A dynamic optimization system and method based on sewage treatment

Abstract

The present invention relates to the technical field of sewage treatment, and specifically discloses a dynamic optimization system and method based on sewage treatment, which specifically includes the following steps: Step 1: extract the influencing parameter values ​​of sewage treatment and the sewage treatment effect value in the historical data, and process them. The present invention calculates the similarity characterization value of the historical analysis period and the historical calculation period, calculates the similarity characterization mean and variance value based on the similarity characterization values ​​of multiple historical calculation periods and historical analysis periods, calculates the similarity evaluation value, compares the similarity evaluation value with the similarity evaluation threshold, and marks the corresponding influencing key parameters when the similarity evaluation value is less than the similarity evaluation threshold as parameters to be optimized, so that which influencing key parameters can be accurately identified in the historical data presenting abnormal behavior patterns that are highly similar to the current analysis period, and these parameters are used as key values ​​to be optimized, which is helpful to screen out the parameters to be optimized and improve the efficiency of the optimization work.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and in particular to a dynamic optimization system and method based on sewage treatment. Background Art

[0002] Sewage treatment refers to the treatment of wastewater containing various harmful substances to remove or reduce the pollutants in it so that it meets the prescribed discharge standards. It can effectively reduce the pollution of sewage to water bodies and protect the sustainable use of water resources. With the acceleration of urbanization, sewage treatment has become an important part of urban environmental protection.

[0003] However, traditional treatment methods often rely on fixed process flows and parameter settings, which are difficult to adapt to the complex and changeable sewage quality and treatment needs, and cannot be dynamically optimized. At the same time, there are many factors that affect the treatment effect, including but not limited to the chemical composition of sewage, temperature, pH value, flow rate, and operating parameters of treatment equipment. Existing methods cannot select the influencing factors that need to be optimized from the many factors. Summary of the invention

[0004] The object of the present invention is to provide a dynamic optimization system and method based on sewage treatment to solve the technical problems in the above background.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] The present invention provides a dynamic optimization method based on sewage treatment, which specifically includes the following steps: Step 1: extracting the influencing parameter values ​​and sewage treatment effect values ​​of historical data, and processing them to calculate the influencing related values; comparing the influencing related values ​​with the influencing related thresholds, if the influencing related values ​​are greater than the influencing related thresholds, generating a large influencing signal, obtaining the influencing parameter values ​​corresponding to the generated large influencing signal, and marking them as influencing key parameters; Step 2: calculating a similarity evaluation value through data processing based on the influencing key values ​​obtained by screening; comparing the similarity evaluation value with the similarity evaluation threshold, if the similarity evaluation value is less than the similarity evaluation threshold, generating a high similarity signal, obtaining the influencing key values ​​corresponding to the high similarity signal, and marking them as parameter values ​​to be optimized;

[0007] Step 3: Based on the obtained parameter values ​​to be optimized, construct the parameter value to be optimized - sewage treatment effect value curve, and analyze it to obtain the change type of the parameter value to be optimized - sewage treatment effect value curve; Step 4: Based on the change type of the parameter value to be optimized - sewage treatment effect value curve obtained above, train the target parameter optimization model.

[0008] As a further solution of the present invention: the process of obtaining the impact-related value is:

[0009] Obtain the total value of abnormal time of influencing parameters, the total value of abnormal time of sewage treatment effect and the total value of coincidence time;

[0010] Substitute into the formula , the impact-related value XG is calculated, where T1 represents the total value of the abnormal time of the impact parameter, T2 represents the total value of the abnormal time of the sewage treatment effect, T3 represents the total value of the overlap time, and a1 is the preset proportional coefficient, and the value of a1 is 2.86.

[0011] As a further solution of the present invention: the acquisition process of the total value of abnormal time of the influencing parameter, the total value of abnormal time of the sewage treatment effect and the total value of the overlap time is:

[0012] Preset the analysis period, extract the analysis period from the historical data, mark it as the historical analysis period, and obtain the influencing parameter values ​​and sewage treatment effect values ​​of the historical analysis period;

[0013] Compare the influencing parameter value of the sewage treatment with the influencing parameter threshold range, and if the influencing parameter value of the sewage treatment is not within the influencing parameter threshold range, generate an influencing parameter abnormal signal;

[0014] In the historical analysis period, the time period in which the influence parameter abnormality signal is generated is obtained, marked as the influence parameter abnormality time period, and the time values ​​of all the influence parameter abnormality time periods are calculated and summed to obtain the total value of the influence parameter abnormality time;

[0015] The sewage treatment effect value is compared with the sewage treatment effect threshold value, and if the sewage treatment effect value is less than the sewage treatment effect threshold value, a sewage treatment effect abnormal signal is generated;

[0016] During the historical analysis period, the time period in which the abnormal sewage treatment effect signal is generated is obtained and marked as the abnormal sewage treatment effect time period. All the abnormal sewage treatment effect time periods are summed up.

[0017] Obtain the total value of abnormal time of sewage treatment effect;

[0018] During the historical analysis period, the overlapping time periods of the abnormal time periods of influencing parameters and the abnormal time periods of sewage treatment effects are extracted, and all the overlapping time periods are summed up to obtain the total value of the overlapping time.

