the present invention provide a method and device for predicting acid-base neutralization in a wastewater reaction tank.

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0037] Partial least squares (PLS) demonstrates excellent performance in addressing multiple correlations and small sample sizes, and has been widely used in industrial process modeling. It combines multivariate statistical regression, principal component analysis, and canonical correlation analysis to extract comprehensive variables that best explain the system. By projecting the high-dimensional data space between the independent and dependent variables into the corresponding low-dimensional feature space, it effectively eliminates the collinearity of the original independent variables. However, PLS is a linear regression method, and the relationships within most industrial processes are often nonlinear rather than linear. Therefore, PLS struggles to achieve satisfactory results for such problems. To address complex nonlinear processes, researchers have proposed nonlinear partial least squares methods. Generally speaking, these methods can be divided into two categories: the first category is nonlinear partial least squares methods based on external sample transformations; the second category is nonlinear partial least squares methods based on internal nonlinear mappings. Among them, the most common form of external sample transformation method is kernel partial least squares. The principle is to use functions such as polynomial kernel function or Gaussian kernel function to map the independent variable samples from the original low-dimensional space to the high-dimensional feature space, and then use the linear PLS method to build the model. Some nonlinear mappings are also achieved by expanding the nonlinear terms of the independent variable sample matrix. Common nonlinear terms are square terms, cubic terms, and cross terms. In addition, there are also sample transformation methods based on mechanisms; another type of internal nonlinear mapping nonlinear partial least squares method uses nonlinear functions to describe the correspondence between independent variables and dependent variable components, while keeping the external linear PLS framework unchanged.

[0038] In order to solve the problem of difficulty in establishing a high-precision data model when there is a small amount of sample data and a large correlation between variables when establishing a data model, the present invention proposes a new nonlinear partial least squares algorithm. The algorithm uses the extreme learning machine (ELM) as the internal mapping function, keeps the linear PLS external framework unchanged, introduces an error-based input weight update step, and constructs a nonlinear PLS model based on the extreme learning machine.

[0039] like Figure 1 The flowchart of a method for predicting acid-base neutralization of a wastewater reaction tank according to an embodiment of the present invention is shown. The execution subject of the method for predicting acid-base neutralization of a wastewater reaction tank provided by the embodiment of the present invention includes but is not limited to a computer. Figure 1 The methods shown include:

[0040] Step S1, obtaining the historical inflow, historical outflow, and historical pH value of the reaction tank, and dividing the historical inflow, historical outflow, and historical pH value into a training data set and a test data set;

[0041] Step S2, using the training data set to train the pre-established initial partial least squares model, and using the test data set to test the trained initial partial least squares model to obtain an acid-base neutralization prediction model;

[0042] Step S3: input the acquired inflow rate and outflow rate of the reaction tank into the acid-base neutralization prediction model to obtain the base neutralization prediction result.

[0043] As an embodiment of the present invention, the historical inflow includes the historical strong acid inflow, the historical buffer flow inflow and the historical strong base inflow.

[0044] Among them, the historical data of the wastewater reaction tank is obtained, including historical inflow, historical outflow and historical pH value. Figure 4 Figure 1 shows a schematic diagram of acid-base neutralization in a reaction tank. The neutralization reaction occurs between a strong acid (q1, such as nitric acid), a buffer flow (q2, such as sodium bicarbonate), and a strong base (q3, such as sodium hydroxide) in the reaction tank. The pH value of the output flow is determined by the flow rates at the inlet (q1-q3) and outlet (q4). To maintain a constant liquid level in the tank, the outlet flow rate varies with the inlet flow rate. The historical data for the strong acid, buffer flow, and strong base are used as the historical inflow, the historical data for the output flow as the historical outflow, and the historical data for the pH value of the output flow as the historical pH value. The historical inflow, historical outflow, and historical pH values ​​are divided into a training dataset and a test dataset.

[0045] As an embodiment of the present invention, the initial partial least squares model includes an internal extreme learning machine mapping function and an external partial least squares method.

[0046] Among them, the pre-established initial partial least squares model is as follows Figure 3 As shown in the figure, the partial least squares model uses the extreme learning machine (ELM) as the internal mapping function, namely the ELM internal model, keeps the linear partial least squares (PLS) external framework unchanged, namely the PLS external model, referred to as the ELMPLS model, introduces the error-based input weight update step, and constitutes a nonlinear PLS model based on the extreme learning machine, namely the partial least squares model.

