Intelligent control method for wastewater treatment devices at dry bulk cargo terminal

Abstract

The present invention relates to the technical field of the control of wastewater treatment devices. Disclosed is an intelligent control method for wastewater treatment devices at a dry bulk cargo terminal, which is used for solving the problem of poor control of wastewater treatment devices at a terminal. The method comprises the following steps: installing a plurality of types of sensors at key locations of a dry bulk cargo terminal, and using edge computing nodes to perform real-time data collection and preprocessing; on the basis of historical features and temporal features, using a machine learning model to perform wastewater type classification, thereby realizing efficient dynamic adjustment of operating parameters of wastewater treatment devices; then, by means of weighted voting and confidence evaluation, integrating a plurality of classification results to ensure an optimal treatment effect; and analyzing actual wastewater treatment conditions to continuously optimize device control, thereby preventing faults, extending the service life of devices, and improving the wastewater treatment effect.

Description

Intelligent control method of dry bulk cargo wharf sewage treatment equipment TECHNICAL FIELD

[0001] The present application relates to the technical field of sewage treatment equipment control, in particular to an intelligent control method of dry bulk cargo wharf sewage treatment equipment. BACKGROUND

[0002] Dry bulk cargo wharfs are usually equipped with large-scale handling equipment such as belt conveyors, grab cranes, loaders, etc. to efficiently load and unload and transport goods. In addition, these wharfs also have corresponding storage facilities such as yards and silos to meet the needs of temporary storage and transit of goods. Especially in the import and export of mineral resources and agricultural commodities, through dry bulk cargo wharfs, goods can be efficiently transported from production areas to consumption areas, ensuring the smooth and stable supply chain.

[0003] The prior art has the following disadvantages: In the loading and unloading area of the dry bulk cargo wharf, frequent flushing is often required. These flushing waters can carry away the mud, oil stains and bulk materials (such as ore dust, grain residues, etc.) on the ground, forming sewage. The ships docked at the wharf can also produce sewage, including domestic sewage (such as crew's domestic wastewater), oil-containing sewage from the engine room, and cleaning wastewater from the cargo hold, etc. The existing dry bulk cargo wharf sewage treatment equipment has the problems of incomplete data collection, insufficient intelligent control strategy, insufficient fault prediction and prevention, and low energy consumption and resource utilization rate. These deficiencies result in the inability to monitor and record sewage-related data in real time, making it difficult for the sewage treatment control system to adapt to complex and changing demands, resulting in increased equipment maintenance costs and low operating efficiency.

[0004] The above information disclosed in the background section is only intended to enhance the understanding of the background of the present disclosure, and therefore it can include information that does not constitute the prior art known to those of ordinary skill in the art. SUMMARY

[0005] The purpose of the present application is to provide an intelligent control method of dry bulk cargo wharf sewage treatment equipment. The present application installs multiple types of sensors at key positions in the dry bulk cargo wharf and uses edge computing nodes for real-time data collection and preprocessing. Combined with historical characteristics and time characteristics, a machine learning model is used for sewage type classification, achieving efficient dynamic adjustment of sewage treatment equipment operating parameters. Through weighted voting and confidence assessment, multiple classification results are integrated to ensure the best treatment effect, thereby solving the deficiencies in the background art.

[0006] In order to achieve the above-mentioned purpose, the present application provides the following technical scheme: An intelligent control method of dry bulk cargo wharf sewage treatment equipment, comprising the following steps:

[0007] A plurality of types of sensors are installed at different positions of the dry bulk cargo terminal, and sewage data are collected in real time, preliminary data preprocessing and analysis are performed on the sewage treatment, and a preliminary classified sewage type is obtained;

[0008] Historical feature extraction and time feature extraction are performed on historical data, the extracted historical features and time features are combined with the preprocessed data to form a complete feature vector for input of a machine learning model;

[0009] The model classifies according to the input feature vector, outputs the prediction probability or classification label of each sewage type, combines the rule-based classification and machine learning classification results, combines the weighted voting and confidence evaluation to obtain the final classification result, and dynamically adjusts the operation parameters of the sewage treatment equipment according to the final classification result;

[0010] The operation processing information generated after the operation parameter adjustment of the sewage treatment equipment is obtained, the sewage treatment situation after the adjustment of the operation parameters of the equipment according to the final classification result is determined, and the adjustment and optimization of the sewage equipment control are performed according to the obtained processing result.