[0019] As a further solution of the present invention: the process of obtaining the similarity evaluation value is:

[0020] The similarity characterization values ​​of all historical calculation periods are summed up and averaged to obtain the similarity characterization mean, and the similarity characterization values ​​of all historical calculation periods are substituted into the variance formula to calculate the variance value of the similarity characterization value;

[0021] Substitute into the formula , the similarity evaluation value PG is calculated, where XJ represents the mean of the similarity representation, FC represents the variance of the similarity representation value, p1 and p2 are preset proportional coefficients, the value of p1 is 1.73, and the value of p2 is 1.27.

[0022] As a further solution of the present invention: the process of obtaining the similarity characterization value is:

[0023] Obtain the abnormal time deviation ratio and the abnormal number deviation ratio;

[0024] Substitute into the formula , the similarity characterization value XS is calculated, where YS represents the abnormal time deviation ratio, YC represents the abnormal number deviation ratio, s1 and s2 are preset proportional coefficients, the value of s1 is 2.63, and the value of s2 is 2.37.

[0025] As a further solution of the present invention: the process of obtaining the abnormal time deviation ratio and the abnormal number deviation ratio is as follows:

[0026] Extract several calculation periods before the historical analysis period in the historical data and mark them as historical calculation periods, where the time value of the historical calculation period is the same as that of the historical analysis period;

[0027] Obtain the abnormal time period and number of abnormalities of the influencing parameters corresponding to the influencing key values ​​of the historical calculation period;

[0028] In the historical calculation period, all abnormal time periods affecting the parameters are summed and averaged to obtain the average abnormal time of the historical calculation period;

[0029] Obtain the number of abnormalities of the influencing parameters in the time period to be analyzed, sum and average the time periods of the abnormal signals of the influencing parameters in the time period to be analyzed, and obtain the abnormal time average of the historical analysis period;

[0030] The mean abnormal time of the historical calculation period is calculated by the difference with the mean abnormal time of the historical analysis period, and the absolute value of the difference is taken to obtain the abnormal time deviation value. The abnormal time deviation value is processed by the ratio of the abnormal time mean of the historical analysis period to obtain the abnormal time deviation ratio;

[0031] The difference between the number of abnormalities in the historical calculation period and the number of abnormalities in the historical analysis period is calculated, and the absolute value of the difference is taken to obtain the abnormality deviation value. The ratio of the abnormality deviation value to the abnormality number in the historical analysis period is calculated to obtain the abnormality deviation ratio.

[0032] As a further solution of the present invention: the process of obtaining the change type of the parameter value to be optimized-sewage treatment effect value curve is:

[0033] Obtaining a change characterization value, and comparing the change characterization value with a change characterization threshold;

[0034] If the change characterization value is greater than or equal to the change characterization threshold, the change type of the parameter value to be optimized-sewage treatment effect value curve is linear change;

[0035] If the change characterization value is less than the change characterization threshold, the change type of the parameter value to be optimized - sewage treatment effect value curve is nonlinear change.

[0036] As a further solution of the present invention: the process of obtaining the change characterization value is:

[0037] Extracting N groups of data in a historical analysis period and multiple historical calculation periods, wherein each group of data includes a parameter value to be optimized and a sewage treatment effect value;

[0038] A two-dimensional model is established with the parameter value to be optimized as the X-axis and the sewage treatment effect value as the Y-axis. The coordinate points of each set of data are plotted in the two-dimensional model, and then a curve of the parameter value to be optimized-sewage treatment effect value is drawn. The curve of the parameter value to be optimized-sewage treatment effect value is divided into several curve segments, and the slopes of the curve segments are measured;

[0039] Connect the beginning and end of the curve with a straight line to obtain the analysis segment, and calculate the slope of the analysis segment;

[0040] If the slope of the curve segment is of the same sign as the slope of the analysis line segment, the curve segment is marked as a curve segment of the same direction;

[0041] If the slope of the curve segment is not the same as the slope of the line connecting the curve endpoints, the curve segment is marked as a reverse curve segment;

[0042] Obtain the number of same-direction curve segments, and perform ratio processing on the number of curve segments to obtain the same-direction ratio of the curve segments;

[0043] The slope of the same-direction curve segment is subjected to difference processing with the slope of the analysis line segment, and the absolute value of the difference is taken to obtain the slope deviation value of the same-direction curve segment, and the slope deviation values ​​of all the same-direction curve segments are summed and averaged to obtain the slope deviation mean;

[0044] Compare the slope deviation values ​​of the same-direction curve segments with the mean slope deviation value;

[0045] If the slope deviation value of the same-direction curve segment is less than or equal to the slope deviation mean value, the same-direction curve segment is marked as a synchronous curve segment;

[0046] The number of synchronous curve segments is obtained, and the ratio is processed with the number of same-direction curve segments to obtain the curve segment synchronization ratio;

[0047] Substitute into the formula , the change characterization value BH is calculated, where TX represents the curve segment isotropic ratio, TB represents the curve segment synchronization ratio, XL represents the slope deviation mean, m1, m2, and m3 are weight coefficients, the value of m1 is 0.23, the value of m2 is 0.42, and the value of m3 is 0.25.