[0047] Furthermore, the initial partial least squares model is trained using the training data set. Specifically, the PLS external model extracts feature information to obtain feature information, and the internal extreme learning machine mapping function updates the weights of the initial partial least squares model based on the feature information to obtain a trained initial partial least squares model. The trained initial partial least squares model is then tested using the test data to obtain an acid-base neutralization prediction model.

[0048] Furthermore, real-time data from the wastewater reaction tank, including inflow and outflow, is collected. These inflow and outflow are then fed into an acid-base neutralization prediction model. This model then calculates and processes the data to predict the alkali neutralization of the wastewater reaction tank. This alkali neutralization prediction is then used to adjust the acid and alkali inflows into the reaction tank, ensuring precise control of wastewater treatment.

[0049] In this embodiment, if Figure 2 As shown, using the training data set, training the pre-established initial partial least squares model includes:

[0050] Step S21, using the external partial least squares method in the initial partial least squares model to extract feature information from the training data set to obtain feature information;

[0051] Step S22 , using the internal extreme learning machine mapping function in the initial partial least squares model, updating the weights of the initial partial least squares model according to the feature information, to obtain a trained initial partial least squares model.

[0052] Among them, the ELMPLS model adopts the method of combining the external PLS framework with the internal ELM function to obtain a new nonlinear PLS method. The model uses external PLS to extract characteristic information from the input variables, decomposes the multi-input and single-output variables (X, Y) into multiple groups of single-input and single-output principal component vectors (t, u), reduces the dimension and collinearity of the input vector, and uses the internal ELM model to establish nonlinear relationships in the latent space. The overall structure of ELMPLS is as follows: Figure 3 shown.

[0053] Furthermore, the main input and output components t and u are obtained from the PLS external model, and then each pair of components is fitted according to formula (1).

[0054] u=f(t)+e (1)

[0055] When the nonlinear function used to fit the intrinsic relationship between the input and output latent variables is continuously differentiable with respect to the input weight w, an error-based input weight update process can be performed. For Equation (1), the relationship between the output and input principal components can be approximated using the Newton-Raphson linearization method.

[0056]

[0057] in, f0 is the estimated value of the output principal component, which can be obtained through the ELM function, w0 is the current w value, and the partial derivative is expanded.

[0058]

[0059] Step 1, parameter initialization: perform data normalization processing on the multi-input single-output data (X, y), E0 = Zscore(X), F0 = Zscore(y), A = 1.

[0060] Step 2, PLS external extraction:

[0061] (1) Let u A =F A-1 ;

[0062] (2) Calculate the weight vector of E0: Formula

[0063] (3) Weight normalization: w A =w A / ||w A ||;

[0064] (4) Calculate the principal component vector of E0: t A =E A-1 w A .

[0065] Step 3, ELM internal mapping:

[0066] (1) According to the input and output principal components (t A ,u A ) Train the ELM model and obtain the internal mapping model f A (·);

[0067] (2) According to the internal mapping model f A (·) Calculate the estimated value of u,

[0068] Step 4, weight update:

[0069] (1) Calculate the weight increment according to formula (4-13): Δw A =(D T D) -1 D T e A ;

[0070] (2) New weight: w A =w A +Δw A ;

[0071] (3) Weight normalization: w A =w A / ||w A ||.

[0072] Step 5, calculate the new input principal component vector:

[0073] (1) Calculate the input principal component vector using the new weights: t A =E A-1 w A ;

[0074] (2) Check whether the changes between the new principal component and the original principal component in step 2 converge. If so, execute step 6; if not, jump to step 3.

[0075] Step 6: Calculate the model output parameters:

[0076] (1) Calculate the model input load vector:

[0077] (2) Calculate the estimated value of the output principal component u:

[0078] (3) Calculate the input and output residual terms:

[0079] Step 7, model accuracy test: determine whether the accuracy requirements are met based on the residual information or the principal component information. If so, stop the iteration. If not, set A=A+1 and repeat steps 2 to 7.

