Method and system for automatically recommending catalyst dosage based on wastewater detection data

Abstract

The present application proposes a method and system for automatically recommending catalyst dosage based on wastewater detection data, which belongs to the field of wastewater treatment. Among them, the method includes: collecting a sample wastewater treatment data set; clustering the sample wastewater treatment data set to obtain multiple clustering results, and calculating multiple clustering cost coefficients; calculating the treatment stability coefficient according to multiple wastewater indicators based on the clustered sample processing data in each clustering result; according to the clustering cost coefficient and the treatment stability coefficient, the clustering results are trained with weights assigned, and an integrated catalyst dosage recommender is trained, the current wastewater data is input into the recommender, and the recommended catalyst dosage is output. The present application solves the technical problem in the prior art of recommending catalyst dosage based on historical wastewater treatment data, which leads to inaccurate recommendations due to data randomness and errors, and achieves the technical effect of improving the accuracy of catalyst dosage recommendations and the stability of wastewater treatment effects.

Description

Technical Field

[0001] The present invention relates to the field of wastewater treatment, and in particular to a method and system for automatically recommending catalyst dosage based on wastewater detection data. Background Art

[0002] With the acceleration of industrialization and increasingly stringent environmental protection requirements, wastewater treatment technology has become an important research direction in the field of environmental protection. Catalysts are key components in wastewater treatment, and accurate control of their dosage directly affects the treatment effect.

[0003] At present, the determination of catalyst dosage in wastewater treatment mainly depends on historical wastewater treatment data. By adopting recommendation methods based on empirical formulas or statistical analysis, by analyzing indicator parameters such as pollutant concentration in historical wastewater detection data, combined with past catalyst usage and treatment effect data, a corresponding recommendation model is established to guide actual production.

[0004] However, historical wastewater treatment data often exhibits randomness and errors. Different batches of wastewater vary in composition and concentration, and measurement errors are inevitable during wastewater testing. These factors lead to inaccurate recommendations in existing catalyst dosage methods, impacting the effectiveness and stability of wastewater treatment. Summary of the Invention

[0005] The present invention aims to solve the technical problem in the prior art of inaccurate recommendations due to randomness and errors in data when recommending catalyst dosage based on historical wastewater treatment data, and provides a method and system for automatically recommending catalyst dosage based on wastewater detection data.

[0006] The technical solution of the present invention to solve the above technical problems is as follows:

[0007] In a first aspect, the present invention provides an automatic catalyst dosage recommendation method based on wastewater detection data, comprising: collecting a sample wastewater data set, a sample catalyst dosage set and a sample treatment data set based on wastewater treatment detection data within a historical period, and combining them to obtain a sample wastewater treatment data set, wherein the wastewater data and the treatment data include indicator parameters of multiple wastewater indicators; clustering the sample wastewater treatment data set to obtain multiple clustering results, and calculating multiple clustering cost coefficients; calculating a treatment stability coefficient according to multiple wastewater indicators based on the clustered sample treatment data in each clustering result to obtain multiple treatment stability coefficients; performing training weight allocation on the multiple clustering results according to the multiple clustering cost coefficients and the multiple treatment stability coefficients, training an integrated catalyst dosage recommender, inputting the current wastewater data into the integrated catalyst dosage recommender, and outputting a recommended catalyst dosage.

[0008] In a second aspect, the present invention provides an automatic catalyst dosage recommendation system based on wastewater detection data, comprising: a data acquisition module for collecting a sample wastewater data set, a sample catalyst dosage set and a sample treatment data set based on wastewater treatment detection data within a historical period, and combining them to obtain a sample wastewater treatment data set, wherein the wastewater data and the treatment data include indicator parameters of multiple wastewater indicators; a clustering analysis module for clustering the sample wastewater treatment data set to obtain multiple clustering results, and calculating multiple clustering cost coefficients; a stability calculation module for calculating the treatment stability coefficient according to multiple wastewater indicators based on the clustered sample treatment data in each clustering result to obtain multiple treatment stability coefficients; a catalyst recommendation module for performing training weight allocation on the multiple clustering results according to the multiple clustering cost coefficients and the multiple treatment stability coefficients, performing integrated catalyst dosage recommender training, inputting the current wastewater data into the integrated catalyst dosage recommender, and outputting the recommended catalyst dosage.

[0009] The beneficial effects of the present invention are:

[0010] Based on historical wastewater treatment inspection data, a sample wastewater dataset, a sample catalyst dosage dataset, and a sample treatment dataset are collected and combined to obtain a sample wastewater treatment dataset. The wastewater data and treatment data include indicator parameters for multiple wastewater indicators. Through systematic data collection, a complete data foundation is provided for subsequent analysis. The sample wastewater treatment dataset is clustered to obtain multiple clustering results, and multiple clustering cost coefficients are calculated. Similar wastewater treatment samples are classified through cluster analysis, and the degree of data fluctuation within each cluster is quantified by the cost coefficients. Based on the clustered sample treatment data within each clustering result, a treatment stability coefficient is calculated according to multiple wastewater indicators to obtain multiple treatment stability coefficients, which can be used to evaluate the stability of the wastewater treatment effect within each cluster. According to the multiple clustering cost coefficients and multiple treatment stability coefficients, training weights are assigned to the multiple clustering results to train an integrated catalyst dosage recommender. The current wastewater data is input into the integrated catalyst dosage recommender, and the recommended catalyst dosage is output. The weight assignment mechanism ensures that high-quality data plays a leading role in model training, thereby improving the accuracy of catalyst dosage recommendations.

[0011] Through the above technical solution, the adverse effects of randomness and errors in historical wastewater treatment data on the accuracy of catalyst dosage recommendations were effectively solved, and intelligent catalyst dosage recommendations based on data quality assessment were realized, achieving the technical effect of improving the accuracy of catalyst dosage recommendations and the stability of wastewater treatment effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1A schematic flow chart of the method for automatically recommending catalyst dosage based on wastewater testing data provided by the present invention;

[0013] Figure 2 This is a schematic diagram of the structure of the automatic catalyst dosage recommendation system based on wastewater detection data provided by the present invention.

[0014] In the accompanying drawings, the components represented by the reference numerals are as follows:

[0015] Data collection module 11, cluster analysis module 12, stability calculation module 13, catalyst recommendation module 14. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0017] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0018] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or illustration". Any embodiment of the present invention described as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed herein.

[0019] Example 1, as Figure 1 As shown, an embodiment of the present invention provides a method for automatically recommending catalyst dosage based on wastewater detection data, comprising:

[0020] S1. Based on wastewater treatment detection data in historical time, a sample wastewater data set, a sample catalyst usage set, and a sample treatment data set are collected and combined to obtain a sample wastewater treatment data set, wherein the wastewater data and the treatment data include indicator parameters of multiple wastewater indicators.

[0021] Specifically, first, all wastewater treatment test data recorded by the wastewater treatment plant over a historical period (e.g., the past year or several months) is collected through the wastewater treatment plant's data management system. Then, based on the acquired wastewater treatment test data, wastewater data, catalyst usage, and sample processing data are collected for each wastewater treatment. This yields a sample wastewater dataset, a sample catalyst usage dataset, and a sample processing dataset, which are then combined to form a sample wastewater processing dataset. The wastewater and processing data include indicator parameters for multiple wastewater indicators, primarily wastewater quality parameters such as pollutant concentrations, which comprehensively reflect the water quality before and after wastewater treatment.

