Sewage hardness prediction method, device, electronic device and storage medium

Abstract

The present application discloses a method, device, electronic device, and storage medium for predicting the hardness of sewage. The method includes: obtaining the ion concentration parameters that cause an increase in the hardness of the sewage and the pH value of the sewage; determining the types of agents for reducing the hardness of the sewage and the dosage of each type according to the ion concentration parameters and the pH value of the sewage; using the ion concentration parameters, the types of agents, and / or the dosage of each type as inputs, and predicting whether the hardness of the sewage after treatment is qualified by using a pre-trained error backpropagation neural network prediction model. By introducing an artificial intelligence algorithm model in the treatment of sewage hardness, the automation and continuous self-learning optimization of sewage hardness prediction and treatment are realized, the prediction of the hardness of the sewage after treatment is accurate, the feedback of agent dosing is more timely, and the dosage of the agent is more accurate.

Description

Technical Field

[0001] The present application relates to the technical field of water treatment, and particularly to a method, a system, an electronic device and a storage medium for predicting the hardness of sewage. Background Art

[0002] During the process of treating and reducing industrial sewage, it is often necessary to remove hardness. Due to the lack of real-time monitoring means, the calculation of the chemical dosage is relatively rough. Usually, a fixed chemical dosage is set according to the design value, and the method of regular sampling and testing for adjustment is adopted. The disadvantage of this method is that it has no real-time adjustment ability, and the prediction and control accuracy of hardness removal are poor. Summary of the Invention

[0003] In view of the above problems, the present application is proposed to provide a method, a system, an electronic device and a storage medium for predicting the hardness of sewage that can overcome or at least partially solve the above problems.

[0004] According to one aspect of the present application, a method for predicting the hardness of sewage is provided. The method includes:

[0005] Obtaining the ion concentration parameters that cause the increase in hardness in the sewage and the pH value of the sewage;

[0006] Determining the types of chemicals for reducing the hardness of the sewage and the dosage of each type according to the ion concentration parameters and the pH value of the sewage;

[0007] Using the ion concentration parameters, the types of chemicals and / or the dosage of each type as inputs, and predicting whether the hardness of the treated sewage is qualified by using a pre-trained error inverse feedback neural network prediction model.

[0008] Optionally, the ion concentration parameters include the calcium ion concentration value, the magnesium ion concentration value, and the bicarbonate ion concentration value. Obtaining the ion concentration parameters that cause the increase in hardness in the sewage and the pH value of the sewage includes:

[0009] Monitoring the calcium ion concentration value, the magnesium ion concentration value, and the bicarbonate ion concentration value of the sewage in the regulating tank by using an on-line detection device;

[0010] Monitoring the pH value of the sewage in the regulating tank by using a pH probe.

[0011] Optionally, the on-line detection device includes a hyperspectral detection device. According to the ion concentration parameters and the pH value of the sewage, determining the types of chemicals for reducing the hardness of the sewage and the dosage of each type includes:

[0012] If the initial pH value of the sewage is pH in , the calcium ion concentration is Ca 2+ in , the magnesium ion concentration is Mg 2+in , the concentration of bicarbonate ion M in , the limit value of calcium ion concentration in the treated effluent Ca 2+ max , the limit value of magnesium ion concentration in the effluent Mg 2+ max , then the target pH of the effluent max is calculated as:

[0013]

[0014] the target value of bicarbonate ion concentration in the effluent M min is calculated as:

[0015]

[0016] If the quality of the sewage meets:

[0017]

[0018] then it is determined that only lime is added, and the lime dosage D1 is calculated as:

[0019]

[0020] If this water quality meets:

[0021]

[0022] It is determined to add sodium hydroxide and sodium carbonate, and the sodium carbonate dosage concentration D2 is calculated as:

[0023]

[0024] The sodium hydroxide dosage concentration D3 is calculated as:

[0025]

[0026] If this water quality meets:

[0027]

[0028] It is determined to add sodium hydroxide, and the sodium hydroxide dosage concentration D4 is calculated as:

[0029]

[0030] Among them, α1, α2, and α3 are comprehensive adjustment coefficients considering activity coefficients, reagent purity, temperature influence, reaction completeness, and correction coefficients.

[0031] Optionally, predicting whether the hardness of the treated sewage is qualified using a pre-trained error inverse feedback neural network prediction model includes:

[0032] Using the error inverse feedback neural network prediction model, the predicted values of the sewage ion concentration parameters and the pH value are obtained. If the predicted values of the sewage ion concentration parameters and the pH value are less than the preset limit values, it indicates that the sewage treatment is qualified and meets the standard. If it is unqualified, at least one of the types of chemicals and the dosage of each type is adjusted and predicted again until the prediction result meets the standard; or,

[0033] Using the error inverse feedback neural network prediction model, directly obtain the result of whether the predicted values of the sewage ion concentration parameters and the pH value are qualified. If it is unqualified, at least one of the types of chemicals and the dosage of each type is adjusted and predicted again until the prediction result meets the standard.

[0034] Optionally, if it is unqualified, at least one of the types of chemicals and the dosage of each type is adjusted and predicted again until the prediction result meets the standard, including:

[0035] When the predicted value of the sewage ion concentration parameter exceeds the preset limit value, at least one of α1, α2 or α3 is adjusted, the stepping ratio is adjusted, and the adjustment direction is positively correlated with the positive and negative of the difference between the predicted value and the preset limit value;

[0036] After stepping a preset number of times, if the difference between the predicted value and the preset limit value increases, return to the state before adjustment and adjust in the opposite direction;

[0037] Repeat until the predicted value is less than the preset limit value.

[0038] Optionally, the error inverse feedback neural network prediction model includes at least two hidden layers and a sigmoid activation function, where the number of nodes, the number of partitions and the learning rate of the hidden layer are determined according to the input parameters, output parameters and training conditions.

[0039] Optionally, the method further includes:

[0040] Using an on-line detection device to monitor and obtain the processed value of the ion concentration parameter of the sewage after treatment;

[0041] Feed back the processed value of the ion concentration parameter to the error inverse feedback neural network prediction model;

[0042] Use the processed value of the ion concentration parameter to optimize the error inverse feedback neural network prediction model.