[0080] The above steps 1 to 7 are the data training part of the ELMPLS model. After training, a series of input load vectors P = [p1, p2, ..., p A ], output principal component vector Weight coefficient vector W=[w1,w2,...,w A ] and the mapping function [f1,f2,...,f A]; For single output data, the output load vector is q=[1,1,...,1] 1×A , then the model's prediction value for the training data is the denormalized result of the following formula:

[0081]

[0082] For the test data X t , the corresponding input principal component T can be calculated based on the load vector P and the weight coefficient vector W t =[t1,t2,...,t A ]:

[0083] T t =X t W(PW) -1 (5)

[0084] Calculate T using the above formula t And the series of internal mapping function test data obtained during training corresponding to [f1,f2,...,f A ]Calculate the principal components of the predicted output corresponding to the test data The predicted output value corresponding to the test data can be obtained by denormalizing the following results:

[0085]

[0086] In the above formula, the vector q is still a vector whose elements are all 1.

[0087] In a specific embodiment of the present invention, a currently common partial least squares modeling method includes dimensionality reduction of multivariate input data based on a partial least squares method and selecting an appropriate score vector as the input of a Gaussian process regression model. Subsequently, by selecting and combining covariance functions, different types of Gaussian process regression soft-sensor models are constructed to predict the output data. Finally, the predictive ability of the model is evaluated using test data. This method uses Gaussian process regression as its internal mapping function. Gaussian process regression is essentially a probabilistic model. Therefore, when the number of modeling samples is small, the modeling accuracy of the partial least squares Gaussian regression soft-sensor modeling method based on partial least squares will decrease. Moreover, the computational principles of probabilistic models are complex, making it difficult to apply to industrial sites with limited computing conditions. Therefore, the present invention addresses the problem of difficulty in establishing a high-precision data model when establishing a data model with a small number of sample data and large correlations between variables. A new nonlinear partial least squares algorithm is proposed. This algorithm uses an extreme learning machine (ELM) as the internal mapping function, maintains the linear PLS external framework unchanged, and introduces an error-based input weight update step to form a nonlinear PLS model based on the extreme learning machine.

[0088] In this embodiment, in order to verify the performance of the ELMPLS model, the soft sensing model was established using ELMPLS and pH neutralization simulation data. Figure 4 The figure below is a schematic diagram of the acid-base neutralization reaction tank. The neutralization reaction of the strong acid (q1), buffer flow (q2), and strong base (q3) in the reaction tank is shown. The pH value of the output flow is determined by the flow rates at the inlet (q1-q3) and outlet (q4). To maintain the liquid level in the tank, the outlet flow rate varies with the inlet flow rate. A data set of 81 sampling points was obtained under different inlet flow conditions. The relationship between the pH value and the liquid flow rates q1-q4 was verified. The sample data was divided into a training data set and a test data set. 65 data sets were selected as the training data set, and the remaining 16 data sets were selected as the test data set.

[0089] Among them, the pH neutralization soft sensor model is established with q1, q2, q3 and q4 as input variables and pH value as output variable. In order to verify the performance of ELMPLS proposed in this paper, three other models are also established for comparison, namely linear PLS regression model (PLSR), ANNPLS model and *ELMPLS model. *ELMPLS uses ELM as the internal mapping function, but does not update the weights; the four models constructed in this invention use the same linear PLS to extract the principal components externally, and four different internal mapping models are used internally. The root mean square error (RMSE) and mean relative error (MRE) are used as two indicators to evaluate the performance of the model.

[0090] Furthermore, Table 1 shows the modeling results of various models for the pH neutralization process, that is, the prediction results of each model for the pH neutralization process. Tr and RMSE Te are the root mean square errors of the training samples and the test samples, respectively. Similarly, MRE Tr and MREs Teare the average relative errors of the training samples and test samples, respectively. As can be seen from the table, the PLSR (partial least squares regression) model has the worst effect, with an average relative error of 1.0768 for the training data set and 1.1994 for the test data set, exceeding 12%. This shows that linear PLSR cannot accurately describe and process complex nonlinear object characteristics. The root mean square error values ​​of the ANNPLS (artificial neural network-partial least squares method) model for the training data and test data are 0.573 and 0.5940, respectively, showing good prediction accuracy. The ELMPLS method has the best fitting and generalization capabilities, with the root mean square error and average relative error values ​​for the test data being 0.4681 and 4.89%, respectively. It can be seen from the prediction results that the ELMPLS model modeling effect is significantly better than the *ELMPLS model due to the addition of the error-based input weight update step. This experimental result shows that the constructed ELMPLS model has strong nonlinear fitting and generalization performance.