[0022] Among them, the sample wastewater data set refers to the parameter records of various wastewater indicators before each wastewater treatment extracted from the wastewater treatment test data. For example, the chemical oxygen demand (COD) value of a sample wastewater data is 250mg / L, the biochemical oxygen demand (BOD) value is 120mg / L, the pH value is 6.8, the total suspended solids (TSS) is 180mg / L, the ammonia nitrogen content is 35mg / L, etc. The sample catalyst dosage set refers to the record of the catalyst dosage used in each wastewater treatment process extracted from the wastewater treatment test data. For example, the sample catalyst dosage used when processing the sample wastewater data in the above example is recorded as iron oxide catalyst. , activated carbon catalyst is The sample processing data set refers to the records of various index parameters after each wastewater treatment, which are extracted from the wastewater treatment test data. These index parameters after treatment correspond to the index parameters before treatment. For example, in the above example, the sample wastewater data corresponding to the sample processing data has a COD of 30 mg / L, a BOD of 15 mg / L, a TSS of 20 mg / L, a pH of 7.2, and an ammonia nitrogen content of 5 mg / L.

[0023] Through data collection and combination, a complete sample wastewater treatment dataset was established. This dataset contains the correspondence between the initial state of the wastewater, catalyst dosage and treatment effect, laying the foundation for subsequent data analysis and integrated catalyst dosage recommender training.

[0024] S2. Clustering the sample wastewater treatment data set to obtain multiple clustering results, and calculating multiple clustering cost coefficients.

[0025] Specifically, cluster analysis was first performed on the generated sample wastewater treatment data set, grouping sample wastewater treatment data with similar characteristics into one category, forming multiple cluster results. Each cluster result contains multiple sample wastewater treatment data with similar characteristics. These similarity characteristics are calculated based on the combined characteristics of the wastewater data, catalyst dosage, and treatment data. Through cluster analysis, the originally dispersed sample wastewater treatment data were grouped according to the degree of similarity between the data, resulting in high similarity within the same cluster and large differences between the data in different clusters.

[0026] Then, based on the multiple clustering results obtained, the clustering cost coefficient of each cluster is calculated. Among them, the clustering cost coefficient refers to the value calculated based on the deviation amplitude of the sample wastewater data and the sample catalyst dosage in the multiple sample wastewater treatment data in the clustering results. Specifically, the maximum deviation amplitude of the sample wastewater data and the sample catalyst dosage in each cluster is calculated respectively, and then the clustering cost coefficient is calculated based on these deviation amplitudes. The clustering cost coefficient reflects the degree of fluctuation of the sample wastewater treatment data in the same cluster. The greater the difference, the greater the clustering cost coefficient, indicating that the data fluctuation is greater and the data value is lower. For example, if the wastewater characteristics within a cluster are similar, but the corresponding catalyst dosages vary greatly, the clustering cost coefficient of the cluster will be higher, indicating that the data consistency within the cluster is poor, which is not conducive to the subsequent catalyst dosage recommendation.

[0027] Through cluster analysis and calculation of clustering cost coefficients, we can identify data patterns and regularities in the sample wastewater treatment data set, and evaluate the data quality of each clustering result, providing more refined data support for subsequent recommendations on appropriate catalyst dosage based on wastewater characteristics.

[0028] S3. Calculate the treatment stability coefficient according to the cluster sample processing data in each clustering result and the multiple wastewater indicators to obtain multiple treatment stability coefficients.

[0029] Specifically, first, cluster sample processing data is extracted from each clustering result. The cluster sample processing data refers to the sample processing data portion of the multiple sample wastewater processing data contained in each cluster, that is, the various index parameter records of the wastewater after catalyst treatment.

[0030] Then, for each cluster sample processing data, the standard deviation of each parameter is calculated for multiple wastewater indicators (such as COD, BOD, pH, TSS, and ammonia nitrogen), resulting in a discrete parameter for the treatment indicator. This discrete parameter reflects the degree of fluctuation of each parameter after the wastewater in the same cluster is treated with the catalyst. The smaller the discrete parameter, the more stable the treatment effect.

[0031] Next, the preset processing discrete parameters are compared with the calculated processing discrete parameters to calculate the processing stability coefficient. The processing stability coefficient refers to the ratio of the preset processing discrete parameters to the actual processing discrete parameters, which reflects the stability of the catalyst treatment effect. The larger the processing stability coefficient, the more stable the wastewater treatment effect and the higher the data value. According to the above method, the processing stability coefficients of the multiple clustering results obtained are calculated respectively to obtain multiple processing stability coefficients. These processing stability coefficients provide an important basis for the subsequent training weight allocation. Clusters with high stability coefficients will obtain higher weights in training.

[0032] By calculating the treatment stability coefficient, we can evaluate the stability of the catalyst treatment effect within each clustering result, screen out sample data with stable treatment effects and high data value, and provide more reliable data support for the subsequent integrated catalyst dosage recommender training.

[0033] S4. According to the multiple clustering cost coefficients and the multiple processing stability coefficients, the multiple clustering results are trained with weights, and an integrated catalyst dosage recommender is trained. The current wastewater data is input into the integrated catalyst dosage recommender, and the recommended catalyst dosage is obtained by output.

[0034] Specifically, first, based on the obtained multiple clustering cost coefficients, multiple clustering quality coefficients are calculated. The clustering quality coefficient is obtained by subtracting the clustering cost coefficient from 1, reflecting the quality of the clustering result. The smaller the clustering cost coefficient, the larger the corresponding clustering quality coefficient, indicating that the data consistency within the cluster is better and the data value is higher.

[0035] Then, based on the calculated cluster quality coefficients and the obtained processing stability coefficients, multiple training priority coefficients are calculated and assigned accordingly. The training priority coefficient is a comprehensive evaluation metric derived by weighting the cluster quality coefficients and the processing stability coefficients, reflecting the contribution of each clustering result to the catalyst dosage recommender training. Clusters with both high cluster quality coefficients and high processing stability coefficients receive higher training weights and are prioritized during training.

[0036] Next, multiple catalyst dosage recommenders are trained using the clustering results according to the assigned training weights, ultimately yielding an integrated catalyst dosage recommender. This integrated catalyst dosage recommender is a model integration system composed of multiple catalyst dosage recommenders. By making comprehensive decisions based on the outputs of these recommenders, the accuracy and reliability of the recommendation results are improved. Subsequently, the current wastewater data to be treated is obtained and fed into the trained integrated catalyst dosage recommender, which then outputs the optimal recommended catalyst dosage.

[0037] Through step S4, intelligent catalyst dosage recommendation based on historical wastewater treatment data is achieved, overcoming the problem of inaccurate recommendation caused by the randomness and errors of historical wastewater treatment data in traditional methods, and improving the efficiency and effect of wastewater treatment.

