A method and system for automatic optimization of water quality model parameters

Abstract

The present invention discloses a method and system for automatic optimization of water quality model parameters, including: obtaining a pre-constructed water quality model based on the relationship between water quality parameters and water quality targets; calculating the sensitivity index SI for the water quality parameters in the water quality model, sorting the parameters according to the sensitivity, and determining the sensitivity order of the water quality parameters; optimizing the parameters one by one according to the sensitivity order of the water quality parameters, and determining the optimal value of each water quality parameter. Advantages: The present invention constructs a BP neural network proxy model between the water quality model parameters and the water quality targets, calculates the sensitivity index SI by a one-variable-at-a-time method, and optimizes the parameters one by one according to the order of sensitivity, so that the model simulation error is gradually reduced during the parameter optimization process. The present invention quantitatively calculates the degree to which the parameters affect the model results through parameter sensitivity analysis, and gradually improves the model accuracy during the parameter optimization process, which can provide a better solution for water quality model parameter calibration.

Description

Technical Field

[0001] The invention relates to a method and system for automatically optimizing water quality model parameters, belonging to the technical field of water environment mathematical model parameter optimization. Background Art

[0002] Accurately capturing the current status of water environment is a common problem faced by environmental researchers. For common environmental water bodies, environmental factors have great heterogeneity in time and space. As conventional methods for environmental monitoring, field monitoring, indoor experiments and numerical simulation are three commonly used methods. Among them, for water bodies with large temporal and spatial heterogeneity, field monitoring and indoor experiments often require large monitoring costs. At this time, numerical simulation has become an indispensable tool for obtaining continuous temporal and spatial changes in environmental factors due to its low cost, strong operability and the ability to achieve scenario prediction functions. However, there are a large number of parameters that need to be calibrated in the numerical simulation process, and the parameter values ​​will directly affect the simulation accuracy of the model.

[0003] The highly nonlinear characteristics of water quality model parameters and the tendency to have the phenomenon of "different parameters with the same effect" make the parameter optimization process the focus and difficulty in the modeling process. On the one hand, the sensitivity of each parameter to the model output results is different, and the order of parameter adjustment is difficult to determine. It is of great significance to perform sensitivity analysis on the parameters; on the other hand, improving model accuracy is the ultimate goal of parameter adjustment. How to make the simulation error in the parameter adjustment process move in the direction of improving accuracy is the difficulty of the parameter calibration process. Therefore, it is necessary to provide a method to determine the parameters that have a greater impact on the model results, and make the model simulation error move in the direction of reduction during the parameter adjustment process. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide a method and system for automatically optimizing water quality model parameters.

[0005] In order to solve the above technical problems, the present invention provides a method for automatically optimizing water quality model parameters, comprising:

[0006] Access to pre-built water quality models based on the relationship between water quality parameters and water quality objectives;

[0007] Obtaining the actual value of a water quality parameter before and after a change in a water quality parameter at the observation site;

[0008] Utilizing the water quality model to calculate the simulated water quality values ​​before and after the water quality parameter of the observation site changes while other water quality parameters do not change;

[0009] According to the actual value of the water quality parameter before and after the change and the water quality simulation value before and after the change, the sensitivity index SI of the water quality parameter is calculated to obtain the sensitivity index SI calculation result of the water quality parameter;

[0010] Obtain the sensitivity index SI calculation result of each water quality parameter; sort the water quality parameters according to the sensitivity index SI calculation result to determine the sensitivity order of the water quality parameters;

[0011] The parameters are optimized one by one according to the order of sensitivity of water quality parameters to determine the optimal value of each water quality parameter.

[0012] Furthermore, the water quality targets include: COD, ammonia nitrogen, total nitrogen, and total phosphorus, and each water quality target also includes water quality parameters corresponding to the water quality target.