[0048] As a further solution of the present invention: the acquisition process of the target parameter optimization model is:

[0049] If the change type of the parameter value to be optimized-sewage treatment effect value curve is linear, the linear regression model is used for training;

[0050] Use the data in the training set to train the linear regression model, and use the least squares method to obtain the linear regression model;

[0051] If the change type of the parameter value to be optimized - sewage treatment effect value curve is nonlinear, a nonlinear regression model or machine learning algorithm is used for training;

[0052] In a possible embodiment, a random forest model is adopted for training, and the random forest model is trained using the data in the training set to obtain a conforming random forest model.

[0053] The present invention provides a dynamic optimization system based on sewage treatment, which specifically includes:

[0054] Influencing key parameter acquisition module: extract the influencing parameter values ​​and sewage treatment effect values ​​in the historical data, process them, calculate the impact-related value, compare the impact-related value with the impact-related threshold, and if the impact-related value is greater than the impact-related threshold, generate a large impact signal, obtain the impact parameter value corresponding to the large impact signal, and mark it as the influencing key parameter;

[0055] Module for obtaining parameters to be optimized: Based on the influencing key values ​​obtained through screening, the similarity evaluation value is calculated through data processing; the similarity evaluation value is compared with the similarity evaluation threshold. If the similarity evaluation value is less than the similarity evaluation threshold, a high similarity signal is generated, and the influencing key value corresponding to the high similarity signal is obtained and marked as the parameter value to be optimized;

[0056] Curve change type judgment module: based on the obtained parameter values ​​to be optimized, construct the parameter value to be optimized-sewage treatment effect value curve, and analyze it to obtain the change type of the parameter value to be optimized-sewage treatment effect value curve;

[0057] Target parameter optimization model acquisition module: Based on the above-obtained parameter value to be optimized-sewage treatment effect value curve change type, the target parameter optimization model is trained.

[0058] Beneficial effects of the present invention:

[0059] (1) The present invention obtains the influencing parameters and sewage effect values ​​of sewage treatment, obtains the influencing related values ​​through data processing, compares the influencing related values ​​with the influencing related thresholds, and marks the influencing parameter values ​​corresponding to the influencing related values ​​greater than the influencing related thresholds as influencing key parameters, thereby obtaining the parameters that have a greater impact on the sewage effect from among multiple influencing parameters, thereby helping the staff to conduct targeted analysis;

[0060] (2) The present invention calculates the similarity characterization values ​​of the historical analysis period and the historical calculation period, calculates the similarity characterization mean and variance value based on the similarity characterization values ​​of multiple historical calculation periods and the historical analysis period, calculates the similarity evaluation value, compares the similarity evaluation value with the similarity evaluation threshold, and marks the corresponding key influencing parameters when the similarity evaluation value is less than the similarity evaluation threshold as parameters to be optimized, so as to accurately identify which key influencing parameters present abnormal behavior patterns that are highly similar to those in the current analysis period in the historical data, and use these parameters as key values ​​to be optimized, which is helpful to screen out the parameters to be optimized and improve the efficiency of the optimization work;

[0061] (3) The present invention analyzes N groups of data in the historical data to determine whether the change type is linear or nonlinear, and selects a suitable model based on the determined type, thereby improving the accuracy of model selection. The model is then trained to obtain a target parameter optimization model, which is applied to the actual sewage treatment process, thereby achieving dynamic optimization of the sewage treatment effect and improving the efficiency and effect of sewage treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention will be further described below in conjunction with the accompanying drawings.

[0063] Figure 1 It is a flow chart of a dynamic optimization method based on sewage treatment of the present invention;

[0064] Figure 2 It is a flow chart of a dynamic optimization system based on sewage treatment according to the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] Embodiment 1:

[0067] See also Figure 1 As shown, a dynamic optimization method based on sewage treatment according to an embodiment of the present invention comprises the following steps:

[0068] Step 1: Extract the influencing parameter values ​​and sewage treatment effect values ​​from the historical data, process them, and select the key parameters that have a great impact on the sewage treatment effect;

[0069] Exemplarily, the influencing parameter values ​​of sewage treatment include, but are not limited to: treatment process parameter values ​​(such as aeration volume, stirring intensity, and sedimentation time), equipment operation status parameter values ​​(such as equipment operation time and failure rate);

[0070] The sewage treatment effect value includes the effluent water quality value, wherein the effluent water quality value includes but is not limited to chemical oxygen demand and biochemical oxygen demand;

[0071] In some embodiments, an analysis period is preset, an analysis period is extracted from historical data, marked as a historical analysis period, and an influencing parameter value and a sewage treatment effect value of the historical analysis period are obtained;

[0072] The analysis period is set by those skilled in the art based on their work experience;

[0073] Compare the influencing parameter value of sewage treatment with the influencing parameter threshold range, wherein the influencing parameter threshold range is summarized and set by technical personnel in this field based on historical multiple experimental data and work experience;

[0074] If the influencing parameter value of the sewage treatment is within the influencing parameter threshold range, it means that the influencing parameter value is in a normal state, and a normal influencing parameter signal is generated;

[0075] If the influencing parameter value of the sewage treatment is not within the influencing parameter threshold range, it means that the influencing parameter value is in an abnormal state, and an influencing parameter abnormal signal is generated;