[0011] Preferably, the preliminary data preprocessing and analysis of the sewage treatment are performed to obtain the preliminary classified sewage type, including:

[0012] The sensor is connected to an edge computing node, the data collection frequency and data format are set through the edge computing node, and mean filtering is used for denoising of the data;

[0013] The data are standardized, and the data are converted into a unified unit and scale;

[0014] The preset water quality parameter threshold is determined according to various water quality parameters of the sewage collected in real time, and the water quality parameter threshold is used as a classification condition;

[0015] According to the preset water quality parameter threshold, each data sample is traversed, each water quality parameter of each data sample is compared with the water quality parameter threshold, and a preliminary classification is performed according to the comparison result, and a classification result of each data sample is marked.

[0016] Preferably, the historical feature extraction and time feature extraction are performed on the historical data, the extracted historical features and time features are combined with the preprocessed data to form a complete feature vector for input of a model, and the specific steps are as follows:

[0017] According to the time feature, the change of the sewage feature in different time periods is determined, a time label is added to each data sample, and the data sample is divided into different time periods according to the collection time;

[0018] Calculate the moving average and moving standard deviation in the time window in the historical data, and perform historical feature extraction according to the calculation results;

[0019] After extracting and constructing the time features and historical features, the preprocessed real-time data and the extracted time features and historical features are combined and input into the model;

[0020] The model includes random forest, support vector machine, neural network.

[0021] Preferably, the weighted voting and confidence evaluation are combined to obtain the final classification result in combination with the rule-based classification and machine learning classification results, and the specific steps are as follows:

[0022] According to the preset water quality parameter threshold, the rule classification of the data sample is performed, and one or more sewage types are obtained for each data sample according to the rule classification;

[0023] For each data sample, the classification results of the machine learning model and the rule-based classification results are weighted voted, and different weights are determined for each classification method according to the accuracy;

[0024] The data is classified using a machine learning model, and the total confidence of each category is obtained according to the confidence evaluation of each category in the rule classification;

[0025] According to the weighted voting combined with the confidence evaluation method, the final sewage type of each data sample is determined, and the final classification result of each data sample is output.

[0026] Preferably, the running processing information generated after the running parameter adjustment of the sewage treatment equipment is obtained, and the sewage treatment situation after the equipment running parameter adjustment according to the final classification result is determined, and the specific steps are as follows:

[0027] The running processing information generated after the running parameter adjustment of the sewage treatment equipment is obtained, and the running processing information includes equipment running information and settlement influence information;

[0028] The equipment running information includes a running equipment response stability index, and the settlement influence information includes a multi-dimensional settlement deviation index:

[0029] The running equipment response stability index and the multi-dimensional settlement deviation index are generated together to form an equipment control coefficient;

[0030] The running equipment response stability index is inversely proportional to the equipment control coefficient, and the multi-dimensional settlement deviation index is proportional to the equipment control coefficient.

[0031] Preferably, the sewage equipment control is adjusted and optimized according to the obtained processing result, and the specific steps include:

[0032] The generated device control coefficient is compared with the intelligent control threshold value;

[0033] If the device control coefficient is greater than the intelligent control threshold value, a sewage treatment abnormal signal is generated, and the sewage treatment control scheme is adjusted;

[0034] If the device control coefficient is less than or equal to the intelligent control threshold value, a sewage treatment stable signal is generated, and no sewage treatment control is needed, and monitoring and maintenance are continued.

