Effluent prediction and operation effect evaluation method of reclaimed water plant

Abstract

The present application discloses a method for predicting the effluent of a reclaimed water plant and evaluating its operation effect. The method may include: determining process parameters; reducing and screening the process parameters based on the partial least squares method to determine the optimization parameters; constructing a partial least squares regression model of the optimization parameters and the effluent data, recorded as an effluent prediction model; and predicting the effluent of the reclaimed water plant and evaluating its operation effect through the effluent prediction model. The present invention targets the key operation links of the reclaimed water plant, screens out the measured variables that are most relevant to the effluent, predicts the effluent, optimizes the process parameters, and establishes an inlet warning system.

Description

Technical Field

[0001] The present invention relates to the field of water plant operation effect evaluation, and more specifically, to a method for predicting effluent from a reclaimed water plant and evaluating operation effect. Background Art

[0002] Sustainable development has become the focus of many industries. With the continuous increase in urban water consumption and the number of urban sewage treatment plants, sewage treatment capacity has also been continuously improved. How to evaluate the operating status of sewage treatment plants and propose reasonable improvement measures to guide the sustainable development of sewage treatment is a major issue facing the sewage treatment industry.

[0003] Wastewater treatment is a complex system engineering, which is closely related to society, economy, environment, technology and other aspects. In the evaluation process, it is particularly important to select scientific, objective and representative evaluation indicators that reflect the performance level of sewage treatment plants according to different goals and standards.

[0004] Therefore, it is necessary to develop a method for predicting effluent from reclaimed water plants and evaluating their operation performance.

[0005] The information disclosed in the background technology section of the present invention is only intended to deepen the understanding of the general background technology of the present invention, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0006] The present invention proposes a method, electronic equipment and medium for predicting effluent from a reclaimed water plant and evaluating its operation effect, which can screen out the measured variables most relevant to effluent from key operation links of the reclaimed water plant, predict effluent, optimize process parameters and establish an early warning system for water inlet.

[0007] The embodiment of the present invention provides a method for predicting effluent from a reclaimed water plant and evaluating operating effects, comprising:

[0008] Determine process parameters;

[0009] Based on the partial least squares method, the process parameters are reduced in dimension and screened to determine the optimization parameters;

[0010] A partial least squares regression model of the optimization parameters and the water discharge data is constructed, which is denoted as the water discharge prediction model;

[0011] The effluent prediction model is used to predict the effluent of the reclaimed water plant and evaluate the operation effect.

[0012] Preferably, the process parameters are reduced in dimension and screened based on the partial least squares method, and the optimization parameters are determined to include:

[0013] Establishing a process parameter variable set and an effluent variable set, and then establishing a linear combination of the process parameter variable set and the effluent variable set to obtain candidate process parameters;

[0014] A partial least squares regression model of the candidate process parameters and effluent data is constructed, and the candidate process parameters whose standard regression coefficient ratio is greater than a set threshold are selected as optimization parameters.

[0015] Preferably, establishing a linear combination of the process parameter variable set and the effluent variable set includes component extraction and cross-validation.

[0016] Preferably, constructing a partial least squares regression model of alternative process parameters and effluent data includes:

[0017] The number of candidate process parameters is m, that is, the extracted candidate process parameters are x 1 、x 2 ,…x m , then the partial least squares regression is used to realize the relationship between the water discharge data and x 1 、x 2 ,…x m The partial least squares regression model of the alternative process parameters and effluent data is constructed as follows:

[0018]

[0019] Among them, p i For x i The corresponding regression coefficient.

[0020] Preferably, the candidate process parameters whose proportion of the screening standard regression coefficient is greater than the set threshold as optimization parameters include:

[0021] The formula (1) is standardized as follows:

[0022]

[0023] Among them, P i For X i * The corresponding standardized regression coefficients;

[0024] Assign the standard regression coefficient in formula (2) as the weight value of the corresponding candidate process parameter, and sort the absolute values ​​of the weight values ​​of the candidate process parameters;

[0025] Select the candidate process parameters whose standard regression coefficient ratio is greater than the set threshold in order as the optimization parameters, that is:

[0026]

[0027] Among them, n is the number of optimization parameters and z is the set threshold.