[0038] Furthermore, based on the wastewater treatment detection data in the historical period, a sample wastewater data set, a sample catalyst usage set, and a sample treatment data set are collected and combined to obtain sample wastewater treatment data, including:

[0039] S11. Based on the wastewater treatment test data in the historical period, the indicator parameters of multiple wastewater indicators in the wastewater before each wastewater treatment are collected to obtain a sample wastewater data set;

[0040] S12. Based on historical wastewater treatment test data, collect the catalyst dosage during each wastewater treatment and the indicator parameters of multiple wastewater indicators in the treated wastewater to obtain a sample catalyst dosage set and a sample treatment data set;

[0041] S13. Combining the sample wastewater data set, the sample catalyst usage set, and the sample treatment data set to obtain a sample wastewater treatment data set, wherein each piece of sample wastewater treatment data includes sample wastewater data, sample catalyst usage, and sample treatment data.

[0042] In one feasible implementation, raw test data for all wastewater treatment batches over a historical period (e.g., the past year or several months) is first extracted from the wastewater treatment plant's data management system. This data, known as wastewater treatment test data, is then extracted for each batch of wastewater treatment, along with various pre-treatment wastewater parameters, to form a sample wastewater dataset. These wastewater parameters include, but are not limited to, chemical oxygen demand (COD), biochemical oxygen demand (BOD), pH, total suspended solids (TSS), ammonia nitrogen content, heavy metal concentrations, and other pollutant indicators. These indicators comprehensively reflect the initial contamination state and characteristics of the wastewater being treated.

[0043] Then, for each batch of wastewater treatment, we extract data on catalyst usage during the treatment process to form a sample catalyst usage dataset. This data records the amount of each catalyst added, reflecting catalyst usage. Simultaneously, we extract various parameters of the treated wastewater to form a sample treatment dataset. These post-treatment wastewater parameters correspond to pre-treatment wastewater parameters and are used to evaluate treatment effectiveness.

[0044] Subsequently, the obtained sample wastewater dataset, sample catalyst usage dataset, and sample treatment dataset are associated and combined to establish a complete sample wastewater treatment dataset. In this sample wastewater treatment dataset, each sample wastewater treatment data record contains three parts: Sample wastewater data, i.e. various index parameters of wastewater before treatment; Sample catalyst dosage, i.e., the amount of catalyst used during the treatment process; Sample processing data, namely the various indicators and parameters of the treated wastewater. These three parts of data form a complete record of the wastewater treatment process, establishing the corresponding relationship between the initial state of the wastewater, the amount of catalyst used, and the treatment effect.

[0045] Through the above steps, the collection and organization of historical wastewater treatment test data was achieved, providing a structured data foundation for subsequent cluster analysis and catalyst dosage recommender training, thereby accurately recommending catalyst dosage.

[0046] Furthermore, clustering is performed on the sample wastewater treatment data set to obtain multiple clustering results, and multiple clustering cost coefficients are calculated, including:

[0047] S21, clustering the sample wastewater treatment data set to obtain multiple clustering results, wherein each clustering result includes multiple sample wastewater treatment data;

[0048] S22. Calculate the maximum deviation amplitudes of the sample wastewater data and the sample catalyst usage in the plurality of sample wastewater treatment data in each clustering result to obtain a plurality of cluster deviation amplitude sets, wherein the maximum deviation amplitude includes the maximum wastewater deviation amplitude of the sample wastewater data and the maximum usage deviation amplitude of the sample catalyst usage;

[0049] S23. Calculate the mean of the maximum wastewater deviation amplitude and the maximum usage deviation amplitude in each cluster deviation amplitude set to obtain multiple cluster cost coefficients.

[0050] In a preferred embodiment, cluster analysis is first performed on the obtained sample wastewater treatment data set, grouping sample wastewater treatment data with similar characteristics into a single category to form multiple clustering results. Each clustering result contains multiple sample wastewater treatment data with similar characteristics, which are determined based on the combined characteristics of the wastewater data, catalyst dosage, and treatment data. Through cluster analysis, the previously dispersed sample wastewater treatment data are grouped according to the degree of similarity between the data, resulting in high similarity within the same cluster and high variability between the data in different clusters.

[0051] Then, the maximum deviation amplitudes of the sample wastewater data and the sample catalyst dosage in the multiple sample wastewater treatment data within each clustering result are calculated to obtain multiple clustering deviation amplitude sets. The maximum deviation amplitude includes the maximum wastewater deviation amplitude of the sample wastewater data and the maximum dosage deviation amplitude of the sample catalyst dosage. Therefore, each clustering deviation amplitude set includes a maximum wastewater deviation amplitude and a maximum dosage deviation amplitude. Specifically, the various indicator parameters and sample catalyst dosage of the multiple sample wastewater treatment data in each clustering result are first normalized so that the indicator data can be compared on the same scale. Among them, the normalization process can use the maximum and minimum value normalization method to map each indicator data to the [0,1] interval to eliminate the dimensional differences between different indicators. For each cluster's normalized wastewater data, the difference between the maximum and minimum values ​​of each parameter (such as COD, BOD, pH, and TSS) was calculated to determine the deviation for each parameter. The average of all the deviations was then calculated as the maximum wastewater deviation for that cluster. For each cluster's normalized catalyst dosage, the difference between the maximum and minimum values ​​was calculated to determine the deviation for each catalyst dosage. The average of all the deviations was then calculated as the maximum deviation for that cluster. For example, if the normalized iron oxide catalyst dosage within a cluster has a minimum of 0.5 and a maximum of 1.0, and the activated carbon catalyst dosage has a minimum of 0.3 and a maximum of 0.7, then the maximum deviation for that cluster is ((1.0 - 0.5) + (0.7 - 0.3)) / 2 = 0.45. The maximum wastewater deviation and maximum dosage deviation for that cluster were summed to form the cluster deviation set for that cluster. Each clustering result is processed separately in the above manner to obtain multiple cluster deviation amplitude sets.

[0052] Next, the mean of the maximum wastewater deviation and maximum dosage deviation within each cluster deviation set is calculated to obtain multiple cluster cost coefficients. This cluster cost coefficient reflects the degree of fluctuation in the wastewater treatment data of samples within the same cluster. The greater the difference, the larger the cluster cost coefficient, indicating greater data fluctuation and lower data value. For example, if the wastewater characteristics within a cluster are similar but the corresponding catalyst dosages vary significantly, the cost coefficient for that cluster will be high, indicating poor data consistency within the cluster, which is not conducive to subsequent catalyst dosage recommendations.

[0053] Through the above steps, cluster analysis and calculation of clustering cost coefficients were performed on the sample wastewater treatment dataset, providing a quantitative indicator for evaluating the data quality of each clustering result. Compared to traditional methods, calculating the clustering cost coefficient can identify clustering results with large data fluctuations and poor consistency, preventing the adverse impact of these low-quality data on subsequent catalyst dosage recommendations, and improving the accuracy and reliability of the recommendation results.

[0054] Furthermore, clustering is performed on the sample wastewater treatment data set to obtain multiple clustering results, including:

[0055] S211, calculating the similarity between each sample wastewater treatment data and other sample wastewater treatment data in the sample wastewater treatment data set, wherein the similarity between the sample wastewater data, the sample catalyst dosage, and the sample treatment data is calculated and weighted to obtain the similarity;

[0056] S212, adding multiple sample wastewater treatment data with the greatest similarity into a clustering result;

[0057] S213. Complete clustering of all sample wastewater treatment data to obtain multiple clustering results.