[0043] According to another aspect of the present application, a sewage hardness prediction device is provided, and the device includes:

[0044] An acquisition module, adapted to acquire the ion concentration parameters that cause the hardness to increase in the sewage and the pH value of the sewage;

[0045] A determination module, adapted to determine the types of agents for reducing the hardness of sewage and the dosage of each type according to the ion concentration parameter and the sewage pH value;

[0046] A prediction module, adapted to use a pre-trained error backpropagation neural network prediction model to predict whether the hardness of the treated sewage is qualified with the ion concentration parameter, the type of agent, and the dosage of each type as inputs.

[0047] According to another aspect of the present application, an electronic device is provided, including: a processor; and a memory arranged to store computer-executable instructions, the executable instructions, when executed, causing the processor to execute the sewage hardness prediction method as described in any one of the above.

[0048] According to still another aspect of the present application, a computer-readable storage medium is provided, the computer-readable storage medium storing one or more programs, the one or more programs, when executed by a processor, implementing the sewage hardness prediction method as described in any one of the above.

[0049] In summary, the technical solution of the present application obtains the following technical effects:

[0050] The present invention provides a reasonable and reliable method for intelligent hardness prediction and removal of industrial sewage, solving the problems of current industrial sewage hardness removal relying on human experience, being unable to be fully automated, having a lag in agent dosing feedback, and inaccurate agent dosage.

[0051] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the specific embodiments of the present application are hereinafter specifically exemplified. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present application. And throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0053] Figure 1 A flowchart showing the process of a sewage hardness prediction method according to an embodiment of the present application;

[0054] Figure 2 A structural diagram showing the structure of a sewage hardness prediction device according to an embodiment of the present application;

[0055] Figure 3 A structural diagram showing the structure of an electronic device according to an embodiment of the present application;

[0056] Figure 4 Shows a schematic structural diagram of a computer-readable storage medium according to an embodiment of the present application;

[0057] Figure 5 Shows a schematic flow diagram of a sewage treatment method applied to the total discharge outlet of a steel plant according to an embodiment of the present application;

[0058] Figure 6 Shows a schematic flow diagram of a method for pre-treating hardening removal of concentrated brine in a steel plant according to an embodiment of the present application;

[0059] Figure 7 Shows a schematic flow diagram of a sewage treatment method applied to the total discharge outlet of a steel plant according to another embodiment of the present application;

[0060] Figure 8 Shows a schematic flow diagram of a method for pre-treating hardening removal of concentrated brine in a steel plant according to still another embodiment of the present application. Detailed implementation manners

[0061] Hereinafter, exemplary embodiments of the present application will be described in more detail with reference to the drawings. Although the exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present application can be more thoroughly understood and the scope of the present application can be completely conveyed to those skilled in the art.

[0062] Figure 1 Shows a method for predicting sewage hardness according to an embodiment of the present application, the method comprising:

[0063] Step 110: Obtain the ion concentration parameters that cause an increase in hardness in the sewage and the sewage pH value.

[0064] Wherein the sewage in this step is preferably industrial sewage, the ion concentration parameters include the ion names and ion concentration values that can cause an increase in hardness, and can be obtained through various means such as sampling and on-line detection equipment, while the sewage pH value can be detected by equipment such as a pH meter.

[0065] Step 120: Determine the types of agents for reducing sewage hardness and the dosage of each type according to the ion concentration parameters and the sewage pH value.

[0066] In this step, the types of agents and the dosage of each type can be determined according to the experience of workers, and are preferably obtained through scientific calculation methods.

[0067] Step 130: Using the ion concentration parameter, chemical agent type, and / or dosage of each type as inputs, predict whether the hardness of the treated sewage is qualified using a pre-trained error backpropagation neural network prediction model.

[0068] The error backpropagation neural network, also known as the BP neural network, is a neural network that propagates errors in the reverse direction. It includes an input layer, a hidden layer, and an output layer. Its basic idea is still the gradient descent method. There is a certain mathematical relationship between the error in the middle hidden layer and the error in the last layer, just like the error is propagated back, so it is called the BP neural network. In this step, after being trained with historical data, this neural network is used to predict the relevant parameters of the hardness of the treated sewage or directly draw a conclusion on whether it is qualified. It should be noted that the input ion concentration parameter can be partial or all, and the chemical agent type and the dosage of the chemical agent are also selectively input according to the prediction needs.

[0069] In summary, through Figure 1 the disclosed method, it is possible to realize the automation of sewage hardness prediction and treatment, and improve the accuracy and precision of sewage hardness prediction and treatment.

[0070] In one embodiment, the ion concentration parameter includes the calcium ion concentration value, the magnesium ion concentration value, and the bicarbonate ion concentration value. Then, obtaining the ion concentration parameter and the sewage pH value that cause the hardness to increase in step 110 includes:

[0071] Using an on-line detection device to monitor the calcium ion concentration value, the magnesium ion concentration value, and the bicarbonate ion concentration value of the sewage in the regulating tank;

[0072] Using a pH probe to monitor the pH value of the sewage in the regulating tank.

[0073] Preferably, the calcium hardness, magnesium hardness, and bicarbonate ion concentration of industrial sewage can be monitored by a hyperspectral device, etc., or the detection values can be obtained by sampling detection, etc. The database of the hyperspectral monitoring device is regularly calibrated and updated with the model, and the pH probe is regularly calibrated or replaced.