[0028] Preferably, the water output prediction model is:

[0029]

[0030] Preferably, predicting the effluent from a reclaimed water plant using the effluent prediction model includes:

[0031] The values ​​of the optimization parameters corresponding to the water output data to be measured are substituted into the water output prediction model to predict the water output data.

[0032] Preferably, evaluating the operation effect of the reclaimed water plant by using the effluent prediction model includes:

[0033] The formula (4) is standardized as follows:

[0034]

[0035] Among them, P i For X i * The corresponding standardized regression coefficients;

[0036] Sort the absolute values ​​of the standard regression coefficients of formula (5) and assign them as weight values ​​of the optimization parameters;

[0037] According to the weight values ​​of the optimization parameters, the priority of the process control parameters for abnormal water discharge conditions is determined from high to low.

[0038] Preferably, evaluating the operation effect of the reclaimed water plant by using the effluent prediction model includes:

[0039] According to the sewage treatment process principle, determine whether the optimization parameters and their weight value ranking are contrary to the sewage treatment process principle. If so, optimize the optimization parameters according to the sewage treatment process principle.

[0040] Preferably, evaluating the operation effect of the reclaimed water plant by using the effluent prediction model includes:

[0041] According to the effluent prediction model, the optimization parameters under different working conditions are empirically assigned, and the effluent data under the corresponding working conditions and the maximum removal capacity of the biological treatment system are calculated;

[0042] If the total amount of pollutants in the influent exceeds the maximum removal capacity of the biological treatment system, external agents are used to remove the pollutants, thereby formulating the boundary conditions for agent addition;

[0043] If the total amount of effluent pollutants exceeds the pollutant emission standard of the reclaimed water plant, the corresponding influent pollutant input concentration will be used as the influent warning value, and an influent early warning system will be established.

[0044] Its beneficial effects are:

[0045] Aiming at the key operating links of the reclaimed water plant, the measured variables most relevant to the effluent are screened out, the effluent is predicted, the process parameters are optimized, and an inlet early warning system is established. It can be used for the evaluation of the effects of typical large-scale, multi-process, non-stationary, strongly coupled, and nonlinear sewage treatment processes.

[0046] The method of the present invention has other characteristics and advantages, which will be apparent from the accompanying drawings and subsequent detailed descriptions incorporated herein, or will be described in detail in the accompanying drawings and subsequent detailed descriptions incorporated herein, which together serve to explain the specific principles of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0048] Figure 1 A flow chart showing the steps of a method for predicting effluent from a reclaimed water plant and evaluating operating effects according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0050] Figure 1 A flow chart showing the steps of a method for predicting effluent from a reclaimed water plant and evaluating operating effects according to an embodiment of the present invention.

[0051] like Figure 1 As shown, the present invention provides a method for predicting effluent from a reclaimed water plant and evaluating operating effects, including:

[0052] Step 101, determining process parameters;

[0053] Step 102, reducing the dimension and screening the process parameters based on the partial least squares method to determine the optimization parameters;

[0054] Step 103, constructing a partial least squares regression model of the optimization parameters and the water discharge data, which is recorded as a water discharge prediction model;

[0055] Step 104, predicting the effluent of the reclaimed water plant and evaluating the operation effect through the effluent prediction model.

[0056] In one example, the process parameters are reduced in dimension and screened based on the partial least squares method, and the optimization parameters are determined to include:

[0057] Establishing a process parameter variable set and an effluent variable set, and then establishing a linear combination of the process parameter variable set and the effluent variable set to obtain candidate process parameters;

[0058] A partial least squares regression model of alternative process parameters and effluent data was constructed, and alternative process parameters whose standard regression coefficient ratio was greater than the set threshold were selected as optimization parameters.

[0059] In one example, establishing a linear combination of a process parameter variable set and a effluent variable set includes component extraction and cross-validation.

[0060] In one example, building a partial least squares regression model of candidate process parameters and effluent data includes:

[0061] The number of candidate process parameters is m, that is, the extracted candidate process parameters are x 1 、x 2 ,…x m , then the partial least squares regression is used to realize the relationship between the water discharge data and x 1 、x 2 ,…x m The partial least squares regression model of the alternative process parameters and the effluent data is constructed as follows:

[0062]

[0063] Among them, p i For x i The corresponding regression coefficient.