[0058] In a feasible implementation, for each sample wastewater treatment data in the sample wastewater treatment data set, its similarity with each other sample wastewater treatment data in the sample wastewater treatment data set is calculated respectively. The similarity calculation includes three parts: sample wastewater data similarity, sample catalyst dosage similarity and sample treatment data similarity. For the sample wastewater data similarity, the similarity of each wastewater index parameter (such as COD, BOD, pH value, etc.) is calculated, and then the average value is taken as the sample wastewater data similarity. For example, the Euclidean distance method is used to calculate the normalized distance between the two sample wastewater data on each index parameter, and the distance is converted into the corresponding similarity. The smaller the distance, the higher the similarity. For the catalyst dosage similarity, the similarity of various catalyst dosages is calculated, and then the average value is taken as the sample catalyst dosage similarity. For example, the Euclidean distance method is also used to calculate the distance between the two sample catalyst dosage vectors and convert it into the corresponding similarity. For sample processing data similarity, the similarity of each post-treatment indicator parameter is similarly calculated, and the average is taken as the sample processing data similarity. For example, similar to obtaining sample wastewater data similarity, the Euclidean distance method is used to calculate the normalized distance between the two sample processing data on each indicator parameter and convert it into the corresponding similarity. Finally, these three similarities are weighted to obtain a comprehensive similarity. For example, the sample wastewater data similarity weight can be set to 0.4, the sample catalyst dosage similarity weight to 0.3, and the sample processing data similarity weight to obtain the final comprehensive similarity.

[0059] Then, multiple sample wastewater treatment data with the greatest similarity are added to a clustering result. Specifically, based on the comprehensive similarity calculated in the previous step, multiple sample wastewater treatment data with the highest similarity are found, and these sample wastewater treatment data are grouped into the same cluster. For example, data with a similarity greater than a certain threshold (such as 0.8) are selected to be grouped into the same cluster, ensuring that the data in the clustering results have a high similarity, providing a basis for subsequent analysis. Afterwards, the above clustering process is repeated until all data in the sample wastewater treatment data set are assigned to a cluster. In this way, the entire sample wastewater treatment data set is divided into multiple clustering results, the data within each cluster has a high similarity, and the data between different clusters have a large difference.

[0060] Through the above steps, we achieved effective clustering of the sample wastewater treatment dataset. Compared to traditional clustering methods, this method considers the comprehensive similarity of wastewater data, catalyst usage, and treatment data, resulting in more comprehensive and reasonable clustering results. Furthermore, by grouping the data with the greatest similarity into one category, we ensured the compactness and consistency of the clustering results, providing a sound data foundation for the subsequent calculation of the clustering cost coefficient and the treatment stability coefficient.

[0061] Furthermore, based on the cluster sample processing data in each clustering result, the treatment stability coefficient is calculated according to multiple wastewater indicators to obtain multiple treatment stability coefficients, including:

[0062] S31, extracting multiple sample processing data in each clustering result, and dividing them according to multiple wastewater indicators to obtain multiple processing indicator parameter sets;

[0063] S32, calculating the mean of the discrete parameters of the processing indicators in the plurality of processing indicator parameter sets to obtain the discrete parameters of the processing indicators;

[0064] S33, obtaining a preset processing discrete parameter, and calculating a ratio of the preset processing discrete parameter to the processing discrete parameter as a processing stability coefficient;

[0065] S34. Continue to calculate multiple processing stability coefficients to obtain multiple clustering results.

[0066] In a preferred embodiment, for each clustering result, the sample processing data portion of the wastewater treatment data for all samples within that cluster is extracted, namely, the various parameter records of the wastewater after catalytic treatment. These processed parameter records are then categorized according to different wastewater indicators (e.g., COD, BOD, pH, TSS, ammonia nitrogen, etc.) to form multiple treatment parameter sets. For example, the COD values ​​of all sample processing data within the cluster are combined into a COD treatment parameter set, all BOD values ​​into a BOD treatment parameter set, and so on, to form multiple treatment parameter sets for different indicators.

[0067] Then, the mean of the discrete parameters of the treatment indicators in multiple treatment indicator parameter sets is calculated to obtain the treatment discrete parameters. Specifically, for each treatment indicator parameter set, its standard deviation and other statistics are calculated as the treatment indicator discrete parameters of the indicator, which are used to measure the degree of fluctuation of the indicator after treatment. Since the dimensions of different wastewater indicators are different, the treatment indicator discrete parameters of each indicator need to be normalized to make them comparable. Normalization can be achieved by dividing by the average or maximum value of the indicator to convert different treatment indicator discrete parameters to the same scale. The smaller the normalized treatment indicator discrete parameter is, the more stable the treatment effect is. Subsequently, the mean of all normalized treatment indicator discrete parameters is calculated to obtain the treatment discrete parameters of the entire cluster. For example, if the normalized treatment indicator discrete parameters after treatment in a cluster include a COD standard deviation of 0.05, a BOD standard deviation of 0.04, and a pH standard deviation of 0.03, then the treatment discrete parameter of the cluster is (0.05+0.04+0.03) / 3=0.04.

[0068] Next, the preset processing discrete parameter is obtained, and the ratio of the preset processing discrete parameter and the processing discrete parameter is calculated as the processing stability coefficient. The preset processing discrete parameter is a processing indicator fluctuation threshold value pre-set by experts based on historical experience or industry standards, which represents an acceptable fluctuation range of the processing effect. The processing stability coefficient is obtained by calculating the ratio of the preset processing discrete parameter and the processing discrete parameter. When the processing discrete parameter is less than the preset processing discrete parameter, the ratio is greater than 1, indicating that the processing effect is more stable than expected; when the processing discrete parameter is greater than the preset processing discrete parameter, the ratio is less than 1, indicating that the processing effect fluctuates greatly. For example, if the preset processing discrete parameter is 0.08, and the actual processing discrete parameter of a cluster is 0.04, the processing stability coefficient of the cluster is 0.08 / 0.04=2, indicating that the processing effect of the cluster is more stable than expected. Afterwards, for each clustering result obtained, the above steps S31-S33 are repeated to calculate the processing stability coefficient of each cluster, and obtain multiple processing stability coefficients corresponding to multiple clustering results. These treatment stability coefficients reflect the stability of the catalyst treatment effect within each cluster and will serve as an important basis for subsequent training weight allocation.

[0069] Through the above steps, we achieved a quantitative assessment of the processing stability of each clustering result, providing reliable data support for subsequent catalyst dosage recommendations. By calculating the processing stability coefficient, we can identify clustering results with stable processing effects and high data value. These high-quality data are prioritized for training, improving the accuracy and reliability of catalyst dosage recommendations.