[0013] Furthermore, the process of constructing the water quality model includes:

[0014] Determine the water quality target of the model to be constructed and the water quality parameters related to the water quality target, and construct a BP neural network proxy water quality model between the water quality parameters and the water quality target according to the water quality target and the water quality parameters related to the water quality target; the BP neural network proxy water quality model adopts a three-layer structure, including 1 input layer, 1 hidden layer, and 1 output layer, and is a two-layer feedforward network with a sigmoid activation function and linear output neurons.

[0015] Furthermore, the calculation of the sensitivity index SI of the water quality parameter includes:

[0016] The sensitivity index SI was calculated using the one variable at a time method. The calculation formula of the sensitivity index SI is:

[0017]

[0018] Where, SI x is the sensitivity index of water quality parameter x; V x represents the value of the x-th water quality parameter, x=1,…,m,f i (V1,...,V x ,...,V m ) is the water quality simulation value of the ith observation station, n represents the number of water quality targets, Δ represents the change in the value of the water quality parameter at the ith observation station at the xth observation station, and the sensitivity index SI x The larger the value, the more sensitive the water quality parameter x is.

[0019] Furthermore, the optimization of the parameters one by one according to the sensitivity order of the water quality parameters includes:

[0020] Select the value V of the water quality parameter with the highest water quality parameter sensitivity x , using Latin hypercube sampling method to extract several groups of parameter values ​​within the range of water quality parameters, keeping the values ​​of other water quality parameters unchanged;

[0021] The values ​​of all groups of water quality parameters obtained by sampling are respectively brought into the BP neural network proxy water quality model to simulate the water quality target, and the error between the water quality target simulation value and the water quality measured value is calculated using the root mean square error RMSE calculation method and the Nash coefficient NSE calculation method respectively;

[0022] Select the value V of water quality parameter x when RMSE is minimum and NSE is maximum x As the optimal value of this water quality parameter;

[0023] The value of the most sensitive water quality parameter V x After determination, keep it at the optimal value, select the water quality parameter with the next highest sensitivity index, calculate the optimal value of the water quality parameter corresponding to the sensitivity index, and so on, until the optimal values ​​of all water quality parameters are determined.

[0024] Furthermore, the Latin hypercube sampling method includes:

[0025] If the number of pre-sampling groups is l, the value range of each water quality parameter is divided into l small intervals according to the pre-acquired prior distribution data, and a value is randomly selected from each small interval.

[0026] Furthermore, the calculation formula of the root mean square error RMSE calculation method is:

[0027]

[0028] The calculation formula of the Nash coefficient NSE is:

[0029]

[0030] In the formula, O t represents the measured water quality value at the tth moment; P t represents the simulated water quality value at the tth moment, O tavg represents the average value of the measured values, and N represents the number of time series.

[0031] A water quality model parameter automatic optimization system, comprising:

[0032] An acquisition module, for acquiring a pre-built water quality model based on the relationship between water quality parameters and water quality targets;

[0033] The first calculation module is used to obtain the actual value of a water quality parameter before and after a change in a water quality parameter of the observation site; use the water quality model to calculate the water quality simulation value before and after the water quality parameter of the observation site changes while other water quality parameters do not change; according to the actual value of the water quality parameter before and after the change and the water quality simulation value before and after, calculate the sensitivity index SI of the water quality parameter to obtain the sensitivity index SI calculation result of the water quality parameter;

[0034] The sorting module is used to obtain the sensitivity index SI calculation result of each water quality parameter; sort the water quality parameters according to the sensitivity index SI calculation result to determine the sensitivity order of the water quality parameters;

[0035] The determination module is used to optimize the parameters one by one according to the sensitivity order of water quality parameters and determine the optimal value of each water quality parameter.

[0036] Furthermore, the acquisition module includes a model building module,

[0037] Used to determine the water quality target of the model to be constructed and the water quality parameters related to the water quality target, and construct a BP neural network proxy water quality model between the water quality parameters and the water quality target according to the water quality target and the water quality parameters related to the water quality target;

[0038] The BP neural network proxy water quality model adopts a three-layer structure, including one input layer, one hidden layer, and one output layer, and is a two-layer feedforward network with a sigmoid activation function and linear output neurons.