[0064] In one example, the candidate process parameters whose proportion of the standard regression coefficient is greater than a set threshold are selected as optimization parameters, including:

[0065] The formula (1) is standardized as follows:

[0066]

[0067] Among them, P i For X i * The corresponding standardized regression coefficients;

[0068] Assign the standard regression coefficient in formula (2) as the weight value of the corresponding candidate process parameter, and sort the absolute values ​​of the weight values ​​of the candidate process parameters;

[0069] Select the candidate process parameters whose standard regression coefficient ratio is greater than the set threshold in order as the optimization parameters, that is:

[0070]

[0071] Among them, n is the number of optimization parameters and z is the set threshold.

[0072] In one example, the water output prediction model is:

[0073]

[0074] In one example, the effluent prediction model is used to predict the effluent of a reclaimed water plant, including:

[0075] Substitute the values ​​of the optimized parameters corresponding to the water output data to be measured into the water output prediction model to predict the water output data.

[0076] In one example, the evaluation of the operation effect of a reclaimed water plant through an effluent prediction model includes:

[0077] The formula (4) is standardized as follows:

[0078]

[0079] Among them, P i For X i * The corresponding standardized regression coefficients;

[0080] Sort the absolute values ​​of the standard regression coefficients of formula (5) and assign them as weight values ​​of the optimization parameters;

[0081] According to the weight value of the optimization parameter, the priority of the process control parameters for abnormal water discharge conditions is determined from high to low.

[0082] In one example, the evaluation of the operation effect of a reclaimed water plant using an effluent prediction model includes:

[0083] According to the principles of sewage treatment technology, determine whether the order of optimization parameters and their weight values ​​is contrary to the principles of sewage treatment technology. If so, optimize the optimization parameters according to the principles of sewage treatment technology.

[0084] In one example, the evaluation of the operation effect of a reclaimed water plant through an effluent prediction model includes:

[0085] According to the effluent prediction model, the optimization parameters under different working conditions are empirically assigned, and the effluent data under the corresponding working conditions and the maximum removal capacity of the biological treatment system are calculated;

[0086] If the total amount of pollutants in the influent exceeds the maximum removal capacity of the biological treatment system, external agents are used to remove the pollutants, thereby formulating the boundary conditions for agent addition;

[0087] If the total amount of effluent pollutants exceeds the pollutant emission standard of the reclaimed water plant, the corresponding influent pollutant input concentration will be used as the influent warning value, and an influent early warning system will be established.

[0088] Specifically, firstly, based on the process mechanism of sewage treatment and the measurability of influent and process parameters, data on effluent-related variables are collected. Based on the multicollinearity characteristics of influent and process parameters, the partial least squares principle is used to perform weight analysis on the collected data, and the process parameters that are significantly related to the important effluent water quality items in the influent and reclaimed water plant sewage treatment process are screened out, which are called optimization parameters. Secondly, the selected optimization parameters are used to construct the effluent prediction model of the reclaimed water plant. Thirdly, the optimization parameters are weighted according to the model results to determine the priority of the process control parameters for abnormal effluent conditions. Then, the rationality of the optimization parameter priority is evaluated according to the sewage treatment process principle, and the process operation parameters are optimized. Finally, different working conditions are assigned values ​​according to the effluent prediction model, the boundary conditions for the addition of reagents and the warning value of the reclaimed water plant influent are calculated, and an influent early warning system is established.

[0089] According to the technological mechanism of the sewage treatment process and the measurability of the influent and process parameters, data on effluent-related variables are collected, and all process parameters related to important effluent water quality items are listed. Based on the multicollinearity characteristics, the partial least squares principle is used for analysis, and the process parameters are reduced in dimension to screen out the process parameters that are significantly correlated with important effluent water quality items in the sewage treatment process of the reclaimed water plant, which are called optimization parameters.

[0090] Collect data based on the process mechanism of sewage treatment and the measurability of influent and process parameters, such as influent water quality, biological process control variables such as sludge concentration, biological pool dissolved oxygen, ORP, water temperature, sludge discharge, etc. Establish process parameter variable set X = {x 1 、x 2 , …, x p} r and the water output variable set Y = {y 1 ,y 2 , …, y q} r , call the zscore function in the MATLAB software to standardize the process parameter variable set and the effluent variable set.

[0091] The linear combinations of process parameter variable sets and effluent variable sets are established as follows:

[0092]

[0093]

[0094] Among them, t 1 is a linear combination of the process parameter variable set X, u 1is a linear combination of the water output variable set Y, X = {x 1 、x 2 , …, x p} r , Y={y 1 ,y 2 , …, y q} r .