[0070] Furthermore, according to the multiple clustering cost coefficients and the multiple processing stability coefficients, training weights are assigned to the multiple clustering results, and an integrated catalyst dosage recommender is trained. The current wastewater data is input into the integrated catalyst dosage recommender, and the recommended catalyst dosage is output, including:

[0071] S41. Calculating and obtaining a plurality of clustering quality coefficients according to the plurality of clustering cost coefficients;

[0072] S42: Calculate and obtain multiple training priority coefficients based on the multiple clustering quality coefficients and the multiple processing stability coefficients, and allocate and calculate to obtain multiple first training weights;

[0073] S43, training a first catalyst dosage recommender using multiple clustering results according to the multiple first training weights, and continuing to train multiple catalyst dosage recommenders to obtain an integrated catalyst dosage recommender;

[0074] S44. Obtain current wastewater data, randomly configure catalyst dosage and input it into an integrated catalyst dosage recommender, output the obtained processing data, optimize the catalyst dosage, and obtain the recommended catalyst dosage.

[0075] In a preferred embodiment, first, multiple clustering quality coefficients are calculated based on multiple clustering cost coefficients. Specifically, the clustering quality coefficient is obtained by subtracting the clustering cost coefficient from 1. The clustering quality coefficient reflects the data consistency and reliability of the clustering results. The smaller the clustering cost coefficient, the larger the corresponding clustering quality coefficient, indicating that the data consistency within the cluster is better and the data value is higher. For example, if the cost coefficient of a cluster is 0.3, then its corresponding clustering quality coefficient is 1-0.3=0.7, indicating that the cluster has high data consistency and reliability.

[0076] Then, based on multiple clustering quality coefficients and multiple processing stability coefficients, multiple training priority coefficients are calculated and multiple first training weights are obtained by allocation and calculation. Specifically, the clustering quality coefficient and the processing stability coefficient are weighted and combined to calculate the training priority coefficient. The training priority coefficient comprehensively reflects the data quality and processing effect stability of the clustering results, and is used to determine the importance of each cluster in subsequent training. Based on the training priority coefficient, a first training weight is assigned to each clustering result. The higher the training priority coefficient, the greater the assigned training weight. For example, using a normalization method, the training priority coefficient of each cluster is divided by the sum of the training priority coefficients of all clusters to obtain the first training weight; specifically, the sum of the training priority coefficients of all clusters is first calculated, and then the training priority coefficient of each cluster is divided by the sum to obtain the first training weight of each cluster. If the cluster quality coefficient of a cluster is 0.7 and the processing stability coefficient is 2, then its training priority coefficient is 0.7×2=1.4; if the training priority coefficients of all clusters are 1.4, 1.2, 0.8, and 0.6 respectively, and the total is 4.0, then the first training weight of the cluster is 1.4 / 4.0=0.35, indicating that the probability of the data in this cluster being selected during training is 35%.

[0077] Then, according to multiple first training weights, multiple clustering results are used to train the first catalyst dosage recommender, and multiple catalyst dosage recommenders are continued to be trained to obtain an integrated catalyst dosage recommender. Specifically, the first training weight is used as the selection probability, and training data is randomly selected from multiple clustering results to train the first catalyst dosage recommender. The catalyst dosage recommender is a model based on machine learning, whose input features are wastewater data and catalyst dosage, and whose output features are treatment data, which is used to predict the treatment effect under given wastewater characteristics and catalyst dosage conditions. Through multiple trainings, multiple catalyst dosage recommenders are obtained, and they are integrated into a comprehensive recommendation system, namely an integrated catalyst dosage recommender. The integration method can be simple averaging, weighted averaging or other integrated learning techniques, which improves the accuracy and stability of the prediction by integrating the outputs of multiple recommenders.

[0078] Then, according to multiple first training weights, multiple clustering results are used to train the first catalyst dosage recommender, and multiple catalyst dosage recommenders are continued to be trained to obtain an integrated catalyst dosage recommender. Specifically, the catalyst dosage recommender is trained based on the sample wastewater treatment data in the multiple clustering results. During the training process, training data are selected from each cluster according to multiple first training weights, and the data in the cluster with a high first training weight is more likely to be selected, ensuring that high-quality data plays a greater role in training. The catalyst dosage recommender is a machine learning model, such as a neural network, a random forest or a support vector machine, etc., whose input is wastewater data and catalyst dosage, and the output is predicted treatment data. By adjusting the model parameters, the predicted treatment data is made as close as possible to the actual treatment data, thereby establishing a mapping relationship between wastewater characteristics, catalyst dosage and treatment effect. In order to improve generalization ability and prediction accuracy, an integrated learning method is adopted to train multiple different catalyst dosage recommenders, such as using different model structures or different initial parameters to obtain multiple catalyst dosage recommenders with differences. Then, these catalyst dosage recommenders are combined into an integrated catalyst dosage recommender, wherein the integration method can be voting method, averaging method or stacking method, etc., which improves the overall prediction performance by integrating the advantages of multiple catalyst dosage recommenders.

[0079] Afterward, the current wastewater data is obtained, and the catalyst dosage is randomly configured and input into the integrated catalyst dosage recommender. The resulting processed data is then output, and the catalyst dosage is optimized to obtain the recommended catalyst dosage. Specifically, for the current wastewater data to be treated, the various indicator parameters are first obtained. Then, within the feasible range of catalyst dosage, multiple sets of different catalyst dosage configurations are randomly generated. For each configuration, the current wastewater data and catalyst dosage are used as input and fed into a trained integrated catalyst dosage recommender to predict the water quality indicators after treatment. The integrated catalyst dosage recommender integrates the prediction results of multiple catalyst dosage recommenders to provide more accurate predicted treatment data. By comparing the treatment data under different catalyst dosage configurations, the configuration that meets the treatment standard requirements (such as the compliance of various pollutant indicators) and the most economical catalyst dosage is found as the final recommended catalyst dosage. If multiple configurations all meet the treatment standard, the one with the lowest total catalyst dosage is selected to achieve the optimal balance between economic and environmental benefits.

[0080] Through the above steps, intelligent catalyst dosage recommendations based on historical wastewater treatment data are realized, which effectively solves the problem of inaccurate recommendations caused by the randomness and errors of historical wastewater treatment data, obtains accurate and reliable catalyst dosage recommendations, and improves the efficiency and effectiveness of wastewater treatment.

[0081] Furthermore, according to the multiple first training weights, the multiple clustering results are used to train the first catalyst dosage recommender, and the multiple catalyst dosage recommenders are continuously trained to obtain an integrated catalyst dosage recommender:

[0082] S431, respectively calculating the mean of multiple sample wastewater data, multiple sample catalyst usages, and multiple sample treatment data in multiple clustering results to obtain multiple training wastewater treatment data;

[0083] S432: Using the plurality of first training weights as a plurality of first selection probabilities, randomly selecting a preset number of training wastewater treatment data from the plurality of training wastewater treatment data, and training a first catalyst dosage recommender based on machine learning, wherein the input features of the catalyst dosage recommender are the wastewater data and the catalyst dosage, and the output features are the treatment data;

[0084] S433: multiplying the first training weight of the selected training wastewater treatment data by a preset weight adjustment coefficient to obtain a plurality of second training weights;

[0085] S433. Continue training the second catalyst dosage recommender according to the multiple second training weights until multiple catalyst dosage recommenders are obtained through training, thereby obtaining an integrated catalyst dosage recommender.