[0074] In one embodiment, the on-line detection device is preferably a hyperspectral detection device. Then, determining the chemical agent type and the dosage of each type for reducing the sewage hardness in step 120 includes:

[0075] If the initial sewage pH value is pH in , the calcium ion concentration is Ca 2+ in , the magnesium ion concentration is Mg 2+ in , the bicarbonate ion concentration M in , the limit value of the calcium ion concentration of the treated effluent Ca2+ max , the limit value of magnesium ion concentration in the effluent, Mg 2+ max ,

[0076] The target pH of the effluent max Calculated as:

[0077]

[0078] The target value of bicarbonate concentration in the effluent, M min Calculated as:

[0079]

[0080] Judge the type of chemical addition. Since the water quality meets:

[0081]

[0082] If the water quality of the sewage meets:

[0083]

[0084] Then it is determined that only lime is added, and the lime dosage D1 is calculated as:

[0085]

[0086] If the water quality meets:

[0087]

[0088] It is determined that sodium hydroxide and sodium carbonate are added, and the sodium carbonate dosage concentration D2 is calculated as:

[0089]

[0090] The sodium hydroxide dosage concentration D3 is calculated as:

[0091]

[0092] If the water quality meets:

[0093]

[0094] It is determined that sodium hydroxide is added, and the sodium hydroxide dosage concentration D4 is calculated as:

[0095]

[0096] Among them, α1, α2, and α3 are comprehensive adjustment coefficients considering activity coefficients, reagent purity, temperature influence, reaction completeness, and correction factors.

[0097] In one embodiment, predicting whether the hardness of the treated sewage is qualified in step 130 by using the pre-trained error inverse feedback neural network prediction model includes:

[0098] Using the error inverse feedback neural network prediction model to predict the predicted value of the sewage ion concentration parameter and the predicted value of pH. If the predicted value of the sewage ion concentration parameter and the predicted value of pH are less than the preset limit values, it indicates that the sewage treatment is qualified and meets the standard. If it is unqualified, at least one of the types of chemicals and the dosage of each type is adjusted and predicted again until the prediction result meets the standard.

[0099] As an alternative to the above method, it is also possible to directly obtain the result of whether the predicted value of the sewage ion concentration parameter and the predicted value of pH are qualified by using the error inverse feedback neural network prediction model. If it is unqualified, at least one of the types of chemicals and the dosage of each type is adjusted and predicted again until the prediction result meets the standard.

[0100] In one embodiment, if it is unqualified, at least one of the types of chemicals and the dosage of each type is adjusted and predicted again until the prediction result meets the standard, including:

[0101] When the predicted value of the sewage ion concentration parameter exceeds the preset limit value, at least one of α1, α2, or α3 is adjusted, the stepping ratio is adjusted, and the adjustment direction is positively correlated with the positive and negative of the difference between the predicted value and the preset limit value;

[0102] After stepping through the preset number of times, if the difference between the predicted value and the preset limit value increases, it returns to the state before adjustment and adjusts in the opposite direction;

[0103] Repeat until the predicted value is less than the preset limit value.

[0104] Furthermore, if the number of repeated predictions is greater than the preset number, such as 100 times, an alarm is generated and recorded, and the training can be adjusted by using a training optimizer or manual intervention.

[0105] In one embodiment, the error inverse feedback neural network prediction model includes at least two hidden layers and a sigmoid activation function, where the number of nodes, the number of partitions, and the learning rate of the hidden layer are determined according to the input parameters, output parameters, and training situation.

[0106] In a preferred embodiment, the input item is pH in , Ca 2+ in , Mg 2+ in , M in , Ca 2+ max , Mg 2+max , D2, D3, with 4 output items, and the predicted value of the pH of the treated effluent is pH pre , and the predicted value of the calcium ion concentration in the treated effluent is Ca 2+ pre , and the predicted value of the magnesium ion concentration in the treated effluent is Mg 2+ pre , and the predicted value of the bicarbonate concentration in the treated effluent is M pre , with 10 nodes in the first layer of the hidden layer, divided into 4 regions in the second layer, with 9 nodes in each region, and the learning rate n of the two layers is set to 0.1. It runs in the way of using historical data for offline training and updating the model regularly.

[0107] After this method is applied for a period of time, based on the set hardness removal target, the hardness removal effect of this concentrated brine is stably up to 99%.

[0108] In one embodiment, the method further includes:

[0109] Using an on-line detection device to monitor the processed values of the ion concentration parameters of the sewage after treatment;

[0110] Feeding back the processed values of the ion concentration parameters into the error inverse feedback neural network prediction model;

[0111] Using the processed values of the ion concentration parameters to optimize the error inverse feedback neural network prediction model.

[0112] See Figure 2 The shown embodiment discloses a sewage hardness prediction device 200, and the device includes:

[0113] An acquisition module 210, adapted to acquire the ion concentration parameters that cause the increase of hardness in the sewage and the sewage pH value;

[0114] A determination module 220, adapted to determine the types of agents for reducing the sewage hardness and the dosage of each type according to the ion concentration parameters and the sewage pH value;

[0115] A prediction module 230, adapted to use the ion concentration parameters, the types of agents and the dosage of each type as inputs, and use a pre-trained error inverse feedback neural network prediction model to predict whether the hardness of the treated sewage is qualified.

[0116] By setting the above device in an electronic device, the problems of the current industrial sewage hardness removal relying on human experience, being unable to be fully automated, the feedback of agent dosing being lagged, and the inaccuracy of the agent dosing amount can be solved.

[0117] In one embodiment, if the ion concentration parameters include the calcium ion concentration value, the magnesium ion concentration value, and the bicarbonate concentration value, then the acquisition module 210 is adapted to:

[0118] The calcium ion concentration value, magnesium ion concentration value, and bicarbonate concentration value of the sewage in the regulating tank are monitored by using an on-line detection device;

[0119] The pH value of the sewage in the regulating tank is monitored by using a pH probe.

[0120] In one embodiment, the on-line detection device includes a hyperspectral detection device, and the determination module 220 is adapted to:

[0121] If the initial sewage pH value is pH in , the calcium ion concentration is Ca 2+ in , the magnesium ion concentration is Mg 2+ in , the bicarbonate concentration M in , the limit value of the calcium ion concentration of the treated effluent is Ca 2+ max , the limit value of the magnesium ion concentration of the effluent is Mg 2+ max ,

[0122] The target pH of the effluent max is calculated as:

[0123]

[0124] The target value of the bicarbonate concentration of the effluent M min is calculated as:

[0125]

[0126] Judge the type of chemical added. Since the water quality satisfies:

[0127]

[0128] If the water quality of the sewage satisfies:

[0129]

[0130] Then it is determined that only lime is added, and the lime dosage D1 is calculated as:

[0131]

[0132] If the water quality satisfies:

[0133]

[0134] It is determined that sodium hydroxide and sodium carbonate are added, and the sodium carbonate dosage concentration D2 is calculated as:

[0135]

[0136] The dosing concentration D3 of sodium hydroxide is calculated as:

[0137]

[0138] If the water quality meets:

[0139]

[0140] It is determined to dose sodium hydroxide, and the dosing concentration D4 of sodium hydroxide is calculated as:

[0141]

[0142] Where α1, α2, and α3 are comprehensive adjustment coefficients considering activity coefficients, reagent purity, temperature influence, reaction completeness, and correction factors.