[0039] Furthermore, the determination module includes:

[0040] Sampling module, used to select the value V of the water quality parameter with the highest sensitivity to water quality parameters x , using Latin hypercube sampling method to extract several groups of parameter values ​​within the range of water quality parameters, keeping the values ​​of other water quality parameters unchanged;

[0041] The second calculation module is used to bring the values ​​of all groups of water quality parameters obtained by sampling into the BP neural network proxy water quality model to simulate the water quality target, and calculate the error between the water quality target simulation value and the water quality measured value using the root mean square error RMSE calculation method and the Nash coefficient NSE calculation method respectively; select the value V of the water quality parameter x when the RMSE is the smallest and the NSE is the largest x As the optimal value of this water quality parameter;

[0042] The circulation module is used to obtain the value of the most sensitive water quality parameter V x After determination, keep it at the optimal value, select the water quality parameter with the next highest sensitivity index, calculate the optimal value of the water quality parameter corresponding to the sensitivity index, and so on, until the optimal values ​​of all water quality parameters are determined.

[0043] The beneficial effects achieved by the present invention are:

[0044] The present invention constructs a BP neural network proxy model between water quality model parameters and water quality targets, calculates the sensitivity index SI by a variable-at-a-time method, and optimizes the parameters one by one according to the order of sensitivity, so that the model simulation error is gradually reduced during the parameter optimization process. The present invention quantitatively calculates the degree to which the parameters affect the model results through parameter sensitivity analysis, and gradually improves the model accuracy during the parameter optimization process, which can provide a better solution for water quality model parameter calibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 Schematic diagram of the analytical method of the present invention;

[0046] Figure 2 This figure shows the changes in model simulation errors (RMSE and NSE) during parameter optimization in the implementation case of the present invention. DETAILED DESCRIPTION

[0047] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0048] like Figure 1 As shown, a method for automatic optimization of water quality model parameters includes the following steps:

[0049] A1. Construct a water quality model. Determine the model output water quality target and influencing parameters, and construct a BP neural network proxy water quality model between water quality parameters and water quality targets.

[0050] A2. Parameter sensitivity analysis: The sensitivity index SI was calculated using the one variable at a time method (OAT), and the parameters were ranked according to their sensitivity.

[0051] A3. Optimize parameters one by one. Optimize the parameters one by one according to the order of parameter sensitivity, and calculate the change of model simulation error during the optimization process.

[0052] The water quality targets in step A1 include: COD, ammonia nitrogen, total nitrogen, and total phosphorus.

[0053] The BP neural network proxy water quality model in step A1 adopts a three-layer structure, including one input layer, one hidden layer, and one output layer, and is a two-layer feedforward network with a sigmoid activation function and linear output neurons.

[0054] The influencing parameter information of the water quality model determined in step A1 includes its value range and prior distribution.

[0055] The calculation formula of the sensitivity index SI to be calculated using the one variable at a time method in step A2 is:

[0056]

[0057] In the formula, SI x is the sensitivity index of water quality parameter x; V x represents the value of the x-th water quality parameter, x=1,…,m,f i (V1,...,V x ,...,V m ) is the water quality simulation value of the ith observation station, n represents the number of water quality targets, Δ represents the change in the value of the water quality parameter at the ith observation station at the xth observation station, and the sensitivity index SI x The larger the value, the more sensitive the water quality parameter x is.

[0058] The model simulation error in step A3 is characterized by root mean square error (RMSE) and Nash coefficient (NSE):

[0059]

[0060] In the formula, O t represents the measured water quality value at the tth moment; P t represents the simulated water quality value at the tth moment, O tavg represents the average value of the measured values, and N represents the number of time series.

[0061] The specific steps of optimizing the parameters one by one in step A3 are as follows:

[0062] B1. First select the parameter V with the highest sensitivity x (Its SI x The highest value) is selected, and 200 sets of parameter values ​​are extracted using Latin hypercube sampling within its value range.