[0076] In the historical analysis period, the time period in which the influence parameter abnormality signal is generated is obtained, marked as the influence parameter abnormality time period, and the time values ​​of all the influence parameter abnormality time periods are calculated and summed to obtain the total value of the influence parameter abnormality time;

[0077] Compare the sewage treatment effect value with the sewage treatment effect threshold, wherein the sewage treatment effect threshold is set by technicians in this field based on historical experimental data and work experience;

[0078] If the sewage treatment effect value is greater than or equal to the sewage treatment effect threshold, it means that the sewage treatment effect is good, and a normal sewage treatment effect signal is generated;

[0079] If the sewage treatment effect value is less than the sewage treatment effect threshold, it means that the sewage treatment effect is poor, and a sewage treatment effect abnormal signal is generated;

[0080] During the historical analysis period, the time period in which the abnormal sewage treatment effect signal is generated is obtained, marked as the abnormal sewage treatment effect time period, and all the abnormal sewage treatment effect time periods are summed to obtain the total value of the abnormal sewage treatment effect time;

[0081] During the historical analysis period, the overlapping time periods of the abnormal time periods of the influencing parameters and the abnormal time periods of the sewage treatment effects are extracted, and all the overlapping time periods are summed up to obtain the total value of the overlapping time;

[0082] Substitute into the formula , calculate the impact related value XG, where T1 represents the total value of the impact parameter abnormal time, T2 represents the total value of the sewage treatment effect abnormal time, T3 represents the total value of the overlap time, a1 is the preset proportional coefficient, and the value of a1 is 2.86;

[0083] It should be explained that the meaning reflected by the impact correlation value is: the impact correlation value is calculated by the proportion of the total value of overlap time in the total value of the impact parameter abnormal time and the proportion of the total value of overlap time in the total value of the sewage treatment effect abnormal time. Among them, the greater the proportion of the total value of overlap time in the total value of the impact parameter abnormal time and the proportion of the total value of overlap time in the total value of the sewage treatment effect abnormal time, the greater the correlation between the impact parameter abnormality and the sewage treatment effect abnormality.

[0084] Compare the impact-related value with the impact-related threshold, wherein the impact-related threshold is set by a person skilled in the art based on a summary of multiple historical experimental data;

[0085] If the impact correlation value is less than or equal to the impact correlation threshold, it means that the impact parameter has little impact on the sewage treatment effect and generates a small impact signal;

[0086] If the impact correlation value is greater than the impact correlation threshold, it means that the impact parameter has a greater impact on the sewage treatment effect and generates a large impact signal;

[0087] Based on the generated large-influence signal, an influencing parameter value corresponding to the generated large-influence signal is obtained and marked as an influencing key parameter;

[0088] The technical solution of the embodiment of the present invention is mainly as follows: first, by obtaining the influencing parameters and sewage effect values ​​of sewage treatment, data processing is performed to obtain the impact-related value, and the impact-related value is compared with the impact-related threshold. If the impact-related value is greater than the impact-related threshold, a large impact signal is generated, and the influencing parameter value corresponding to the generated large impact signal is marked as an influencing key parameter, so that the parameters that have a greater impact on the sewage effect can be obtained from multiple influencing parameters, which can help staff to conduct targeted analysis.

[0089] Embodiment 2:

[0090] Based on Example 1, please refer to Figure 1 As shown, a dynamic optimization method based on sewage treatment according to an embodiment of the present invention further includes the following steps:

[0091] Step 2: Based on the influencing key values ​​obtained through screening, obtain the parameter values ​​to be optimized in the influencing key values ​​through data processing;

[0092] In some embodiments, a number of calculation periods before the historical analysis period in the historical data are extracted and marked as historical calculation periods, wherein the time value of the historical calculation period is the same as that of the historical analysis period;

[0093] Obtain the abnormal time period and number of abnormalities of the influencing parameters corresponding to the influencing key values ​​of the historical calculation period;

[0094] In the historical calculation period, all abnormal time periods affecting the parameters are summed and averaged to obtain the average abnormal time of the historical calculation period;

[0095] Obtain the number of abnormalities of the influencing parameters in the time period to be analyzed, sum and average the time periods of the abnormal signals of the influencing parameters in the time period to be analyzed, and obtain the abnormal time average of the historical analysis period;

[0096] The mean abnormal time of the historical calculation period is calculated by the difference with the mean abnormal time of the historical analysis period, and the absolute value of the difference is taken to obtain the abnormal time deviation value. The abnormal time deviation value is processed by the ratio of the abnormal time mean of the historical analysis period to obtain the abnormal time deviation ratio;

[0097] The number of abnormalities in the historical calculation period is calculated by difference with the number of abnormalities in the historical analysis period, the absolute value of the difference is taken to obtain the deviation value of the number of abnormalities, and the ratio of the deviation value of the number of abnormalities to the number of abnormalities in the historical analysis period is calculated to obtain the deviation ratio of the number of abnormalities;

[0098] Substitute into the formula , the similarity characterization value XS is calculated, where YS represents the abnormal time deviation ratio, YC represents the abnormal number deviation ratio, s1 and s2 are preset proportional coefficients, the value of s1 is 2.63, and the value of s2 is 2.37;