[0095] According to the principle of partial least squares, the components are extracted and the first component t 1 Extraction:

[0096] The two sets of variables X and Y constitute the observation matrix X α , Y α , the score vector of the first pair of components can be calculated separately for:

[0097]

[0098]

[0099] Among them, for The calculation formula of the combination coefficient Yes X α The first axis of for The calculation formula of the combination coefficient Yes α The first axis of This requirement is equivalent to requiring t 1 with u 1 The covariance of To achieve the maximum, the problem comes down to calculating the matrix X a T Y α Y a T X α The eigenvalues ​​and eigenvectors of . That is, the following three conditions are met:

[0100]

[0101]

[0102]

[0103] The calculations show:

[0104]

[0105]

[0106] Among them, r(xp ,y) represents x p Correlation coefficient with y; w 1 Yes X a T Y α Y a T X α The maximum eigenvalue θ 1 2 The corresponding unit eigenvector is

[0107] Get w 1 , we can get t 1 , so we can get X a , Y a t 1 The regression equation is:

[0108]

[0109] Y a =t 1 s 1 +Y β (16)

[0110] in, X β , Y β is the residual matrix of the regression equation.

[0111] For the second component t 2 Extraction:

[0112] After the first component is extracted, perform X on t 1 The regression of Y on u 1 If the regression equation reaches the required satisfactory accuracy, the algorithm is terminated; otherwise, the second pair of components is extracted and X is used to 1 The remaining information after interpretation replaces X, and Y is replaced by u 1 The explained remaining information replaces Y and the regression steps are repeated until the required satisfactory accuracy is achieved.

[0113] X β Replace X α , with Y β Replace Y α , use t 1 The method to find the second axis w 2 and the second component t 2 , where the combination coefficient is:

[0114]

[0115] Likewise, Xβ , Y β For t 2 Perform regression and get X β , Y β t 2 The regression equation is:

[0116]

[0117] Y β =t 2 s 2 +Y β (19)

[0118] For the extraction of the hth component:

[0119] The extraction principle of the hth component is the same as that of the first and second components. And in the process of extracting each component, it is necessary to ensure that the correlation between each two variable groups is maximized. The number of h needs to be less than the rank of the observation matrix, and the number of h can be identified using the cross-validation principle.

[0120] Cross-validation test:

[0121] The operation process of the cross-validation test is to add one more component when the number of components is determined to be h when the model is available, so that the total number of components becomes h + 1. When the number of components becomes h + 1, if it makes the model unusable or the results are not significantly improved, it can be proved that the number of components should be h.

[0122] There are m extracted components (t 1 ,t 2 ,…,t m ), It is the fitting value of the i-th sample point obtained after regression using all sample points and m components, which is recorded as y g(i-1) is the fitting value of the i-1th sample point obtained after deleting sample point i but retaining m components for regression, denoted as

[0123] Based on the above theory, we define Used to determine whether the number of components h is valid. The calculation method is:

[0124]

[0125] according to It is easy to see from the calculation method that every time a component is obtained, the cross-test can be used to determine the number of components, avoiding the occurrence of meaningless operations.

[0126] It is generally believed that if there is When , the model accuracy requirement is high and the component extraction can be stopped; When the contribution of the added ingredients is significant, the ingredients should continue to be extracted.

[0127] After determining the final components extracted from X, the candidate process parameters can be determined. The number of candidate process parameters is m, that is, the extracted candidate process parameters are x 1 、x 2 ,…x m , then the partial least squares regression is used to realize the relationship between the water discharge data and x 1 、x 2 ,…x m The partial least squares regression model of the alternative process parameters and the effluent data is constructed as formula (1). Formula (1) is standardized to obtain formula (2). The standard regression coefficient in formula (2) is assigned as the weight value of the corresponding alternative process parameter, and the absolute values ​​of the weight values ​​of the alternative process parameters are sorted; the alternative process parameters whose standard regression coefficient ratio is greater than the set threshold are selected in order as the optimization parameters, which is formula (3). The set threshold can be determined to be 0.9.

[0128] A partial least squares regression model of the optimization parameters and the water discharge data is constructed, which is denoted as the water discharge prediction model as formula (4).