[0063] B2. The 200 sets of parameter values ​​obtained by sampling are respectively brought into the BP neural network proxy water quality model to participate in water quality simulation, and the errors (RMSE and NSE) between the water quality simulation values ​​and the water quality measured values ​​are calculated.

[0064] B3. Select parameter V when RMSE is minimum and NSE is maximum x The value of is taken as the optimal value of this parameter.

[0065] B4, the most sensitive parameter V x After determination, keep it at the optimal value, select the parameter with the next highest sensitivity index, repeat the parameter sampling, model calculation and optimal value selection process in step 1, and so on, until all parameters have determined the optimal values.

[0066] The specific sampling steps of the Latin hypercube sampling method in step B1 are as follows: if the number of pre-sampling groups is n, the value range of each water quality parameter is divided into n small intervals according to the prior distribution, and a value is randomly selected from each small interval.

[0067] Embodiment, the specific implementation steps of the present invention are as follows:

[0068] Step 1: Build a water quality model. Determine the water quality target of the model output. In this implementation, ammonia nitrogen is used as the water quality target, and the influencing parameters are [k MNH4 ,θ NH4 ,k NH4 ,k NO3,1 ,k NO3 ,2], the meaning of each parameter is shown in Table 1. A BP neural network proxy water quality model between water quality parameters and water quality target ammonia nitrogen is constructed. The BP neural network proxy water quality model adopts a three-layer structure, including 1 input layer, 1 hidden layer, and 1 output layer, and is a two-layer feedforward network with a sigmoid activation function and linear output neurons.

[0069] The water quality targets and the corresponding water quality parameters are shown in Table 1.

[0070] Table 1

[0071]

[0072]

[0073] Step 2, determine the influencing parameter information of the water quality model, including the preset value range and prior distribution. Prior distribution refers to the assumed parameter distribution obtained through experience or literature. Some parameter information is shown in Table 2. In this case, the prior distribution of parameters is uniform.

[0074] Table 2 Range of water quality parameters

[0075]

[0076] Step 3: Parameter sensitivity analysis. The sensitivity index SI is calculated using the one variable at a time method (OAT):

[0077]

[0078] Where, SI x is the sensitivity index of water quality parameter x; f i (V1,...,V x ,...,V m ) is the value of the ith observation station when the parameter x changes by Δ = 0.1V x Water quality simulation value at . Sensitivity index SI xThe larger the value, the more sensitive the water quality parameter x is. The sensitivity calculation formula of each parameter in this embodiment is:

[0079]

[0080] According to the sensitivity index, the sensitivity of the five water quality parameters is ranked as follows: θ NH4 >k NH4 >k MNH4 >k NO3,2 >k NO3,1

[0081] Step 4: Optimize parameters one by one. Optimize the parameters one by one according to the order of parameter sensitivity, and calculate the change of model simulation error during the optimization process. The model simulation error is characterized by the root mean square error (RMSE) and the Nash coefficient (NSE):

[0082]

[0083] In the formula, O t ——The measured value of water quality at the tth moment; P t ——Simulated water quality value at the tth moment.

[0084] Furthermore, the steps of optimizing parameters one by one are:

[0085] B1. First, select the parameter θ with the highest sensitivity NH4 , 200 sets of parameter values ​​are extracted using Latin hypercube sampling within its value range.

[0086] B2. The 200 sets of parameter values ​​obtained by sampling are respectively brought into the BP neural network proxy water quality model to participate in water quality simulation, and the errors (RMSE and NSE) between the water quality simulation values ​​and the water quality measured values ​​are calculated.

[0087] B3, select parameter θ when the minimum RMSE value is 0.0294 and the maximum NSE value is 0.9345 NH4 The value of 1.126 is taken as the optimal value of this parameter.