[0143] In one embodiment, the prediction module 230 is further adapted to:

[0144] Use the error inverse feedback neural network prediction model to predict the predicted values of sewage ion concentration parameters and pH. If the predicted values of sewage ion concentration parameters and pH are less than the preset limit values, it indicates that the sewage treatment is qualified and meets the standard. If not, at least one of the types of chemicals and the dosing amounts of various types are adjusted and predicted again until the prediction result meets the standard.

[0145] As an alternative to the above method, it is also possible to directly obtain the result of whether the predicted values of sewage ion concentration parameters and pH are qualified using the error inverse feedback neural network prediction model. If not, at least one of the types of chemicals and the dosing amounts of various types are adjusted and predicted again until the prediction result meets the standard.

[0146] In one embodiment, in the prediction module 230, if not qualified, at least one of the types of chemicals and the dosing amounts of various types are adjusted and predicted again until the prediction result meets the standard, including:

[0147] When the predicted value of the sewage ion concentration parameter exceeds the preset limit value, at least one of α1, α2, or α3 is adjusted, the adjustment step ratio is adjusted, and the adjustment direction is positively correlated with the positive or negative value of the difference between the predicted value and the preset limit value;

[0148] After the preset number of steps, if the difference between the predicted value and the preset limit value increases, return to the state before adjustment and adjust in the opposite direction;

[0149] Repeat until the predicted value is less than the preset limit value.

[0150] In one embodiment, the error inverse feedback neural network prediction model includes at least two hidden layers and a sigmoid activation function, wherein the number of nodes, the number of partitions, and the learning rate of the hidden layer are determined according to input parameters, output parameters, and training conditions.

[0151] In one embodiment, the device further includes an optimization module, and the optimization module is adapted to:

[0152] Monitor the processed value of the ion concentration parameter of the sewage after treatment by using an on-line detection device;

[0153] Feed back the processed value of the ion concentration parameter into the error inverse feedback neural network prediction model;

[0154] Optimize the error inverse feedback neural network prediction model by using the processed value of the ion concentration parameter.

[0155] Embodiment 1

[0156] Please refer to Figure 5 , an embodiment of the present invention provides a sewage treatment method for the general discharge outlet of a steel plant, including:

[0157] S1. The sewage at the general discharge outlet of the steel plant enters the regulation tank, and a hyperspectral device is used to monitor the calcium hardness, magnesium hardness, and bicarbonate concentration of the industrial sewage;

[0158] S2. The industrial sewage enters the high-density clarification tank, and the types and dosages of the chemicals to be added are calculated and determined according to the monitoring results and the pH value of the industrial sewage;

[0159] S3. Use the error inverse feedback neural network method model to predict the pH value, calcium hardness value, magnesium hardness value, and bicarbonate concentration of the effluent after precipitation. When the prediction result exceeds the limit value of the effluent calcium hardness or magnesium hardness index, adjust the calculated value of one or more chemical dosages and predict again. Repeat this process until the prediction result meets the standard;

[0160] S4. The effluent from the high-density clarification tank of the industrial sewage is discharged, and a hyperspectral device is used to monitor the calcium hardness, magnesium hardness, and bicarbonate concentration of the industrial sewage, and the monitoring results are fed back to the error inverse feedback neural network method for model self-optimization.

[0161] Specifically, in S1, in addition to using the hyperspectral monitoring device, a pH probe is also used to monitor the pH value of the sewage at the general discharge outlet. The database of the hyperspectral monitoring device is regularly calibrated and updated, and the pH probe is regularly calibrated or replaced.

[0162] In S2, according to the monitored pH value pH of the sewage at the general discharge outlet in , calcium ion concentration Ca 2+ in , magnesium ion concentration Mg2+ in , the concentration of bicarbonate ion M in , and the limit value of calcium ion concentration in the treated effluent Ca 2+ max , the limit value of magnesium ion concentration in the treated effluent Mg 2+ max , compare Ca 2+ max with Mg 2+ max , determine that the pH value of the effluent is related to the limit value of magnesium ion concentration in the treated effluent Mg 2+ max , the target pH value of the effluent max is calculated as:

[0163]

[0164] the target value of bicarbonate ion concentration in the effluent M min is calculated as:

[0165]

[0166] Judge the type of chemical addition. Since the water quality always meets:

[0167]

[0168] it is determined that only lime is added;

[0169] The lime dosage D1 is calculated as:

[0170]

[0171] In S3, an artificial intelligence prediction optimization is carried out using the prediction model formula obtained by training with the neural network algorithm of error backpropagation.

[0172] When the predicted value exceeds the effluent concentration limit, adjust α1, adjust the step ratio by 1%, and the adjustment direction is positively correlated with the positive and negative of the difference between the predicted value and the effluent concentration limit. If the difference between the predicted value and the effluent concentration limit increases after predicting and stepping 5 times, then retreat α1 to before stepping 5 times, and adjust the step direction to be negatively correlated with the positive and negative of the difference between the predicted value and the effluent concentration limit; repeat until the predicted value is less than the effluent concentration limit. When the number of repeated predictions is greater than 100 times, an alarm is generated and recorded.