[0129] Predicting the effluent of a reclaimed water plant by using an effluent prediction model includes: substituting the values ​​of the optimization parameters corresponding to the effluent data to be measured into the effluent prediction model to predict the effluent data.

[0130] The evaluation of the operation effect of the reclaimed water plant through the effluent prediction model includes:

[0131] (1) Standardization processing is performed on the effluent prediction model to obtain formula (5), and the absolute values ​​of the standard regression coefficients of formula (5) are sorted and assigned as weight values ​​of the optimization parameters; the weights of the optimization parameters can be used to quantify the influence of the optimization parameters on the effluent under the current operating state, and the optimization parameters with large weights have a greater impact on the effluent. In the actual operation of the reclaimed water plant, if process adjustment is required, the effectiveness of the control schemes of each process parameter can be evaluated according to the calculation results of the present invention, and the control priority of the process control parameters for abnormal effluent conditions can be determined from high to low according to the weight values ​​of the optimization parameters, so as to formulate a more reasonable and effective control scheme.

[0132] (2) According to the sewage treatment process principle, determine whether the optimization parameters and their weight value ranking are contrary to the sewage treatment process principle. If so, it indicates that the operating conditions need to be optimized. Then, according to the sewage treatment process principle, optimize the optimization parameters. In the actual operation of the reclaimed water plant, the rationality of each process parameter can be evaluated according to the calculation results of the present invention. If the calculation results are contrary to the sewage treatment mechanism, the process operation parameters should be optimized according to the actual situation to improve the treatment efficiency of the biological system.

[0133] (3) Based on the effluent prediction model, empirical values ​​are assigned to the optimization parameters under different working conditions, and the effluent data under the corresponding working conditions and the maximum removal capacity of the biological treatment system are calculated; if the total amount of influent pollutants exceeds the maximum removal capacity of the biological treatment system, the pollutants are removed by adding external agents. In the actual operation of the reclaimed water plant, the boundary conditions for agent addition can be formulated based on the calculation results of the present invention to avoid effluent pollution incidents caused by insufficient agent addition and waste of resources caused by excessive agent addition; if the total amount of effluent pollutants exceeds the pollutant emission standard of the reclaimed water plant, the corresponding influent pollutant input concentration is used as the influent warning value, and an influent early warning system is established to avoid environmental pollution caused by effluent.

[0134] This method can be used to predict the effluent from reclaimed water plants, evaluate and improve the effects of sewage treatment processes, and establish an early warning system for the inlet of reclaimed water plants based on different working conditions. It also provides a quantifiable evaluation method and optimization direction for multi-process, non-stationary, strongly coupled, and nonlinear sewage treatment processes.

[0135] To facilitate understanding of the solutions and effects of the embodiments of the present invention, four specific application examples are given below. Those skilled in the art should understand that the examples are only for facilitating understanding of the present invention, and any specific details thereof are not intended to limit the present invention in any way.

[0136] Example 1

[0137] Establish a prediction model for the effluent of a reclaimed water plant to predict the effluent of the reclaimed water plant. Collect data based on the process mechanism of the sewage treatment process and the measurability of the influent and process parameters, such as influent water quality, treated water volume, biological process control variables such as sludge concentration, dissolved oxygen in the biological pool, water temperature, sludge discharge, etc., and list all process parameters related to important effluent water quality items. Perform preliminary data cleaning, call the zscore function in the matlab software, and standardize the sample matrix. Based on the multicollinearity characteristics of the influent and process parameters, the partial least squares principle is used for analysis, and the data is reduced in dimension according to the weight ranking of the optimized parameters. Construct a mathematical model between the optimized parameters and the effluent. When using the mechanism analysis and algorithm to screen variables based on 193 days of real data from a reclaimed water plant, the correlation between the main variables and the effluent total nitrogen index can be obtained as shown in Table 1.