[0099] It should be explained that the meaning of the similarity characterization value is as follows: the similarity characterization value is calculated by the abnormal time deviation ratio and the abnormal number deviation ratio, wherein the smaller the abnormal time deviation ratio is, the smaller the deviation between the abnormal time mean corresponding to the historical calculation period and the abnormal time mean corresponding to the historical analysis period is, and the higher the similarity is; the smaller the abnormal number deviation ratio is, the smaller the deviation between the abnormal number corresponding to the historical calculation period and the abnormal number corresponding to the historical analysis period is, and the higher the similarity is;

[0100] The similarity characterization values ​​of all historical calculation periods are summed up and averaged to obtain the similarity characterization mean, and the similarity characterization values ​​of all historical calculation periods are substituted into the variance formula to calculate the variance value of the similarity characterization value;

[0101] Substitute into the formula , the similarity evaluation value PG is calculated, where XJ represents the mean of the similarity representation, FC represents the variance of the similarity representation value, p1 and p2 are preset proportional coefficients, the value of p1 is 1.73, and the value of p2 is 1.27;

[0102] It should be explained that the meaning of the similarity evaluation value is: the similarity evaluation value is calculated by the mean value of the similarity representation and the variance value of the similarity representation value, wherein the smaller the mean value of the similarity representation is, the higher the similarity between the multiple historical calculation periods and the historical analysis periods is, and the smaller the variance value of the similarity representation value is, the smaller the fluctuation degree of the similarity representation value corresponding to the multiple historical calculation periods is, and the more stable it is;

[0103] Compare the similarity evaluation value with a similarity evaluation threshold, wherein the similarity evaluation threshold is set by a person skilled in the art based on a summary of multiple historical experimental data;

[0104] If the similarity evaluation value is less than the similarity evaluation threshold, a high similarity signal is generated;

[0105] If the similarity evaluation value is greater than or equal to the similarity evaluation threshold, a similarity low signal is generated;

[0106] Based on the generated high similarity signal, the influencing key value corresponding to the high similarity signal is obtained and marked as the parameter value to be optimized;

[0107] The technical solution of the embodiment of the present invention is mainly: by obtaining data of key influencing values ​​in historical analysis periods and historical calculation periods, calculating similar characterization values ​​between historical analysis periods and historical calculation periods, and based on similar characterization values ​​of multiple historical calculation periods and historical analysis periods, calculating similar characterization means and variance values, calculating similar evaluation values, comparing the similar evaluation values ​​with similar evaluation thresholds, and generating a high similarity signal if the similar evaluation value is less than the similar evaluation threshold, so as to accurately identify which key influencing parameters present abnormal behavior patterns that are highly similar to the current analysis period in historical data, and use these parameters as key values ​​to be optimized, which helps to screen out parameters to be optimized and improve the efficiency of optimization work.

[0108] Embodiment 3:

[0109] Based on Example 1 and Example 2, please refer to Figure 1 As shown, a dynamic optimization method based on sewage treatment according to an embodiment of the present invention further includes the following steps:

[0110] Step 3: Based on the obtained parameter values ​​to be optimized, construct a parameter value to be optimized-sewage treatment effect value curve, and analyze it to obtain the change type of the parameter value to be optimized-sewage treatment effect value curve;

[0111] Among them, the curve change types include linear change and nonlinear change;

[0112] In some embodiments, N groups of data are extracted in a historical analysis period and multiple historical calculation periods, wherein each group of data includes a parameter value to be optimized and a sewage treatment effect value;

[0113] A two-dimensional model is established with the parameter value to be optimized as the X-axis and the sewage treatment effect value as the Y-axis. The coordinate points of each set of data are plotted in the two-dimensional model, and then a curve of the parameter value to be optimized-sewage treatment effect value is drawn. The curve of the parameter value to be optimized-sewage treatment effect value is divided into several curve segments, and the slopes of the curve segments are measured;

[0114] Connect the beginning and end of the curve with a straight line to obtain the analysis segment, and calculate the slope of the analysis segment;

[0115] If the slope of the curve segment is of the same sign as the slope of the analysis line segment, the curve segment is marked as a curve segment of the same direction;

[0116] If the slope of the curve segment is not the same as the slope of the line connecting the curve endpoints, the curve segment is marked as a reverse curve segment;

[0117] It should be noted that the same positive and negative means that the slope of the curve segment and the slope of the analysis line segment are of the same positive or negative, and different positive and negative means that the slope of the curve segment and the slope of the analysis line segment are of different positive or different negative, or one is 0 and the other is not 0;

[0118] Obtain the number of same-direction curve segments, and perform ratio processing on the number of curve segments to obtain the same-direction ratio of the curve segments;

[0119] The slope of the same-direction curve segment is subjected to difference processing with the slope of the analysis line segment, and the absolute value of the difference is taken to obtain the slope deviation value of the same-direction curve segment, and the slope deviation values ​​of all the same-direction curve segments are summed and averaged to obtain the slope deviation mean;

[0120] Compare the slope deviation values ​​of the same-direction curve segments with the mean slope deviation value;