[0173] In S4, in addition to using the hyperspectral monitoring equipment, a pH probe is also used to monitor the pH value of the treated effluent from the high-density clarifier. The database of the hyperspectral monitoring equipment and the model are regularly calibrated and updated, and the pH probe is regularly calibrated or replaced. The neural network algorithm of error backpropagation is used to optimize the model formula, with a double hidden layer, and the activation function uses the sigmoid function. The input item is pH in , Ca2+ in , Mg 2+ in , M in , Ca 2+ max , Mg 2+ max , D1, the output items are 4, and the predicted value of the pH of the treated effluent is pH pre , the predicted value of the calcium ion concentration in the treated effluent is Ca 2+ pre , the predicted value of the magnesium ion concentration in the treated effluent is Mg 2+ pre , the predicted value of the bicarbonate concentration in the treated effluent is M pre , there are 8 nodes in the first layer of the hidden layer, the second layer is divided into 4 areas, with 8 nodes in each area, and the learning rate n of the two layers is set to 0.1. It runs in the way of using historical data for offline training and updating the model regularly.

[0174] Example 2

[0175] Please refer to Figure 6 , an embodiment of the present invention provides a method for removing hardness from concentrated brine in a steel plant, including:

[0176] S1, the concentrated brine in the steel plant enters the regulating tank, and a hyperspectral device is used to monitor the calcium hardness, magnesium hardness, and bicarbonate concentration of the concentrated brine;

[0177] S2, the concentrated brine enters the inclined plate sedimentation tank, and the types and dosages of chemical agents are calculated and determined according to the monitoring results and the pH value of the concentrated brine;

[0178] S3, using the error backpropagation neural network method model to predict the pH value, calcium hardness value, magnesium hardness value, and bicarbonate concentration of the effluent after precipitation. When the prediction result exceeds the limit value of the calcium hardness or magnesium hardness index of the effluent, adjust the calculated value of one or more chemical agent dosages and predict again. Repeat this process until the prediction result meets the standard;

[0179] S4, the effluent of the concentrated brine after passing through the inclined plate sedimentation tank is monitored for calcium hardness, magnesium hardness, and bicarbonate concentration of the concentrated brine effluent using a hyperspectral device, and the pH value is monitored using a pH probe, and the monitoring results are fed back to the error backpropagation neural network method for model self-optimization.

[0180] Specifically, in S1, in addition to using a hyperspectral monitoring device, a pH probe is also used to monitor the pH value of the concentrated brine. The database of the hyperspectral monitoring device is regularly calibrated and updated, and the pH probe is regularly calibrated or replaced.

[0181] In S2, according to the monitored pH value pH of the concentrated brine in , calcium ion concentration Ca 2+in , the concentration of magnesium ions, Mg 2+ in , the concentration of bicarbonate ions, M in , and the limit value of the calcium ion concentration in the treated effluent, Ca 2+ max , the limit value of the magnesium ion concentration in the treated effluent, Mg 2+ max , compare Ca 2 + max with Mg 2+ max , determine that the pH value of the effluent is related to the limit value of the magnesium ion concentration in the treated effluent, Mg 2+ max , the target pH value of the effluent max is calculated as:

[0182]

[0183] the target value of the bicarbonate ion concentration in the effluent, M min is calculated as:

[0184]

[0185] , determine the type of chemical added. In a small part of the time, this water quality meets:

[0186]

[0187] , determine that sodium hydroxide and sodium carbonate need to be added at this time.

[0188] , the dosing concentration of sodium carbonate, D2, is calculated as:

[0189]

[0190] where α2 is a comprehensive adjustment coefficient considering activity coefficient, chemical purity, temperature influence, reaction completeness, and correction factor.

[0191] , the dosing concentration of sodium hydroxide, D3, is calculated as:

[0192]

[0193] where α3 is a comprehensive adjustment coefficient considering activity coefficient, chemical purity, temperature influence, reaction completeness, and correction factor.

[0194] , in most cases, this water quality meets:

[0195]

[0196] , determine that sodium hydroxide needs to be added at this time.

[0197] , the dosing concentration of sodium hydroxide, D3, is calculated as:

[0198]

[0199] Among them, α3 is a comprehensive adjustment coefficient considering activity coefficient, reagent purity, temperature influence, reaction completeness, and correction coefficient.

[0200] In S3, an artificial intelligence prediction optimization is performed using the prediction model formula obtained by training with the neural network algorithm of error backpropagation.

[0201] When the calcium hardness in the predicted value exceeds the effluent concentration limit, adjust α2 and α3, adjust the stepping ratio by 1%, and the adjustment direction is positively correlated with the positive or negative of the difference between the predicted value and the effluent concentration limit. After predicting and stepping 5 times, if the difference between the predicted value and the effluent concentration limit increases, then retreat α2 and α3 to before stepping 5 times, and adjust the stepping direction to be negatively correlated with the positive or negative of the difference between the predicted value and the effluent concentration limit; repeat until the predicted value is less than the effluent concentration limit. When the number of repeated predictions is greater than 100 times, an alarm is generated and recorded.

[0202] When the magnesium hardness in the predicted value exceeds the effluent concentration limit, adjust α3, adjust the stepping ratio by 1%, and the adjustment direction is positively correlated with the positive or negative of the difference between the predicted value and the effluent concentration limit. After predicting and stepping 5 times, if the difference between the predicted value and the effluent concentration limit increases, then retreat α3 to before stepping 5 times, and adjust the stepping direction to be negatively correlated with the positive or negative of the difference between the predicted value and the effluent concentration limit; repeat until the predicted value is less than the effluent concentration limit. When the number of repeated predictions is greater than 100 times, an alarm is generated and recorded.

[0203] In S4, in addition to using a hyperspectral monitoring device, a pH probe is also used to monitor the pH value of the effluent from the concentrated brine treatment. The database of the hyperspectral monitoring device and the model are regularly calibrated and updated, and the pH probe is regularly calibrated or replaced. The neural network algorithm of error backpropagation is used to optimize the model formula, with a double hidden layer, the activation function uses the sigmoid function, and the input items are pH in , Ca 2 + in , Mg 2+ in , M in , Ca 2+ max , Mg 2+ max , D2, D3, and the output items are 4 items, the predicted value of the pH of the effluent pH pre , the predicted value of the calcium ion concentration in the effluent Ca 2+ pre , the predicted value of the magnesium ion concentration in the effluent Mg 2+ pre , the predicted value of the bicarbonate concentration in the effluent M pre, there are 10 nodes in the first layer of the hidden layer, the second layer is divided into 4 areas, with 9 nodes in each area, and the learning rate n of the two layers is set to 0.1. It runs in the way of using historical data for offline training and updating the model regularly.