[0088] B4, the most sensitive parameter θ NH4 After determination, keep it at the optimal value and select the parameter k with the second highest sensitivity index. NH4 Repeat the parameter sampling, model calculation and optimal value selection process in step 1, and select the parameter k when the minimum RMSE value is 0.0025 and the maximum NSE value is 0.9995 NH4 The value of 0.1062 is taken as the optimal value of this parameter.

[0089] B5. And so on, determine the parameter k MNH4The optimal value of parameter k is 0.7476 (RMSE = 0.0024, NSE = 0.9996). NO3,2 The optimal value is 0.1976 (RMSE = 0.0019, NSE = 0.9997), and the parameter k NO3,1 The optimal value is 0.2057 (RMSE = 0.0009, NSE = 0.9999). The changes of RMSE and NSE of model simulation error during parameter optimization are shown in Figure 2. Figure 2 .

[0090] Correspondingly, the present invention also provides a water quality model parameter automatic optimization system, comprising:

[0091] An acquisition module, for acquiring a pre-built water quality model based on the relationship between water quality parameters and water quality targets;

[0092] The first calculation module is used to obtain the actual value of a water quality parameter before and after a change in a water quality parameter of the observation site; use the water quality model to calculate the water quality simulation value before and after the water quality parameter of the observation site changes while other water quality parameters do not change; according to the actual value of the water quality parameter before and after the change and the water quality simulation value before and after, calculate the sensitivity index SI of the water quality parameter to obtain the sensitivity index SI calculation result of the water quality parameter;

[0093] The sorting module is used to obtain the sensitivity index SI calculation result of each water quality parameter; sort the water quality parameters according to the sensitivity index SI calculation result to determine the sensitivity order of the water quality parameters;

[0094] The determination module is used to optimize the parameters one by one according to the sensitivity order of water quality parameters and determine the optimal value of each water quality parameter.

[0095] Furthermore, the acquisition module includes a model building module,

[0096] Used to determine the water quality target of the model to be constructed and the water quality parameters related to the water quality target, and construct a BP neural network proxy water quality model between the water quality parameters and the water quality target according to the water quality target and the water quality parameters related to the water quality target;

[0097] The BP neural network proxy water quality model adopts a three-layer structure, including one input layer, one hidden layer, and one output layer, and is a two-layer feedforward network with a sigmoid activation function and linear output neurons.

[0098] Furthermore, the determination module includes:

[0099] Sampling module, used to select the value V of the water quality parameter with the highest sensitivity to water quality parameters x, using Latin hypercube sampling method to extract several groups of parameter values ​​within the range of water quality parameters, keeping the values ​​of other water quality parameters unchanged;

[0100] The second calculation module is used to bring the values ​​of all groups of water quality parameters obtained by sampling into the BP neural network proxy water quality model to simulate the water quality target, and calculate the error between the water quality target simulation value and the water quality measured value using the root mean square error RMSE calculation method and the Nash coefficient NSE calculation method respectively; select the value V of the water quality parameter x when the RMSE is the smallest and the NSE is the largest x As the optimal value of this water quality parameter;

[0101] The circulation module is used to obtain the value of the most sensitive water quality parameter V x After determination, keep it at the optimal value, select the water quality parameter with the next highest sensitivity index, calculate the optimal value of the water quality parameter corresponding to the sensitivity index, and so on, until the optimal values ​​of all water quality parameters are determined.

[0102] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0103] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, 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 generate 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 flowchart and / or block diagram. 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.