[0086] In a preferred embodiment, first, the mean of multiple sample wastewater data, multiple sample catalyst dosages and multiple sample treatment data in multiple clustering results are calculated respectively to obtain multiple training wastewater treatment data. Specifically, for each clustering result, the mean of each index parameter of all sample wastewater data in the cluster, the mean of various catalyst dosages of all sample catalyst dosages, and the mean of each index parameter of all sample treatment data are calculated respectively, and these mean values ​​are combined to form a training wastewater treatment data representing the characteristics of the cluster. In this way, each clustering result is converted into a representative training wastewater treatment data, which not only reduces the amount of data but also retains the core characteristics of the cluster. For example, if there are 10 sample wastewater treatment data in a cluster, the mean of each index of the wastewater data, the mean of various types of catalyst dosage and the mean of each index of the treatment data in these 10 data are calculated to form a new training wastewater treatment data representing the typical characteristics of the cluster.

[0087] Then, using multiple first training weights as multiple first selection probabilities, a preset number of training wastewater treatment data are randomly selected from multiple training wastewater treatment data, and based on machine learning, the first catalyst dosage recommender is trained, wherein the input features of the catalyst dosage recommender are wastewater data and catalyst dosage, and the output features are processing data. Specifically, based on the obtained first training weights, training data is selected by random sampling. The higher the first training weight of a cluster, the greater the probability that its corresponding training wastewater treatment data will be selected. For example, if the first training weight of a cluster is 0.35, then in the random selection process, the probability that the training wastewater treatment data corresponding to the cluster will be selected is 35%. Through this weight-based random selection, high-quality data is guaranteed to be used first in the training process. Subsequently, the selected training wastewater treatment data is used to train the first catalyst dosage recommender based on a machine learning algorithm (such as a neural network, a decision tree, a support vector machine, etc.). The input features of the first catalyst dosage recommender include wastewater data (such as COD, BOD, pH value, etc.) and catalyst dosage (such as the dosage of various catalysts), and the output features are treatment data (such as COD, BOD, pH value, etc. after treatment), which are used to predict the treatment effect under given wastewater characteristics and catalyst dosage conditions.

[0088] Next, a preset weight adjustment coefficient is used to multiply the first training weight of the selected training wastewater treatment data to obtain multiple second training weights. Specifically, to avoid reusing the same data in subsequent training, a preset weight adjustment coefficient (such as 0.9) is introduced. For the training wastewater treatment data selected for training the first catalyst dosage recommender, its corresponding first training weight is multiplied by the preset weight adjustment coefficient to obtain the second training weight. For the training wastewater treatment data not selected for training the first catalyst dosage recommender, its first training weight is directly used as the second training weight, maintaining its original selection probability. This weight adjustment mechanism reduces the probability of used data being selected again in subsequent training, prompting the catalyst dosage recommender to be exposed to more diverse data during training and improving its generalization ability. For example, if a certain training wastewater treatment data has been selected when training the first catalyst dosage recommender, its first training weight is 0.35, and the preset weight adjustment coefficient is 0.9, then its second training weight is 0.35×0.9=0.315, and the probability of being selected in subsequent training is reduced; for training wastewater treatment data that has not been selected, if its first training weight is 0.25, then its second training weight is still 0.25, maintaining the original selection probability.

[0089] Afterwards, the second catalyst dosage recommender is continued to be trained according to the multiple second training weights until multiple catalyst dosage recommenders are obtained through training to obtain an integrated catalyst dosage recommender. Specifically, a similar method to that of training the first catalyst dosage recommender is adopted, but the second training weight is used as the selection probability to select training data to train the second catalyst dosage recommender. Then, the weights are continued to be adjusted to train the third, fourth, and other multiple catalyst dosage recommenders until the preset number of recommenders is reached. Finally, these catalyst dosage recommenders with differences are integrated into a comprehensive system, namely an integrated catalyst dosage recommender. The integration method can be simple averaging, weighted averaging, or stacked integration, etc., which improves the accuracy and stability of the overall prediction by integrating the prediction results of multiple models.

[0090] Through the above steps, training weight allocation based on clustering quality and processing stability, as well as multi-model training and integration based on weight reduction are achieved, making full use of high-quality data while avoiding over-reliance on certain data, thereby improving the accuracy and reliability of catalyst dosage recommendations.

[0091] Furthermore, the current wastewater data is obtained, the catalyst dosage is randomly configured and input into the integrated catalyst dosage recommender, the processing data is output, the catalyst dosage is optimized, and the recommended catalyst dosage is obtained, including:

[0092] S441. Obtain current wastewater data, randomly configure catalyst dosage, input the data into an integrated catalyst dosage recommender, output multiple predicted treatment data, and calculate the mean to obtain average treatment data;

[0093] S442. Continue to randomly configure the catalyst dosage until convergence, output the catalyst dosage with the minimum index parameters of multiple wastewater indicators in the average treatment data, and obtain the recommended catalyst dosage.

[0094] In a preferred embodiment, for the current wastewater that needs to be processed, first obtain its various index parameter data, such as COD, BOD, pH value, TSS, ammonia nitrogen, etc., to obtain the current wastewater data. Then, within the range of catalyst dosage, randomly generate multiple groups of different catalyst dosage configuration schemes. The current wastewater data and each group of catalyst dosage configuration schemes are used as input and sent to the trained integrated catalyst dosage recommender to obtain multiple predicted processing data. Since the integrated catalyst dosage recommender is composed of multiple catalyst dosage recommenders, each catalyst dosage recommender gives slightly different prediction results, it is necessary to calculate the mean of these prediction results to obtain more reliable average processing data. For example, if three groups of different catalyst dosage configuration schemes are used to predict a certain wastewater, and three groups of processing data are obtained respectively, the average value of these three groups of data is calculated as the expected processing effect of the configuration scheme, that is, the average processing data.

[0095] The catalyst dosage is then randomly configured until convergence is achieved. The recommended catalyst dosage is then output, minimizing the average treatment data across multiple wastewater indicators. Specifically, through an iterative optimization process, new catalyst dosage configurations are continuously generated and their treatment effects are predicted until a configuration that meets the treatment requirements and exhibits the optimal catalyst dosage is found. Convergence occurs when, after multiple iterations, the predicted treatment effect and catalyst dosage no longer show significant improvement, or when a preset upper limit on the number of iterations is reached. After convergence, the configuration with the lowest total catalyst dosage is selected from all configurations that meet the treatment criteria (e.g., all pollutant indicators meet the standards) as the final recommended catalyst dosage. If multiple configurations have similar total catalyst dosages, the optimal overall performance can be selected based on treatment stability or other economic factors. For example, if, after multiple iterations, three configurations are found to meet the treated water quality standards, the one with the lowest total catalyst dosage is selected as the recommended configuration, achieving the optimal balance between economic and environmental benefits.

[0096] Through these steps, intelligent optimization of catalyst dosage using an integrated catalyst dosage recommender was achieved. Compared to traditional empirical or trial-and-error methods, this improved the accuracy and efficiency of catalyst dosage recommendations. Through multiple predictions and optimization comparisons, the optimal configuration was found that both ensured treatment effectiveness and conserved catalyst usage. This enabled intelligent and refined management of the wastewater treatment process, with significant economic and environmental benefits.