[0204] After this method has been applied for a period of time, based on the set hard removal target, the hardness removal effect of the concentrated brine reaches 99% stably, and the spot-check and calibration water quality are shown in the following table.

[0205]

[0206] Example 3

[0207] Please refer to Figure 7 , another embodiment of the present invention provides a sewage treatment method for the general discharge outlet of a steel plant, including:

[0208] S1, the sewage from the general discharge outlet of the steel plant enters the regulation tank, and the pH value, calcium hardness, magnesium hardness, and bicarbonate concentration of the industrial sewage are monitored;

[0209] S2, the industrial sewage enters the high-density clarification tank, and the types and dosages of chemical agents to be added are calculated and determined according to the monitoring results;

[0210] S3, use the error inverse feedback neural network method model to predict the pH value, calcium hardness value, magnesium hardness value, and bicarbonate concentration of the effluent after precipitation. When the prediction result exceeds the limit value of the effluent calcium hardness or magnesium hardness index, adjust the calculated value of one or more chemical agent dosages and predict again. Repeat this process until the prediction result meets the standard;

[0211] S4, the industrial sewage flows out after passing through the high-density clarification tank, and the pH value, calcium hardness, magnesium hardness, and bicarbonate concentration of the industrial sewage are monitored, and the monitoring results are fed back to the error inverse feedback neural network method for model self-optimization.

[0212] Specifically, in S1, a pH probe is used to monitor the pH value of the sewage at the general discharge outlet, and the pH probe is calibrated or replaced regularly. The calcium hardness, magnesium hardness, and bicarbonate concentration are detected by manual sampling.

[0213] In S2, according to the monitored pH value pH of the sewage at the general discharge outlet in , calcium ion concentration Ca 2+ in , magnesium ion concentration Mg 2+ in , bicarbonate concentration M in , and the limit value Ca of the calcium ion concentration in the treated effluent 2+ max , the limit value Mg of the magnesium ion concentration in the treated effluent 2+ max , compare Ca 2+ max with Mg2+ max , it is determined that the pH value of the effluent is related to the limit value of the magnesium ion concentration Mg in the treated effluent 2+ max related, the target pH of the effluent max is calculated as:

[0214]

[0215] The target value M of the bicarbonate concentration in the effluent min is calculated as:

[0216]

[0217] Judge the type of chemical added. Since the water quality always satisfies:

[0218]

[0219] It is determined that only lime is added;

[0220] The lime dosage D1 is calculated as:

[0221]

[0222] where α1 is a comprehensive adjustment coefficient considering activity coefficient, reagent purity, temperature influence, reaction completeness, and correction coefficient.

[0223] In S3, an artificial intelligence prediction optimization is performed using the prediction model formula obtained by training with the neural network algorithm of error backpropagation.

[0224] When the predicted value exceeds the effluent concentration limit, adjust α1, adjust the step ratio by 1%, and the adjustment direction is positively correlated with the positive and negative of the difference between the predicted value and the effluent concentration limit. If the difference between the predicted value and the effluent concentration limit increases after predicting and stepping 5 times, then roll back α1 to before stepping 5 times, and adjust the step direction to be negatively correlated with the positive and negative of the difference between the predicted value and the effluent concentration limit; repeat until the predicted value is less than the effluent concentration limit. When the number of repeated predictions is greater than 100 times, an alarm is generated and recorded.

[0225] In S4, a pH probe is used to monitor the pH value of the sewage at the total discharge outlet, and the pH probe is calibrated or replaced regularly. The calcium hardness, magnesium hardness, and bicarbonate concentration are detected using on-line monitoring instruments or manual sampling and chemical analysis. The model formula is optimized using the neural network algorithm of error backpropagation, with a double hidden layer, and the sigmoid function is used as the activation function. The input items are pH in , Ca 2+ in , Mg 2 + in , M in , Ca 2+max , Mg 2+ max , D1, with 4 output items, and the predicted value of the pH of the treated effluent is pH pre , and the predicted value of the calcium ion concentration in the treated effluent is Ca 2+ pre , and the predicted value of the magnesium ion concentration in the treated effluent is Mg 2+ pre , and the predicted value of the bicarbonate concentration in the treated effluent is M pre , there are 8 nodes in the first layer of the hidden layer, the second layer is divided into 4 areas, with 8 nodes in each area, and the learning rate n of the two layers is set to 0.1. It runs in the way of using historical data for offline training and updating the model regularly.

[0226] Example 4

[0227] Please refer to Figure 8 , yet another embodiment of the present invention provides a method for removing hardness from concentrated brine in a steel plant, including:

[0228] S1. The concentrated brine from the steel plant enters the regulation tank, and the pH value, calcium hardness, magnesium hardness, and bicarbonate concentration of the concentrated brine are monitored;

[0229] S2. The concentrated brine enters the inclined plate sedimentation tank, and the types and dosages of chemical agents to be added are calculated and determined according to the monitoring results;

[0230] S3. Using the error backpropagation neural network method model to predict the pH value, calcium hardness value, magnesium hardness value, and bicarbonate concentration of the effluent after sedimentation. When the prediction result exceeds the limit value of the calcium hardness or magnesium hardness index of the effluent, adjust the calculated value of one or more chemical agent dosages and predict again. Repeat this process until the prediction result meets the standard;

[0231] S4. The effluent from the inclined plate sedimentation tank of the concentrated brine is monitored for the pH value, calcium hardness, magnesium hardness, and bicarbonate concentration of the concentrated brine effluent, and the monitoring results are fed back to the error backpropagation neural network method for model self-optimization.

[0232] Specifically, in S1, a pH probe is used to monitor the pH value of the sewage at the total discharge port, and the pH probe is calibrated or replaced regularly. The calcium hardness, magnesium hardness, and bicarbonate concentration are detected using on-line monitoring instruments.