[0091] Table 1

[0092]

[0093] The present invention uses the extreme learning machine as the internal mapping function and keeps the external framework of the linear partial least squares method unchanged, thereby forming a nonlinear partial least squares model based on the extreme learning machine. It can show higher modeling accuracy when the number of modeling samples is small, thereby accurately and efficiently modeling the acid-base neutralization process of wastewater, accurately predicting the pH value of the neutralized wastewater, and avoiding the consequences of inadequate neutralization of the wastewater.

[0094] like Figure 5 The figure shows a schematic structural diagram of a wastewater reaction tank acid-base neutralization prediction device according to an embodiment of the present invention. The device shown in the figure includes:

[0095] A data acquisition module 10 is used to obtain historical inflow, historical outflow, and historical pH values ​​of the reaction tank, and divide the historical inflow, historical outflow, and historical pH values ​​into a training data set and a test data set;

[0096] The model training module 20 is used to train the pre-established initial partial least squares model using the training data set, and to test the trained initial partial least squares model using the test data set to obtain an acid-base neutralization prediction model;

[0097] The prediction result module 30 is used to input the acquired inflow rate and outflow rate of the reaction tank into the acid-base neutralization prediction model to obtain the base neutralization prediction result.

[0098] As an embodiment of the present invention, the historical inflow includes the historical strong acid inflow, the historical buffer flow inflow and the historical strong base inflow.

[0099] As an embodiment of the present invention, the initial partial least squares model includes an internal extreme learning machine mapping function and an external partial least squares method.

[0100] In this embodiment, if Figure 6 As shown, the model training module 20 includes:

[0101] A feature information unit 21 is configured to extract feature information from the training data set using the external partial least squares method in the initial partial least squares model to obtain feature information;

[0102] The weight updating unit 22 is used to use the internal extreme learning machine mapping function in the initial partial least squares model to update the weights of the initial partial least squares model according to the feature information to obtain a trained initial partial least squares model.

[0103] Based on the same application concept as the aforementioned method for predicting acid-base neutralization in a wastewater reaction tank, the present invention also provides the aforementioned device for predicting acid-base neutralization in a wastewater reaction tank. Because the principles underlying the device for predicting acid-base neutralization in a wastewater reaction tank are similar to those of the method for predicting acid-base neutralization in a wastewater reaction tank, the implementation of the device for predicting acid-base neutralization in a wastewater reaction tank can be referenced to the implementation of the method for predicting acid-base neutralization in a wastewater reaction tank, and any repetitions will not be repeated.

[0104] The present invention uses the extreme learning machine as the internal mapping function and keeps the external framework of the linear partial least squares method unchanged, thereby forming a nonlinear partial least squares model based on the extreme learning machine. It can show higher modeling accuracy when the number of modeling samples is small, thereby accurately and efficiently modeling the acid-base neutralization process of wastewater, accurately predicting the pH value of the neutralized wastewater, and avoiding the consequences of inadequate neutralization of the wastewater.

[0105] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above method when executing the program.

[0106] The present invention also provides a computer-readable storage medium storing a computer program for executing the above method.

[0107] like Figure 7 As shown, the electronic device 600 may further include: a communication module 110, an input unit 120, an audio processing unit 130, a display 160, and a power supply 170. It is worth noting that the electronic device 600 does not necessarily have to include Figure 7In addition, the electronic device 600 may also include all components shown in Figure 7 For components not shown, reference may be made to the prior art.

[0108] like Figure 7 As shown, the central processing unit 100 is sometimes also referred to as a controller or an operation control unit, and may include a microprocessor or other processor device and / or logic device. The central processing unit 100 receives inputs and controls the operations of various components of the electronic device 600 .

[0109] Memory 140 may be, for example, one or more of a cache, flash memory, hard drive, removable media, volatile memory, non-volatile memory, or other suitable devices. It may store the aforementioned failure-related information and may also store programs that execute the relevant information. The CPU 100 may execute the programs stored in memory 140 to implement information storage or processing.

[0110] The input unit 120 provides input to the CPU 100. The input unit 120 may be, for example, a keypad or touch input device. The power supply 170 is used to provide power to the electronic device 600. The display 160 is used to display objects such as images and text. The display may be, for example, an LCD display, but is not limited thereto.