[0097] Example 2, as Figure 2 As shown, based on the same inventive concept as the method for automatically recommending catalyst dosage based on wastewater detection data provided in Example 1, an embodiment of the present invention also provides an automatic catalyst dosage recommendation system based on wastewater detection data, including:

[0098] The data acquisition module 11 is used to collect a sample wastewater data set, a sample catalyst usage data set, and a sample treatment data set based on wastewater treatment detection data in a historical period, and combine them to obtain a sample wastewater treatment data set, wherein the wastewater data and the treatment data include indicator parameters of multiple wastewater indicators;

[0099] A cluster analysis module 12 is used to cluster the sample wastewater treatment data set to obtain multiple clustering results and calculate multiple clustering cost coefficients;

[0100] The stability calculation module 13 is used to calculate the treatment stability coefficient according to the cluster sample processing data in each clustering result and the multiple wastewater indicators to obtain multiple treatment stability coefficients;

[0101] The catalyst recommendation module 14 is used to assign training weights to the multiple clustering results according to the multiple clustering cost coefficients and the multiple processing stability coefficients, train the integrated catalyst dosage recommender, input the current wastewater data into the integrated catalyst dosage recommender, and output the recommended catalyst dosage.

[0102] Furthermore, the execution steps of the data acquisition module 11 include:

[0103] Based on the wastewater treatment test data in the historical period, the indicator parameters of multiple wastewater indicators in the wastewater before each wastewater treatment are collected to obtain a sample wastewater data set;

[0104] Based on the wastewater treatment test data in the historical period, the catalyst dosage during each wastewater treatment and the index parameters of multiple wastewater indicators in the treated wastewater are collected to obtain a sample catalyst dosage set and a sample treatment data set;

[0105] The sample wastewater data set, the sample catalyst usage set and the sample treatment data set are combined to obtain a sample wastewater treatment data set, wherein each piece of sample wastewater treatment data includes sample wastewater data, sample catalyst usage and sample treatment data.

[0106] Furthermore, the cluster analysis module 12 executes the following steps:

[0107] Clustering the sample wastewater treatment data set to obtain a plurality of clustering results, wherein each clustering result includes a plurality of sample wastewater treatment data;

[0108] Calculating the maximum deviation amplitudes of the sample wastewater data and the sample catalyst usage in the multiple sample wastewater treatment data in each clustering result to obtain multiple cluster deviation amplitude sets, wherein the maximum deviation amplitude includes the maximum wastewater deviation amplitude of the sample wastewater data and the maximum usage deviation amplitude of the sample catalyst usage;

[0109] The mean of the maximum wastewater deviation amplitude and the maximum usage deviation amplitude in each cluster deviation amplitude set is calculated to obtain multiple cluster cost coefficients.

[0110] Furthermore, the cluster analysis module 12 may further execute the following steps:

[0111] Calculating the similarity between each sample wastewater treatment data and other sample wastewater treatment data in the sample wastewater treatment data set, wherein the similarity between the sample wastewater data, the sample catalyst dosage and the sample treatment data is calculated and weighted to obtain the similarity;

[0112] Add multiple sample wastewater treatment data with the greatest similarity into one clustering result;

[0113] Clustering was completed for all sample wastewater treatment data, and multiple clustering results were obtained.

[0114] Furthermore, the execution steps of the stability calculation module 13 include:

[0115] Extracting multiple sample processing data within each clustering result, and dividing them according to multiple wastewater indicators to obtain multiple processing indicator parameter sets;

[0116] Calculating the mean of discrete parameters of processing indicators in multiple processing indicator parameter sets to obtain discrete processing parameters;

[0117] Obtaining a preset processing discrete parameter, and calculating a ratio of the preset processing discrete parameter to the processing discrete parameter as a processing stability coefficient;

[0118] Continue to calculate multiple processing stability coefficients to obtain multiple clustering results.

[0119] Furthermore, the catalyst recommendation module 14 executes the following steps:

[0120] Calculating and obtaining a plurality of clustering quality coefficients according to the plurality of clustering cost coefficients;

[0121] Calculating multiple training priority coefficients according to multiple clustering quality coefficients and multiple processing stability coefficients, and allocating and calculating to obtain multiple first training weights;

[0122] According to the plurality of first training weights, the plurality of clustering results are used to train a first catalyst dosage recommender, and the plurality of catalyst dosage recommenders are continuously trained to obtain an integrated catalyst dosage recommender;

[0123] Obtain the current wastewater data, randomly configure the catalyst dosage and input it into the integrated catalyst dosage recommender, output the processed data, optimize the catalyst dosage, and obtain the recommended catalyst dosage.

[0124] Furthermore, the catalyst recommendation module 14 may further execute the following steps:

[0125] Calculating the mean of multiple sample wastewater data, multiple sample catalyst dosages, and multiple sample treatment data in multiple clustering results respectively to obtain multiple training wastewater treatment data;

[0126] Using the multiple first training weights as multiple first selection probabilities, randomly selecting a preset number of training wastewater treatment data from the multiple training wastewater treatment data, and training a first catalyst dosage recommender based on machine learning, wherein the input features of the catalyst dosage recommender are the wastewater data and the catalyst dosage, and the output feature is the treatment data;

[0127] Using a preset weight adjustment coefficient, multiplying the first training weight of the selected training wastewater treatment data to obtain a plurality of second training weights;

[0128] The second catalyst dosage recommender is continuously trained according to the plurality of second training weights until a plurality of catalyst dosage recommenders are obtained through training, thereby obtaining an integrated catalyst dosage recommender.

[0129] Furthermore, the catalyst recommendation module 14 may further execute the following steps:

[0130] Obtain current wastewater data, randomly configure catalyst dosage, input into the integrated catalyst dosage recommender, output multiple predicted treatment data, calculate the mean to obtain average treatment data;

[0131] Continue to randomly configure the catalyst dosage until convergence, output the catalyst dosage with the minimum index parameters of multiple wastewater indicators in the average treatment data, and obtain the recommended catalyst dosage.

[0132] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0133] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0134] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0135] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0136] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0137] Although preferred embodiments of the present invention have been described, additional changes and modifications to these embodiments may occur to those skilled in the art once the basic inventive concepts become known.

[0138] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for automatically recommending catalyst dosage based on wastewater testing data, characterized in that: The method comprises: Based on historical wastewater treatment detection data, a sample wastewater data set, a sample catalyst usage set, and a sample treatment data set are collected and combined to obtain a sample wastewater treatment data set, wherein the wastewater data and the treatment data include indicator parameters of multiple wastewater indicators; Clustering the sample wastewater treatment data set to obtain multiple clustering results, and calculating multiple clustering cost coefficients, including: Clustering the sample wastewater treatment data set to obtain a plurality of clustering results, wherein each clustering result includes a plurality of sample wastewater treatment data; Calculating the maximum deviation amplitudes of the sample wastewater data and the sample catalyst usage in the multiple sample wastewater treatment data in each clustering result to obtain multiple cluster deviation amplitude sets, wherein the maximum deviation amplitude includes the maximum wastewater deviation amplitude of the sample wastewater data and the maximum usage deviation amplitude of the sample catalyst usage; Calculate the mean of the maximum wastewater deviation amplitude and the maximum usage deviation amplitude in each cluster deviation amplitude set to obtain multiple cluster cost coefficients; Based on the cluster sample processing data in each clustering result, the treatment stability coefficient is calculated according to multiple wastewater indicators to obtain multiple treatment stability coefficients, including: Extracting multiple sample processing data within each clustering result, and dividing them according to multiple wastewater indicators to obtain multiple processing indicator parameter sets; Calculating the mean of discrete parameters of processing indicators in multiple processing indicator parameter sets to obtain discrete processing parameters; Obtaining a preset processing discrete parameter, and calculating a ratio of the preset processing discrete parameter to the processing discrete parameter as a processing stability coefficient; Continue to calculate multiple processing stability coefficients to obtain multiple clustering results; According to the multiple clustering cost coefficients and the multiple processing stability coefficients, the multiple clustering results are trained with weights allocated, and an integrated catalyst dosage recommender is trained. The current wastewater data is input into the integrated catalyst dosage recommender, and the recommended catalyst dosage is obtained as output.