[0035] In the above technical solution, the present application provides technical effects and advantages:

[0036] The present application installs multiple types of sensors at key positions of the dry bulk cargo terminal, and uses edge computing nodes for real-time data acquisition and preprocessing, combines historical features and time features, uses a machine learning model for sewage type classification, realizes efficient dynamic adjustment of sewage treatment equipment operation parameters, and through weighted voting and confidence evaluation, comprehensive multiple classification results are obtained to ensure the best treatment effect, and through analysis of the actual sewage treatment situation, the equipment control is continuously optimized, thereby preventing faults, prolonging the service life of the equipment, and improving the sewage treatment effect.

[0037] The sewage treatment equipment intelligent control method significantly improves the sewage treatment efficiency and effect, reduces the operation cost, enhances the stability and reliability of the equipment operation, realizes efficient, stable and intelligent management of sewage treatment through intelligent monitoring, data analysis and dynamic adjustment, and ensures that the effluent quality meets the standard and protects the environment. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments or prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0039] FIG. 1 is a flowchart of a sewage treatment equipment intelligent control method for a dry bulk cargo terminal according to the present application. DETAILED DESCRIPTION

[0040] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations may, however, be implemented in many different forms and should not be construed as limited to the examples set forth herein; rather, these example implementations are provided so that this disclosure will be thorough and complete, and will fully convey the inventive aspects to those skilled in the art.

[0041] Embodiment 1: The present application provides a sewage treatment equipment intelligent control method for a dry bulk cargo terminal as shown in FIG. 1, the method comprising the following steps:

[0042] Multiple types of sensors are installed at different locations in the dry bulk cargo terminal to collect wastewater data in real time. Preliminary data preprocessing and analysis are performed on the wastewater treatment to obtain a preliminary classification of wastewater types.

[0043] Historical features and time features are extracted from historical data. The extracted historical and time features are then combined with preprocessed data to form a complete feature vector for input to machine learning models.

[0044] The model classifies based on the input feature vector and outputs the predicted probability or classification label for each type of wastewater. Combining the results of rule-based classification and machine learning classification, the model combines weighted voting with confidence assessment to obtain the final classification result. The operating parameters of the wastewater treatment equipment are dynamically adjusted based on the final classification result.

[0045] The system acquires operational processing information generated after adjusting the operating parameters of the wastewater treatment equipment, determines the wastewater treatment status after adjusting the equipment operating parameters based on the final classification results, and adjusts and optimizes the control of the wastewater equipment based on the obtained processing results.

[0046] Key locations at the wharf were selected for monitoring, including inlets, outlets, treatment ponds, rainwater collection areas, flushing areas, mechanical cleaning areas, and ship sewage discharge outlets. Multiple types of sensors were installed at different locations at the dry bulk cargo wharf to collect comprehensive sewage data. The sensors included water quality monitoring sensors to monitor pH, dissolved oxygen, chemical oxygen demand (COD), biological oxygen demand (BOD), suspended solids concentration, and heavy metal content in the sewage. The water quality monitoring sensors were installed at inlets, outlets, treatment ponds, rainwater collection areas, and flushing areas.

[0047] Flow sensors measure sewage flow and are installed at the inlets and outlets of major drainage pipes. Temperature sensors monitor sewage temperature and are installed at key points such as inlets and treatment tanks. Oil pollution monitoring sensors monitor oil pollution content and are installed in mechanical cleaning areas and ship sewage discharge outlets.

[0048] Connect these sensors to edge computing nodes to ensure that data can be transmitted to edge computing devices in real time. Use IoT technologies such as LoRa and NB-IoT to transmit sensor data to the cloud in real time, store and manage sensor data in the cloud, and ensure data security and accessibility.