[0138] Table 1

[0139] Parameter name unit Weight Weight ratio Whether to filter as optimization parameter Optimization parameter coefficients Water treatment capacity <![CDATA[Ten thousand m 3 > 0.077 3.32％ no Water temperature ℃ -0.485 21.03％ yes -0.11 Aerobic sludge concentration MLSS mg / L -0.429 18.63％ yes -0.00065 Aerobic section dissolved oxygen DO mg / L 0.258 11.21％ yes 0.24 Total influent COD mg / L -0.309 13.40％ yes -0.004 Total water inlet SS mg / L -0.050 2.17％ no Total water inlet TN mg / L 0.600 26.01％ yes 0.15 Total influent ammonia nitrogen mg / L -0.004 0.18％ no Total water inlet TP mg / L 0.093 4.05％ no

[0140] Based on the results of partial least squares algorithm analysis, under the current data operation state, the weight of influent and process parameters accounts for the top 90%. The parameters that should be selected as optimization parameters include: water temperature, aerobic stage sludge concentration (MLSS), aerobic stage dissolved oxygen (DO), total influent COD, total influent TN, and then the effluent prediction model of effluent TN and optimized parameters is constructed as follows:

[0141] Outlet TN (mg / L) = 8.98 + 0.15 * total inlet TN (mg / L) - 0.004 * total inlet COD (mg / L) - 0.00065 * MLSS (mg / L) + 0.24 * DO (mg / L) - 0.11 * water temperature (℃)

[0142] Substitute the values ​​of the optimized parameters corresponding to the water output data to be measured into the water output prediction model to predict the water output data.

[0143] Example 2

[0144] Sort the absolute values ​​of the weights of the optimization parameters in Example 1, quantify the degree of influence of the optimization parameters on the effluent under the current operating state, and evaluate the effectiveness of the control schemes of each process parameter accordingly, determine the priority of the control parameters, and formulate a more reasonable and effective control scheme. Among them, the one with the largest absolute value of the weight is the influent TN concentration, indicating that the effluent TN concentration is most affected by the influent TN concentration value, but the influent TN concentration belongs to the influent receiving value of the reclaimed water plant and is not within the range of controllable parameters. The variable with the largest absolute value of the weight that can be controlled is the aerobic stage sludge concentration (MLSS), and the aerobic stage sludge concentration (MLSS) should be controlled first during regulation.

[0145] Example 3

[0146] When the effluent TN needs to meet the national effluent water quality standard limit of 15 mg / L, the maximum bearing capacity of various influent indicators under different working conditions can be calculated based on the effluent prediction model, and an influent early warning system can be established. The values ​​of various influent concentrations and operation control parameters are assigned based on historical winter experience values ​​as shown in Table 2.

[0147] Table 2

[0148]

[0149]

[0150] If the total amount of influent pollutants exceeds the maximum removal capacity of the biological treatment system, for example, according to the mathematical model calculation of the example effluent TN and the optimization parameters, when the total influent TN is higher than 78.4 mg / L, the effluent TN will exceed the standard (higher than 15 mg / L), and the influent TN concentration is the influent warning value. In actual operation, the reclaimed water plant should calculate the influent warning value under different conditions according to different working conditions and establish an influent early warning system.

[0151] If the effluent concentration exceeds the national effluent water quality standard limit, in addition to adjusting the process parameters, chemicals can be added to the biological system for enhanced treatment to ensure that the effluent meets the standard. For example, according to the mathematical model calculation of the effluent TN and the optimized parameters, in condition 2, when the water temperature is lower than 11.4, the effluent TN will exceed the standard (higher than 15 mg / L), and the critical values ​​at this time are the boundary conditions for chemical addition. In actual operation, the reclaimed water plant should calculate the boundary conditions for chemical addition under different conditions according to different working conditions, establish a system to ensure that the effluent meets the standard, and avoid environmental pollution incidents and chemical waste.

[0152] Example 4

[0153] According to the sewage treatment process principle, the optimization parameter screening results and their weight values ​​are evaluated. If the optimization parameter screening and their weight ranking are contrary to the sewage treatment mechanism, it indicates that the operating conditions need to be optimized. In the actual operation of the reclaimed water plant, the rationality of each process parameter can be evaluated according to the calculation results of the present invention. If the calculation results are contrary to the sewage treatment mechanism, the process operation parameters should be optimized according to the actual situation to improve the treatment efficiency of the biological system. The above optimization parameter priority ranking is reasonable and no parameter optimization is required.

[0154] When screening variables based on 184 days of real data from a reclaimed water plant, the correlation between the main variables and the effluent total nitrogen index was obtained using mechanism analysis and algorithms, as shown in Table 3.

[0155] Table 3

[0156]

[0157]

[0158] The above-mentioned sodium acetate dosage as a carbon source was not selected into the optimization parameters, which is contrary to the principle, indicating that this indicator should be optimized. In the production of this reclaimed water plant, the carbon source dosage should be optimized according to the actual working conditions.