[0111] The memory 140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), or a SIM card. Alternatively, it may be a memory that retains information even when power is off, can be selectively erased, and is provided with more data. Examples of such memory are sometimes referred to as EPROMs. The memory 140 may also be some other type of device. The memory 140 includes a buffer memory 141 (sometimes referred to as a buffer). The memory 140 may include an application / function storage unit 142 for storing application programs and function programs or processes for executing the operations of the electronic device 600 via the central processing unit 100.

[0112] The memory 140 may also include a data storage unit 143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 144 of the memory 140 may include various driver programs for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0113] The communication module 110 is a transmitter / receiver 110 that transmits and receives signals via an antenna 111. The communication module (transmitter / receiver) 110 is coupled to the central processor 100 to provide input signals and receive output signals, which may be the same as in a conventional mobile communication terminal.

[0114] Based on different communication technologies, multiple communication modules 110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module, and / or a wireless local area network module. The communication module (transmitter / receiver) 110 is also coupled to a speaker 131 and a microphone 132 via an audio processor 130 to provide audio output via the speaker 131 and receive audio input from the microphone 132, thereby implementing common telecommunication functions. The audio processor 130 may include any suitable buffer, decoder, amplifier, etc. Furthermore, the audio processor 130 is also coupled to the central processing unit 100, enabling local recording via the microphone 132 and playback of stored audio via the speaker 131.

[0115] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0116] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0117] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1The function specified in one or more boxes.

[0118] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0119] Specific embodiments are used in the present invention to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A method for predicting acid-base neutralization of a wastewater reaction tank, characterized in that: The method comprises: Obtaining historical inflow, historical outflow, and historical pH values ​​of the reaction tank, and dividing the historical inflow, historical outflow, and historical pH values ​​into a training data set and a test data set, wherein the historical inflow includes historical strong acid inflow, historical buffer flow inflow, and historical strong base inflow; Using the training data set, a pre-established initial partial least squares model is trained, and using the test data set, the trained initial partial least squares model is tested to obtain an acid-base neutralization prediction model, wherein the initial partial least squares model includes an internal extreme learning machine mapping function and an external partial least squares method; Inputting the obtained reaction tank inflow and reaction tank outflow into the acid-base neutralization prediction model to obtain a base neutralization prediction result; The step of training the pre-established initial partial least squares model using the training data set includes: Extracting feature information from the training data set using the external partial least squares method to obtain feature information, wherein the feature information includes an input principal component vector and an output principal component vector; The internal extreme learning machine mapping function is trained according to the feature information, and the weights of the initial partial least squares model are updated. The external partial least squares method updates the input principal component vector according to the updated weights, and then checks whether the change between the updated input principal component vector and the original input principal component vector converges. If not, return to the above step of training the internal extreme learning machine mapping function, and train again using the updated input principal component vector. If convergence occurs, the training is completed to obtain the trained initial partial least squares model.

2. A wastewater reaction tank acid-base neutralization prediction device, characterized in that: The device comprises: a data acquisition module for acquiring historical inflow, historical outflow, and historical pH values ​​of the reaction tank, and dividing the historical inflow, historical outflow, and historical pH values ​​into a training data set and a test data set, wherein the historical inflow includes historical strong acid inflow, historical buffer flow inflow, and historical strong base inflow; A model training module is used to train a pre-established initial partial least squares model using the training data set, and to test the trained initial partial least squares model using the test data set to obtain an acid-base neutralization prediction model, wherein the initial partial least squares model includes an internal extreme learning machine mapping function and an external partial least squares method; A prediction result module, used to input the acquired reaction tank inflow and reaction tank outflow into the acid-base neutralization prediction model to obtain a base neutralization prediction result; The model training module includes: A feature information unit, configured to extract feature information from the training data set using the external partial least squares method to obtain feature information, wherein the feature information includes an input principal component vector and an output principal component vector; A weight updating unit is used to train the internal extreme learning machine mapping function according to the feature information, update the weights of the initial partial least squares model, update the input principal component vector according to the updated weights using the external partial least squares method, and then check whether the change between the updated input principal component vector and the original input principal component vector converges. If not, return to the above step of training the internal extreme learning machine mapping function and train again using the updated input principal component vector. If convergence occurs, the training is completed to obtain the trained initial partial least squares model.

3. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to claim 1 is implemented.

4. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program for executing the method according to claim 1 .

Images