[0049] Edge computing nodes are installed at each monitoring point to perform preliminary data processing and analysis. The edge computing nodes process sensor data in real time, perform preprocessing (such as noise reduction and standardization), and classify wastewater types according to preset rules. The specific steps are as follows:

[0050] The data receiving function is configured on the edge computing node to receive real-time data from the sensor, and the data collection frequency and data format are set, a filtering algorithm mean filtering is applied for denoising processing to eliminate noise in the sensor data and ensure the accuracy of the data, and meanwhile, the data is standardized to convert the data into a unified unit and scale;

[0051] The mean filtering is to select a window size k, calculate the average value of each data point in the window, and replace the original data point. For a data sequence x1, x2, …, x n The data y i after mean filtering is calculated as follows: Wherein, k is the window size, and the value range of i is (k+1) / 2 to n-(k-1) / 2.

[0052] The data is standardized, and Min-Max standardization is used for processing. For a data sequence x1, x2, …, x n The standardized data x i is calculated as follows: Wherein, min represents the minimum value of the sequence, and max represents the maximum value of the sequence.

[0053] Rule-based classification, according to the preset water quality parameter threshold, the collected data is classified into multiple data samples, according to the real-time collected various water quality parameters of the sewage, such as pH value, COD concentration, dissolved oxygen (D0), etc., the water quality parameter threshold is determined, for example, the preliminary classification of pH<6 is strong acid sewage, the classification of pH>9 is strong alkaline sewage, the preliminary classification of COD concentration threshold COD>200mg / L is high organic pollution sewage, and DO<2mg / L is classified as low dissolved oxygen sewage.

[0054] According to the real-time collected various water quality parameters of the sewage, the preset water quality parameter threshold is determined, and the water quality parameter threshold is used as the classification condition.

[0055] Each collected data sample is traversed, and whether each water quality parameter meets the classification condition is checked. The water quality parameter threshold of each water quality parameter of each data sample is compared, each water quality parameter of each data sample is checked according to the preset classification condition, and the data sample is preliminarily classified. The classification result of each data sample is marked to ensure that the sewage treatment equipment can adopt the best treatment scheme according to different types of sewage.

[0056] Through the above specific steps, the sewage of the dry bulk cargo terminal can be effectively classified according to the preset water quality parameter threshold, the sewage treatment equipment can adopt the best treatment scheme according to different types of sewage, and the sewage treatment efficiency and effect are improved.

[0057] Extract the time characteristics (such as day, night) and historical characteristics (extracted using moving average, moving standard deviation, etc.) of historical data (historical sewage-related data of dry bulk cargo terminals), enhance data features, and input the preprocessed data into a trained machine learning model (such as random forest, support vector machine, neural network, etc.) for real-time classification. The specific steps are as follows:

[0058] Extract according to time characteristics, define time period, divide a day into different time periods (such as day, night, morning, evening, etc.) to determine the change of sewage characteristics in different time periods. The division of time periods can be based on actual experience and data analysis results. For example, day is from 6:00 to 18:00, night is from 18:00 to 6:00;

[0059] Add time labels to each data sample according to the collection time and divide them into different time periods;

[0060] Extract historical characteristics from historical data, calculate the moving average of the time window, reflect the short-term trend of the sewage parameter, for example, the 5-minute window pH value moving average, also calculate the moving standard deviation of the time window, reflect the volatility of the sewage parameter, for example, the 5-minute window pH value moving standard deviation, and calculate the historical maximum and minimum value in a certain time window, reflect the extreme change of the parameter.

[0061] After extracting and constructing time characteristics and historical characteristics, combine the preprocessed real-time data with the extracted time characteristics and historical characteristics to form a complete feature vector for model input;

[0062] Use recursive feature elimination, principal component analysis, etc. Feature selection algorithm to select the most influential features for classification, reduce feature redundancy, and improve model efficiency;

[0063] Principal Component Analysis (PCA) is a linear transformation that projects data onto a low-dimensional space, extracting principal components that best explain the variance of the data:

[0064] Calculate the data covariance matrix, then calculate the eigenvalues and eigenvectors of the covariance matrix, select the eigenvectors corresponding to the first k largest eigenvalues to form the eigenvector matrix W;

[0065] Project the data onto a low-dimensional space: Z = XW, where Z is the principal component matrix, X is the original data matrix, and W is the eigenvector matrix;