[0159] In summary, the present invention targets the key operating links of the reclaimed water plant, screens out the measured variables most relevant to the effluent, predicts the effluent, optimizes the process parameters, and establishes a water inlet early warning system.

[0160] Those skilled in the art should understand that the purpose of the above description of the embodiments of the present invention is only to exemplarily illustrate the beneficial effects of the embodiments of the present invention, and is not intended to limit the embodiments of the present invention to any given examples.

[0161] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for predicting effluent from a reclaimed water plant and evaluating its operating effect, characterized in that: include: Determine process parameters; Based on the partial least squares method, the process parameters are reduced in dimension and screened to determine the optimization parameters; A partial least squares regression model of the optimization parameters and the water discharge data is constructed, which is denoted as the water discharge prediction model; The effluent prediction model is used to predict the effluent of the reclaimed water plant and evaluate the operation effect; The process parameters are reduced in dimension and screened based on the partial least squares method, and the optimization parameters are determined to include: Establishing a process parameter variable set and an effluent variable set, and then establishing a linear combination of the process parameter variable set and the effluent variable set to obtain candidate process parameters; Constructing a partial least squares regression model of the candidate process parameters and the effluent data, and selecting the candidate process parameters whose standard regression coefficient ratio is greater than a set threshold as optimization parameters; Among them, establishing the linear combination of process parameter variable set and effluent variable set includes component extraction and cross-validation; The evaluation of the operation effect of the reclaimed water plant by using the effluent prediction model includes: According to the effluent prediction model, the optimization parameters under different working conditions are empirically assigned, and the effluent data under the corresponding working conditions and the maximum removal capacity of the biological treatment system are calculated; If the total amount of pollutants in the influent exceeds the maximum removal capacity of the biological treatment system, external agents are used to remove the pollutants, thereby formulating the boundary conditions for agent addition; If the total amount of effluent pollutants exceeds the pollutant discharge standard of the reclaimed water plant, the corresponding influent pollutant input concentration will be used as the influent warning value, and an influent early warning system will be established; Among them, the construction of the partial least squares regression model of alternative process parameters and effluent data includes: The number of candidate process parameters is m, that is, the extracted candidate process parameters are x1, x2, ...x m , then partial least squares regression is used to realize the relationship between the outflow data and x1, x2, …x m The partial least squares regression model of the alternative process parameters and effluent data is constructed as follows: Among them, p i For x i The corresponding regression coefficients; Among them, the candidate process parameters whose proportion of the screening standard regression coefficient is greater than the set threshold are used as optimization parameters, including: The formula (1) is standardized as follows: Among them, P i For X i * The corresponding standardized regression coefficients; Assign the standard regression coefficient of formula (2) as the weight value of the corresponding candidate process parameter, and sort the absolute values ​​of the weight values ​​of the candidate process parameters; Select the candidate process parameters whose standard regression coefficient ratio is greater than the set threshold in order as the optimization parameters, that is: Among them, n is the number of optimization parameters, and z is the set threshold; Wherein, the water output prediction model is:

2. The method for predicting effluent from a reclaimed water plant and evaluating operating results according to claim 1, wherein: The effluent prediction of the reclaimed water plant by using the effluent prediction model includes: The values ​​of the optimization parameters corresponding to the water output data to be measured are substituted into the water output prediction model to predict the water output data.

3. The method for predicting effluent from a reclaimed water plant and evaluating operating results according to claim 1, wherein: The evaluation of the operation effect of the reclaimed water plant through the effluent prediction model includes: The formula (4) is standardized as follows: Among them, P i For X i * The corresponding standardized regression coefficients; Sort the absolute values ​​of the standard regression coefficients of formula (5) and assign them as weight values ​​of the optimization parameters; According to the weight values ​​of the optimization parameters, the priority of the process control parameters for abnormal water discharge conditions is determined from high to low.

4. The method for predicting effluent from a reclaimed water plant and evaluating operating results according to claim 3, wherein: The evaluation of the operation effect of the reclaimed water plant through the effluent prediction model includes: According to the sewage treatment process principle, determine whether the optimization parameters and their weight value ranking are contrary to the sewage treatment process principle. If so, optimize the optimization parameters according to the sewage treatment process principle.

Images