[0121] If the slope deviation value of the same-direction curve segment is greater than the slope deviation mean value, the same-direction curve segment is marked as an asynchronous curve segment;

[0122] If the slope deviation value of the same-direction curve segment is less than or equal to the slope deviation mean value, the same-direction curve segment is marked as a synchronous curve segment;

[0123] The number of synchronous curve segments is obtained, and the ratio is processed with the number of same-direction curve segments to obtain the curve segment synchronization ratio;

[0124] Substitute into the formula , the change characterization value BH is calculated, where TX represents the curve segment isotropic ratio, TB represents the curve segment synchronization ratio, XL represents the slope deviation mean, m1, m2, and m3 are weight coefficients, m1 is 0.23, m2 is 0.42, and m3 is 0.25;

[0125] It should be explained that the meaning reflected by the change characterization value is: the change characterization value is calculated by the curve segment same direction ratio, the curve segment synchronization ratio and the slope deviation mean, wherein the larger the curve segment same direction ratio is, the more curve segments are in the same direction as the analysis line segment, and the higher the degree of linear change is; the smaller the mean value of slope deviation is, the smaller the mean value of absolute deviation between the slope of the curve segment and the slope of the analysis line segment is, and the higher the degree of linear change is; the larger the curve segment synchronization ratio is, the more curve segments are in the same direction with smaller absolute deviations of the slope of the analysis line segment, and the higher the degree of linear change is;

[0126] Comparing the change characterization value with the change characterization threshold, wherein the change characterization threshold is set by a person skilled in the art based on a summary of multiple historical experimental data;

[0127] If the change characterization value is greater than or equal to the change characterization threshold, the change type of the parameter value to be optimized-sewage treatment effect value curve is linear change;

[0128] If the change characterization value is less than the change characterization threshold, the change type of the parameter value to be optimized-sewage treatment effect value curve is nonlinear change;

[0129] Step 4: Based on the obtained parameter value to be optimized - sewage treatment effect value curve change type, train the target parameter optimization model;

[0130] In some embodiments, 0.7N groups of data are extracted from N groups of data as a training set, and the remaining 0.3N groups of data are used as a validation set;

[0131] If the change type of the parameter value to be optimized-sewage treatment effect value curve is linear, the linear regression model is used for training, and the linear regression model can fit the linear relationship well;

[0132] Use the data in the training set to train the linear regression model, use the least squares method to get the linear regression model, and use the validation set data to verify the trained linear regression model;

[0133] For the validation set data, the difference between the predicted value and the actual value is calculated, the absolute value of the difference is taken to obtain the prediction deviation value, and the prediction deviation value is compared with the prediction deviation threshold, wherein the prediction deviation threshold is set by a person skilled in the art based on a summary of multiple historical experimental data;

[0134] If the prediction deviation value is less than or equal to the prediction deviation threshold, the group of data is marked as a valid array; if the prediction deviation value is greater than the prediction deviation threshold, the group of data is marked as an invalid array;

[0135] Get the number of valid array values ​​and the number of invalid data array values, calculate the ratio of the valid array values ​​to the total number of validation set array values ​​to obtain the valid ratio value, and compare the number of invalid array values ​​to the total number of validation set array values ​​to obtain the invalid ratio value;

[0136] If the effective proportion value is greater than the invalid proportion value, the obtained linear regression model is marked as the target parameter optimization model;

[0137] If the effective proportion value is less than or equal to the invalid proportion value, the obtained linear regression model is corrected, including but not limited to adjusting the model parameters, obtaining the corrected linear regression model, and marking it as the target parameter optimization model;

[0138] If the change type of the parameter value to be optimized - sewage treatment effect value curve is nonlinear, a nonlinear regression model or machine learning algorithm is used for training;

[0139] In a possible embodiment, a random forest model is used for training, and the random forest model is trained using the data in the training set to obtain a suitable random forest model;

[0140] The random forest model trained with the validation set is used for validation. The validation method is the same as that for the linear regression model, which will not be described in detail here, to obtain the target parameter optimization model.

[0141] By applying the target parameter optimization model to the actual sewage treatment process, the sewage treatment effect value is predicted according to the parameter values ​​to be optimized obtained in real time, and the parameter values ​​to be optimized are adjusted according to the prediction results to dynamically optimize the sewage treatment effect;

[0142] The technical solution of the embodiment of the present invention is mainly as follows: first, by analyzing N groups of data in historical data, it is determined whether the change type is linear change or nonlinear change, and a suitable model is selected based on the determined type result, thereby improving the accuracy of model selection, and then model training is performed to obtain a target parameter optimization model, and the target parameter optimization model is applied to the actual sewage treatment process, thereby realizing dynamic optimization of the sewage treatment effect and improving the efficiency and effect of sewage treatment.