[0104] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0105] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0106] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for automatic optimization of water quality model parameters, characterized in that: include: Access to pre-built water quality models based on the relationship between water quality parameters and water quality objectives; Obtain the actual value of a water quality parameter before and after the change of a water quality parameter at the observation site; Utilizing the water quality model to calculate the simulated water quality values ​​before and after the water quality parameter of the observation site changes while other water quality parameters do not change; According to the actual value of the water quality parameter before and after the change and the water quality simulation value before and after the change, the sensitivity index SI of the water quality parameter is calculated to obtain the sensitivity index SI calculation result of the water quality parameter; Obtain the sensitivity index SI calculation result of each water quality parameter; sort the water quality parameters according to the sensitivity index SI calculation result to determine the sensitivity order of the water quality parameters; Optimize the parameters one by one according to the sensitivity order of water quality parameters to determine the optimal value of each water quality parameter; The water quality targets include: COD, ammonia nitrogen, total nitrogen, and total phosphorus, and each water quality target also includes water quality parameters corresponding to the water quality target; The process of constructing the water quality model includes: Determine the water quality target of the model to be constructed and the water quality parameters related to the water quality target, and construct a BP neural network proxy water quality model between the water quality parameters and the water quality target according to the water quality target and the water quality parameters related to the water quality target; the BP neural network proxy water quality model adopts a three-layer structure, including one input layer, one hidden layer, and one output layer, and is a two-layer feedforward network with a sigmoid activation function and linear output neurons; The calculation of the sensitivity index SI of the water quality parameter comprises: The sensitivity index SI was calculated using the one variable at a time method. The calculation formula of the sensitivity index SI is: ； In the formula, SI x Is the water quality parameter x sensitivity index; V x represents the value of the x-th water quality parameter, x =1,…, m , It is i The water quality simulation value of each observation station, n represents the number of water quality targets, Δ represents the i The observation station is x The change in the value of the water quality parameter, the sensitivity index SI x The larger the water quality parameter x The more sensitive; The optimization of the parameters one by one according to the sensitivity order of the water quality parameters includes: Select the value of the water quality parameter with the highest water quality parameter sensitivity V x , using Latin hypercube sampling method to extract several groups of parameter values ​​within the range of water quality parameters, keeping the values ​​of other water quality parameters unchanged; The values ​​of all groups of water quality parameters obtained by sampling are respectively brought into the BP neural network proxy water quality model to simulate the water quality target, and the error between the water quality target simulation value and the water quality measured value is calculated using the root mean square error RMSE calculation method and the Nash coefficient NSE calculation method respectively; Select the water quality parameters when RMSE is minimum and NSE is maximum x The value of V x As the optimal value of this water quality parameter; The value of the most sensitive water quality parameter V x After determination, keep it at the optimal value, select the water quality parameter with the next highest sensitivity index, calculate the optimal value of the water quality parameter corresponding to the sensitivity index, and so on, until the optimal values ​​of all water quality parameters are determined.

2. The method for automatic optimization of water quality model parameters according to claim 1, characterized in that: The Latin hypercube sampling method comprises: If the number of pre-sampling groups is l , then the value range of each water quality parameter is divided into l small intervals, and randomly select values ​​from each small interval.

3. The method for automatic optimization of water quality model parameters according to claim 1, characterized in that: The calculation formula of the root mean square error RMSE calculation method is: ； The calculation formula of the Nash coefficient NSE is: ； In the formula, O t Indicates t The measured water quality value at a certain moment; P t Indicates t The simulated water quality value at the moment, O tavg represents the average value of the measured values, N Indicates the number of time series.

4. A system based on the method for automatic optimization of water quality model parameters according to any one of claims 1 to 3, characterized in that: include: An acquisition module, for acquiring a pre-built water quality model based on the relationship between water quality parameters and water quality targets; The first calculation module is used to obtain the actual value of a water quality parameter before and after a change in a water quality parameter of an observation site; use the water quality model to calculate the water quality simulation value before and after the water quality parameter of the observation site changes while other water quality parameters do not change; according to the actual value of the water quality parameter before and after the change and the water quality simulation value before and after, perform sensitivity index SI calculation on the water quality parameter to obtain the sensitivity index SI calculation result of the water quality parameter; The sorting module is used to obtain the sensitivity index SI calculation result of each water quality parameter; sort the water quality parameters according to the sensitivity index SI calculation result to determine the sensitivity order of the water quality parameters; The determination module is used to optimize the parameters one by one according to the sensitivity order of water quality parameters and determine the optimal value of each water quality parameter.

Images