[0233] In S2, according to the monitored pH value pH of the concentrated brine in , calcium ion concentration Ca 2+ in , magnesium ion concentration Mg 2+ in , bicarbonate concentration M in , and the limit value of the calcium ion concentration in the treated effluent Ca 2+ max , and the limit value of the magnesium ion concentration in the treated effluent Mg2+ max , compare Ca 2 + max with Mg 2+ max , determine that the pH value of the effluent is related to the limit value of the magnesium ion concentration Mg in the treated effluent, and the target pH of the effluent 2+ max is calculated as: max Calculate as:

[0234]

[0235] The target value M of the bicarbonate concentration in the effluent min is calculated as:

[0236]

[0237] Determine the type of chemical agent to be added. For a small part of the time, this water quality meets:

[0238]

[0239] Determine that sodium hydroxide and sodium carbonate need to be added at this time.

[0240] The dosing concentration D2 of sodium carbonate is calculated as:

[0241]

[0242] where α2 is a comprehensive adjustment coefficient considering activity coefficient, reagent purity, temperature influence, reaction completeness, and correction factor.

[0243] The dosing concentration D3 of sodium hydroxide is calculated as:

[0244]

[0245] where α3 is a comprehensive adjustment coefficient considering activity coefficient, reagent purity, temperature influence, reaction completeness, and correction factor.

[0246] For most of the time, this water quality meets:

[0247]

[0248] Determine that sodium hydroxide needs to be added at this time.

[0249] The dosing concentration D3 of sodium hydroxide is calculated as:

[0250]

[0251] where α3 is a comprehensive adjustment coefficient considering activity coefficient, reagent purity, temperature influence, reaction completeness, and correction factor.

[0252] In S3, an artificial intelligence prediction optimization is performed using the prediction model formula obtained by training with the neural network algorithm of error backpropagation.

[0253] When the calcium hardness in the predicted value exceeds the effluent concentration limit, adjust α2 and α3, adjust the step ratio by 1%, and the adjustment direction is positively correlated with the positive or negative value of the difference between the predicted value and the effluent concentration limit. After predicting and stepping 5 times, if the difference between the predicted value and the effluent concentration limit increases, then retreat α2 and α3 to before stepping 5 times, and adjust the stepping direction to be negatively correlated with the positive or negative value of the difference between the predicted value and the effluent concentration limit; repeat until the predicted value is less than the effluent concentration limit. When the number of repeated predictions is greater than 100 times, an alarm is generated and recorded.

[0254] When the magnesium hardness in the predicted value exceeds the effluent concentration limit, adjust α3, adjust the step ratio by 1%, and the adjustment direction is positively correlated with the positive or negative value of the difference between the predicted value and the effluent concentration limit. After predicting and stepping 5 times, if the difference between the predicted value and the effluent concentration limit increases, then retreat α3 to before stepping 5 times, and adjust the stepping direction to be negatively correlated with the positive or negative value of the difference between the predicted value and the effluent concentration limit; repeat until the predicted value is less than the effluent concentration limit. When the number of repeated predictions is greater than 100 times, an alarm is generated and recorded.

[0255] In S4, a pH probe is used to monitor the pH value of the sewage at the total discharge outlet, and the pH probe is calibrated or replaced regularly. The calcium hardness, magnesium hardness, and bicarbonate concentration are detected using on-line monitoring instruments. The neural network algorithm of error backpropagation is used to optimize the model formula. There are two hidden layers, the activation function uses the sigmoid function, and the input items are pH in , Ca 2+ in , Mg 2+ in , M in , Ca 2+ max , Mg 2 + max , D2, D3, and the output items are 4 items. The predicted value of the pH of the treated effluent is pH pre , the predicted value of the calcium ion concentration of the treated effluent is Ca 2+ pre , the predicted value of the magnesium ion concentration of the treated effluent is Mg 2+ pre , the predicted value of the bicarbonate concentration of the treated effluent is M pre , there are 10 nodes in the first layer of the hidden layer, the second layer is divided into 4 areas, with 9 nodes in each area, and the learning rate n of the two layers is set to 0.1. It runs in the way of using historical data for offline training and updating the model regularly.

[0256] It should be noted that:

[0257] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general-purpose systems can also be used in conjunction with the teachings based hereon. The structure required to construct such systems will be apparent from the above description. In addition, the present application is not directed to any particular programming language. It should be appreciated that the teachings of the present application can be implemented in a variety of programming languages, and the description of a particular language above is for the purpose of disclosing the best mode of the present application.

[0258] In the specification provided herein, numerous specific details are set forth. However, it is understood that embodiments of the present application may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure an understanding of the present specification.

[0259] Similarly, it should be understood that, in order to streamline the present application and assist in understanding one or more of the various aspects thereof, in the above description of the exemplary embodiments of the present application, the various features of the present application are sometimes grouped together in a single embodiment, figure, or description thereof. However, the disclosed method should not be construed as reflecting an intention that the claimed present application requires more features than are expressly recited in each claim. Rather, as reflected in the following claims, the aspects of the application lie in less than all of the features of the single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into the detailed description, with each claim standing on its own as a separate embodiment of the present application.

[0260] Those skilled in the art will appreciate that the modules in the devices in the embodiments can be adaptively changed and disposed in one or more devices different from those of the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and in addition, they can be divided into multiple sub-modules or sub-units or sub-components. Except that at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all the features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all the processes or units of any method or device so disclosed. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) can be replaced by an alternative feature providing the same, equivalent, or similar purpose.

[0261] In addition, those skilled in the art can understand that although some embodiments described herein include certain features included in other embodiments rather than other features, the combination of features of different embodiments means that it is within the scope of this application and forms different embodiments. For example, in the following claims, any one of the claimed embodiments can be used in any combination.