[0143] Embodiment 4:

[0144] Based on Example 1, Example 2 and Example 3, please refer to Figure 2 As shown, a dynamic optimization system based on sewage treatment according to an embodiment of the present invention specifically includes:

[0145] Influencing key parameter acquisition module: extract the influencing parameter values ​​and sewage treatment effect values ​​in the historical data, process them, and screen the influencing key parameters that have a great impact on the sewage treatment effect;

[0146] Parameter acquisition module for optimization: based on the influencing key values ​​obtained through screening, the parameter values ​​to be optimized in the influencing key values ​​are obtained through data processing;

[0147] Curve change type judgment module: based on the obtained parameter values ​​to be optimized, construct the parameter value to be optimized-sewage treatment effect value curve, and analyze it to obtain the change type of the parameter value to be optimized-sewage treatment effect value curve;

[0148] Among them, the curve change types include linear change and nonlinear change;

[0149] Target parameter optimization model acquisition module: Based on the above-obtained parameter value to be optimized-sewage treatment effect value curve change type, the target parameter optimization model is trained.

[0150] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A dynamic optimization method based on sewage treatment, characterized in that: The specific steps include: Step 1: extract the influencing parameter values ​​and the sewage treatment effect values ​​in the historical data, process them, calculate the impact-related value, compare the impact-related value with the impact-related threshold, and if the impact-related value is greater than the impact-related threshold, generate a large impact signal, obtain the impact parameter value corresponding to the large impact signal, and mark it as the key impact parameter; Step 2: Based on the influencing key values ​​obtained through screening, calculate the similarity evaluation value through data processing; compare the similarity evaluation value with the similarity evaluation threshold, if the similarity evaluation value is less than the similarity evaluation threshold, generate a high similarity signal, obtain the influencing key value corresponding to the high similarity signal, and mark it as the parameter value to be optimized; Step 3: Based on the obtained parameter values ​​to be optimized, construct a parameter value to be optimized-sewage treatment effect value curve, and analyze it to obtain the change type of the parameter value to be optimized-sewage treatment effect value curve; Step 4: Based on the obtained parameter value to be optimized - sewage treatment effect value curve change type, train the target parameter optimization model; The process of obtaining the similarity evaluation value is as follows: The similarity characterization values ​​of all historical calculation periods are summed up and averaged to obtain the similarity characterization mean, and the similarity characterization values ​​of all historical calculation periods are substituted into the variance formula to calculate the variance value of the similarity characterization value; Substitute into the formula , the similarity evaluation value PG is calculated, where XJ represents the mean of the similarity representation, FC represents the variance of the similarity representation value, p1 and p2 are preset proportional coefficients, the value of p1 is 1.73, and the value of p2 is 1.27; The process of obtaining the similarity characterization value is as follows: Obtain the abnormal time deviation ratio and the abnormal number deviation ratio; Substitute into the formula , the similarity characterization value XS is calculated, where YS represents the abnormal time deviation ratio, YC represents the abnormal number deviation ratio, s1 and s2 are preset proportional coefficients, the value of s1 is 2.63, and the value of s2 is 2.37; The process of obtaining the change type of the parameter value to be optimized-sewage treatment effect value curve is as follows: Obtaining a change characterization value, and comparing the change characterization value with a change characterization threshold; If the change characterization value is greater than or equal to the change characterization threshold, the change type of the parameter value to be optimized-sewage treatment effect value curve is linear change; If the change characterization value is less than the change characterization threshold, the change type of the parameter value to be optimized-sewage treatment effect value curve is nonlinear change; The process of obtaining the change characterization value is as follows: Extracting N groups of data in a historical analysis period and multiple historical calculation periods, wherein each group of data includes a parameter value to be optimized and a sewage treatment effect value; A two-dimensional model is established with the parameter value to be optimized as the X-axis and the sewage treatment effect value as the Y-axis. The coordinate points of each set of data are plotted in the two-dimensional model, and then a curve of the parameter value to be optimized-sewage treatment effect value is drawn. The curve of the parameter value to be optimized-sewage treatment effect value is divided into several curve segments, and the slopes of the curve segments are measured; Connect the beginning and end of the curve with a straight line to obtain the analysis segment, and calculate the slope of the analysis segment; If the slope of the curve segment is of the same sign as the slope of the analysis line segment, the curve segment is marked as a curve segment of the same direction; If the slope of the curve segment is not the same as the slope of the line connecting the curve endpoints, the curve segment is marked as a reverse curve segment; Obtain the number of same-direction curve segments, and perform ratio processing on the number of curve segments to obtain the same-direction ratio of the curve segments; The slope of the same-direction curve segment is subjected to difference processing with the slope of the analysis line segment, and the absolute value of the difference is taken to obtain the slope deviation value of the same-direction curve segment, and the slope deviation values ​​of all the same-direction curve segments are summed and averaged to obtain the slope deviation mean; Compare the slope deviation values ​​of the same-direction curve segments with the mean slope deviation value; If the slope deviation value of the same-direction curve segment is less than or equal to the slope deviation mean value, the same-direction curve segment is marked as a synchronous curve segment; The number of synchronous curve segments is obtained, and the ratio is processed with the number of same-direction curve segments to obtain the curve segment synchronization ratio; Substitute into the formula , the change characterization value BH is calculated, where TX represents the curve segment isotropic ratio, TB represents the curve segment synchronization ratio, XL represents the slope deviation mean, m1, m2, and m3 are weight coefficients, the value of m1 is 0.23, the value of m2 is 0.42, and the value of m3 is 0.

25.