2. The method for automatically recommending catalyst dosage based on wastewater detection data according to claim 1, characterized in that: Based on historical wastewater treatment test data, we collected sample wastewater datasets, sample catalyst usage datasets, and sample treatment datasets, and combined them to obtain sample wastewater treatment data, including: Based on the wastewater treatment test data in the historical period, the indicator parameters of multiple wastewater indicators in the wastewater before each wastewater treatment are collected to obtain a sample wastewater data set; Based on the wastewater treatment test data in the historical period, the catalyst dosage during each wastewater treatment and the index parameters of multiple wastewater indicators in the treated wastewater are collected to obtain a sample catalyst dosage set and a sample treatment data set; The sample wastewater data set, the sample catalyst usage set and the sample treatment data set are combined to obtain a sample wastewater treatment data set, wherein each piece of sample wastewater treatment data includes sample wastewater data, sample catalyst usage and sample treatment data.

3. The method for automatically recommending catalyst dosage based on wastewater detection data according to claim 1, characterized in that: Clustering is performed on the sample wastewater treatment dataset to obtain multiple clustering results, including: Calculating the similarity between each sample wastewater treatment data and other sample wastewater treatment data in the sample wastewater treatment data set, wherein the similarity between the sample wastewater data, the sample catalyst dosage and the sample treatment data is calculated and weighted to obtain the similarity; Add multiple sample wastewater treatment data with the greatest similarity into one clustering result; Clustering was completed for all sample wastewater treatment data, and multiple clustering results were obtained.

4. The method for automatically recommending catalyst dosage based on wastewater detection data according to claim 1, characterized in that: According to the multiple clustering cost coefficients and the multiple processing stability coefficients, training weights are assigned to the multiple clustering results, and an integrated catalyst dosage recommender is trained. Current wastewater data is input into the integrated catalyst dosage recommender, and a recommended catalyst dosage is obtained by output, including: Calculating and obtaining a plurality of clustering quality coefficients according to the plurality of clustering cost coefficients; Calculating multiple training priority coefficients according to multiple clustering quality coefficients and multiple processing stability coefficients, and allocating and calculating to obtain multiple first training weights; According to the plurality of first training weights, the plurality of clustering results are used to train a first catalyst dosage recommender, and the plurality of catalyst dosage recommenders are continuously trained to obtain an integrated catalyst dosage recommender; Obtain the current wastewater data, randomly configure the catalyst dosage and input it into the integrated catalyst dosage recommender, output the processed data, optimize the catalyst dosage, and obtain the recommended catalyst dosage.

5. The method for automatically recommending catalyst dosage based on wastewater detection data according to claim 4, characterized in that: According to the multiple first training weights, the multiple clustering results are used to train the first catalyst dosage recommender, and the multiple catalyst dosage recommenders are continuously trained to obtain an integrated catalyst dosage recommender: Calculating the mean of multiple sample wastewater data, multiple sample catalyst dosages, and multiple sample treatment data in multiple clustering results respectively to obtain multiple training wastewater treatment data; Using the multiple first training weights as multiple first selection probabilities, randomly selecting a preset number of training wastewater treatment data from the multiple training wastewater treatment data, and training a first catalyst dosage recommender based on machine learning, wherein the input features of the catalyst dosage recommender are the wastewater data and the catalyst dosage, and the output feature is the treatment data; Using a preset weight adjustment coefficient, multiplying the first training weight of the selected training wastewater treatment data to obtain a plurality of second training weights; The second catalyst dosage recommender is continuously trained according to the plurality of second training weights until a plurality of catalyst dosage recommenders are obtained through training, thereby obtaining an integrated catalyst dosage recommender.

6. The method for automatically recommending catalyst dosage based on wastewater detection data according to claim 4, characterized in that: Obtain current wastewater data, randomly configure catalyst dosage and input it into the integrated catalyst dosage recommender, output the obtained processing data, optimize the catalyst dosage, and obtain the recommended catalyst dosage, including: Obtain current wastewater data, randomly configure catalyst dosage, input into the integrated catalyst dosage recommender, output multiple predicted treatment data, calculate the mean to obtain average treatment data; Continue to randomly configure the catalyst dosage until convergence, output the catalyst dosage with the minimum index parameters of multiple wastewater indicators in the average treatment data, and obtain the recommended catalyst dosage.

7. The automatic catalyst dosage recommendation system based on wastewater detection data is characterized by: The system comprises: A data acquisition module is used to collect a sample wastewater data set, a sample catalyst usage set, and a sample treatment data set based on wastewater treatment detection data within a historical period, and combine them to obtain a sample wastewater treatment data set, wherein the wastewater data and treatment data include indicator parameters of multiple wastewater indicators; The cluster analysis module is used to cluster the sample wastewater treatment data set to obtain multiple clustering results and calculate multiple clustering cost coefficients, including: Clustering the sample wastewater treatment data set to obtain a plurality of clustering results, wherein each clustering result includes a plurality of sample wastewater treatment data; Calculating the maximum deviation amplitudes of the sample wastewater data and the sample catalyst usage in the multiple sample wastewater treatment data in each clustering result to obtain multiple cluster deviation amplitude sets, wherein the maximum deviation amplitude includes the maximum wastewater deviation amplitude of the sample wastewater data and the maximum usage deviation amplitude of the sample catalyst usage; Calculate the mean of the maximum wastewater deviation amplitude and the maximum usage deviation amplitude in each cluster deviation amplitude set to obtain multiple cluster cost coefficients; The stability calculation module is used to calculate the treatment stability coefficient according to the cluster sample processing data in each clustering result according to multiple wastewater indicators, and obtain multiple treatment stability coefficients, including: Extracting multiple sample processing data within each clustering result, and dividing them according to multiple wastewater indicators to obtain multiple processing indicator parameter sets; Calculating the mean of discrete parameters of processing indicators in multiple processing indicator parameter sets to obtain discrete processing parameters; Obtaining a preset processing discrete parameter, and calculating a ratio of the preset processing discrete parameter to the processing discrete parameter as a processing stability coefficient; Continue to calculate multiple processing stability coefficients to obtain multiple clustering results; The catalyst recommendation module is used to assign training weights to the multiple clustering results according to the multiple clustering cost coefficients and multiple processing stability coefficients, train the integrated catalyst dosage recommender, input the current wastewater data into the integrated catalyst dosage recommender, and output the recommended catalyst dosage.

Images