[0262] Each component embodiment of the present application can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) can be used in practice to implement some or all of the functions of some or all of the components in the water quality and water volume warning system according to the embodiments of the present application. The present application can also be implemented as a device or system program (for example, a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0263] For example, Figure 3 The structural schematic diagram of an electronic device according to an embodiment of the present application is shown. The electronic device 300 includes a processor 310 and a memory 320 arranged to store computer-executable instructions (computer-readable program code). The memory 320 can be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk, or a ROM. The memory 320 has a storage space 330 for storing computer-readable program code 331 for executing any method step in the above methods. For example, the storage space 330 for storing computer-readable program code can include respective computer-readable program codes 331 for implementing various steps in the above methods. The computer-readable program code 331 can be read out from or written into one or more computer program products. These computer program products include program code carriers such as hard disks, compact discs (CDs), memory cards, or floppy disks. Such computer program products are generally, for example Figure 4 the computer-readable storage media described above. Figure 4The structural schematic diagram of a computer-readable storage medium according to an embodiment of the present application is shown. The computer-readable storage medium 400 stores computer-readable program code 331 for executing the method steps according to the present application, which can be read by the processor 310 of the electronic device 300. When the computer-readable program code 331 runs on the electronic device 300, it causes the electronic device 300 to execute each step in the method described above. Specifically, the computer-readable program code 331 stored in the computer-readable storage medium can execute the method shown in any of the above embodiments. The computer-readable program code 331 can be compressed in an appropriate form.

[0264] It should be noted that the above embodiments illustrate the present application rather than limit the present application, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in the claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present application can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In the unit claims listing several systems, several of these systems may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names.

Claims

1. A method for predicting the hardness of sewage, the method comprising: Obtaining the ion concentration parameters that cause an increase in hardness and the sewage pH value in the sewage; Determining the types of agents for reducing the hardness of the sewage and the dosage of each type according to the ion concentration parameters and the sewage pH value; Using the pre-trained error inverse feedback neural network prediction model with the ion concentration parameters, the types of agents, and / or the dosage of each type as inputs to predict whether the hardness of the treated sewage is qualified; The ion concentration parameters include the calcium ion concentration value, the magnesium ion concentration value, and the bicarbonate ion concentration value. Obtaining the ion concentration parameters that cause an increase in hardness and the sewage pH value in the sewage includes: Monitoring the calcium ion concentration value, the magnesium ion concentration value, and the bicarbonate ion concentration value of the sewage in the regulating tank using an on-line detection device; Monitoring the pH value of the sewage in the regulating tank using a pH probe; The on-line detection device includes a hyperspectral detection device. Determining the types of agents for reducing the hardness of the sewage and the dosage of each type according to the ion concentration parameters and the sewage pH value includes: If the initial pH value of the sewage is pH in , the calcium ion concentration is Ca 2+ in , the magnesium ion concentration is Mg 2+ in , the bicarbonate concentration is M in , the limit value of the calcium ion concentration in the treated effluent is Ca 2+ max , the limit value of the magnesium ion concentration in the effluent is Mg 2+ max , then the target pH of the effluent max is calculated as: Target value M of bicarbonate concentration in the effluent min Calculated as: If the water quality of the sewage satisfies: It is determined that only lime is added, and the lime dosage D1 is calculated as: If this water quality satisfies: It is determined that sodium hydroxide and sodium carbonate are added, and the sodium carbonate dosage concentration D2 is calculated as: The sodium hydroxide dosage concentration D3 is calculated as: If this water quality satisfies: It is determined that sodium hydroxide is added, and the sodium hydroxide dosage concentration D4 is calculated as: Where α1, α2, and α3 are comprehensive adjustment coefficients considering activity coefficients, agent purity, temperature effects, reaction completeness, and correction factors.

2. The method according to claim 1, characterized in that Using the pre-trained error inverse feedback neural network prediction model to predict whether the hardness of the treated sewage is qualified includes: Using the error inverse feedback neural network prediction model to predict the predicted values of the sewage ion concentration parameters and the pH prediction value. If the predicted values of the sewage ion concentration parameters and the pH prediction value are less than the preset limit values, it indicates that the sewage treatment is qualified and meets the standard. If not, at least one of the types of agents and the dosage of each type is adjusted and predicted again until the prediction result meets the standard; or, Using the error inverse feedback neural network prediction model to directly obtain the result of whether the predicted values of the sewage ion concentration parameters and the pH prediction value are qualified. If not, at least one of the types of agents and the dosage of each type is adjusted and predicted again until the prediction result meets the standard.

3. The method according to claim 2, characterized in that, If not, at least one of the types of agents and the dosage of each type is adjusted and predicted again until the prediction result meets the standard includes: When the predicted value of the sewage ion concentration parameter exceeds the preset limit value, at least one of α1, α2, or α3 is adjusted, the adjustment step ratio is adjusted, and the adjustment direction is positively correlated with the positive and negative difference between the predicted value and the preset limit value; After the preset number of steps, if the difference between the predicted value and the preset limit value increases, return to the state before adjustment and adjust in the opposite direction; Repeat until the predicted value is less than the preset limit value.

4. The method according to any one of claims 1 to 3, characterized in that, The error inverse feedback neural network prediction model includes at least two hidden layers and a sigmoid activation function, wherein the number of nodes, the number of partitions, and the learning rate of the hidden layer are determined according to input parameters, output parameters, and training conditions.

5. The method according to any one of claims 1 to 3, characterized in that, The method further includes: Monitoring, by using an on-line detection device, to obtain a processed value of the ion concentration parameter of the sewage after treatment; Feeding back the processed value of the ion concentration parameter into the error inverse feedback neural network prediction model; Optimizing the error inverse feedback neural network prediction model by using the processed value of the ion concentration parameter.

6. A sewage hardness prediction device for implementing the sewage hardness prediction method according to any one of claims 1-3, the device comprising: An acquisition module, adapted to acquire the ion concentration parameter causing the hardness increase in the sewage and the sewage pH value; A determination module, adapted to determine the types of agents for reducing the sewage hardness and the dosage of each type according to the ion concentration parameter and the sewage pH value; A prediction module, adapted to use the ion concentration parameter, the types of agents, and the dosage of each type as inputs, and predict whether the hardness of the treated sewage is qualified by using a pre-trained error inverse feedback neural network prediction model.

7. An electronic device, characterized in that, The electronic device includes: a processor; and a memory arranged to store computer-executable instructions, the executable instructions, when executed, causing the processor to execute the sewage hardness prediction method according to any one of claims 1-3.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the sewage hardness prediction method according to any one of claims 1-3 is implemented.

Images