2. A dynamic optimization method based on sewage treatment according to claim 1, characterized in that: The process of obtaining the impact-related value is as follows: Obtain the total value of abnormal time of influencing parameters, the total value of abnormal time of sewage treatment effect and the total value of coincidence time; Substitute into the formula , the impact-related value XG is calculated, where T1 represents the total value of the abnormal time of the impact parameter, T2 represents the total value of the abnormal time of the sewage treatment effect, T3 represents the total value of the overlap time, and a1 is the preset proportional coefficient, and the value of a1 is 2.

86.

3. A dynamic optimization method based on sewage treatment according to claim 2, characterized in that: The acquisition process of the total value of the abnormal time of the influencing parameters, the total value of the abnormal time of the sewage treatment effect and the total value of the coincidence time is as follows: Preset the analysis period, extract the analysis period from the historical data, mark it as the historical analysis period, and obtain the influencing parameter values ​​and sewage treatment effect values ​​of the historical analysis period; Compare the influencing parameter value of the sewage treatment with the influencing parameter threshold range, and if the influencing parameter value of the sewage treatment is not within the influencing parameter threshold range, generate an influencing parameter abnormal signal; In the historical analysis period, the time period in which the influence parameter abnormality signal is generated is obtained, marked as the influence parameter abnormality time period, and the time values ​​of all the influence parameter abnormality time periods are calculated and summed to obtain the total value of the influence parameter abnormality time; The sewage treatment effect value is compared with the sewage treatment effect threshold value, and if the sewage treatment effect value is less than the sewage treatment effect threshold value, a sewage treatment effect abnormal signal is generated; During the historical analysis period, the time period in which the abnormal sewage treatment effect signal is generated is obtained, marked as the abnormal sewage treatment effect time period, and all the abnormal sewage treatment effect time periods are summed to obtain the total value of the abnormal sewage treatment effect time; During the historical analysis period, the overlapping time periods of the abnormal time periods of influencing parameters and the abnormal time periods of sewage treatment effects are extracted, and all the overlapping time periods are summed up to obtain the total value of the overlapping time.

4. A dynamic optimization method based on sewage treatment according to claim 1, characterized in that: The process of obtaining the abnormal time deviation ratio and the abnormal number of deviation ratio is as follows: Extract several calculation periods before the historical analysis period in the historical data and mark them as historical calculation periods, where the time value of the historical calculation period is the same as that of the historical analysis period; Obtain the abnormal time period and number of abnormalities of the influencing parameters corresponding to the influencing key values ​​of the historical calculation period; In the historical calculation period, all abnormal time periods affecting the parameters are summed and averaged to obtain the average abnormal time of the historical calculation period; Obtain the number of abnormalities of the influencing parameters in the time period to be analyzed, sum and average the time periods of the abnormal signals of the influencing parameters in the time period to be analyzed, and obtain the abnormal time average of the historical analysis period; The difference between the abnormal time mean of the historical calculation period and the abnormal time mean of the historical analysis period is calculated, the absolute value of the difference is taken to obtain the abnormal time deviation value, and the abnormal time deviation value is compared with the abnormal time mean of the historical analysis period to obtain the abnormal time deviation ratio; The difference between the number of abnormalities in the historical calculation period and the number of abnormalities in the historical analysis period is calculated, and the absolute value of the difference is taken to obtain the abnormality deviation value. The ratio of the abnormality deviation value to the abnormality number in the historical analysis period is calculated to obtain the abnormality deviation ratio.

5. A dynamic optimization method based on sewage treatment according to claim 1, characterized in that: The acquisition process of the target parameter optimization model is as follows: If the change type of the parameter value to be optimized-sewage treatment effect value curve is linear, the linear regression model is used for training; Use the data in the training set to train the linear regression model, and use the least squares method to obtain the linear regression model; If the change type of the parameter value to be optimized - sewage treatment effect value curve is nonlinear, a nonlinear regression model or machine learning algorithm is used for training; The random forest model is used for training, and the data in the training set is used to train the random forest model to obtain a suitable random forest model.

6. A dynamic optimization system based on sewage treatment, characterized in that: The system is used to execute the method described in any one of claims 1 to 5, and the system comprises: Influencing key parameter acquisition module: extract the influencing parameter values ​​and sewage treatment effect values ​​in the historical data, process them, calculate the impact-related value, compare the impact-related value with the impact-related threshold, and if the impact-related value is greater than the impact-related threshold, generate a large impact signal, obtain the impact parameter value corresponding to the large impact signal, and mark it as the influencing key parameter; Module for obtaining parameters to be optimized: Based on the influencing key values ​​obtained through screening, the similarity evaluation value is calculated through data processing; the similarity evaluation value is compared with the similarity evaluation threshold. If the similarity evaluation value is less than the similarity evaluation threshold, a high similarity signal is generated, and the influencing key value corresponding to the high similarity signal is obtained and marked as the parameter value to be optimized; Curve change type judgment module: based on the obtained parameter values ​​to be optimized, construct the parameter value to be optimized-sewage treatment effect value curve, and analyze it to obtain the change type of the parameter value to be optimized-sewage treatment effect value curve; Target parameter optimization model acquisition module: Based on the above-obtained parameter value to be optimized-sewage treatment effect value curve change type, the target parameter optimization model is trained.

Images