[0066] Divide the data set into training set and test set, generally use 80% training set and 20% test set ratio, use cross-validation method (such as k-fold cross-validation) to further divide the training set, to evaluate the stability and generalization ability of the model;

[0067] The training set is divided into k subsets, and the model is trained with k-1 subsets and tested with the remaining one subset, repeated k times, and the average performance index is calculated. The k-fold cross-validation formula is:

[0068] Where M i is the performance index of the i-th validation;

[0069] Select appropriate machine learning models according to specific needs, such as random forest, support vector machine (SVM), neural network, etc. Random forest is suitable for processing multi-feature data, with strong generalization ability and robustness. Support vector machine (SVM) is suitable for high-dimensional data classification and can find the optimal classification hyperplane. Neural network is suitable for modeling complex data relationships and can automatically extract features.

[0070] Train the selected model using the training set data, adjust the model parameters (such as the number of trees in random forest, the kernel function type in SVM, the number of layers and neurons in neural network) to optimize the model performance;

[0071] Optimize the model's hyperparameters using methods such as grid search or random search, for example, the hyperparameters of random forest include the number of trees and the maximum depth of the tree.

[0072] Perform k-fold cross-validation to evaluate the model's performance under different data divisions, ensuring the model's stability and generalization ability.

[0073] Evaluate the model using accuracy, precision, recall, F1 score, etc. to evaluate the model's performance. Accuracy is the proportion of data samples predicted correctly by the model, precision is the proportion of positive data samples predicted by the model that are actually positive data samples, recall is the proportion of positive data samples predicted by the model that are actually positive data samples, and F1 score is the harmonic mean of precision and recall.

[0074] Evaluate the trained model using test set data, calculate various evaluation indicators, and ensure the model's generalization ability and stability.

[0075] The model classifies according to input features, and the model outputs the predicted probability or classification label of each type of wastewater, for example, the model may predict a data sample as "high organic pollution water" or "strong acid / strong alkaline wastewater".

[0076] Combine the results of rule-based classification and machine learning classification, use weighted voting and confidence evaluation to combine, and obtain the final classification result. According to the final classification result, dynamically adjust the operating parameters of the wastewater treatment equipment to ensure the best treatment effect.

[0077] According to the preset water quality parameter threshold, the rule classification is performed. For example, if the pH value is less than 6 or greater than 9, it is preliminarily classified as "strong acid / strong alkaline sewage", and if the COD value exceeds a certain threshold (such as 200 mg / L), it is preliminarily classified as "high organic pollution water";

[0078] For each data sample, one or more possible sewage types are obtained according to the rule classification;

[0079] For each data sample, the classification results from the machine learning model and the rule-based classification results are weighted and voted, and each classification method can be given different weights according to its accuracy or reliability. For example, if the accuracy of the machine learning model is higher, it can be given a higher weight;

[0080] The specific steps of the weighted voting mechanism are as follows: using a machine learning model (such as random forest, SVM, neural network) to classify the data and obtain the prediction probability of each category;

[0081] According to the preset water quality parameter threshold, the rule classification is performed. For example, if the pH value is less than 6 or greater than 9, it is preliminarily classified as "strong acid / strong alkaline sewage", and if the COD value exceeds a certain threshold (such as 200 mg / L), it is preliminarily classified as "high organic pollution water";

[0082] The weights of the machine learning model and the rule-based classification are determined, and the weights are allocated according to the respective accuracy and reliability. Assuming that the weight of the machine learning model is w model , and the weight of the rule-based classification is w ruler ;

[0083] The results of each category are weighted and voted, and the prediction probability of the machine learning model and the vote of the rule-based classification are combined, where argmax represents the input value corresponding to the maximum value, LB is the category, c is the category (such as strong acid / strong alkaline sewage, high organic pollution water, etc.), p model (c) is the probability of the machine learning model predicting the classification as c, p ruler (c) is the vote (1 or 0) of the rule-based classification for the classification c, w model is the weight of the machine learning model, and w ruler is the weight of the rule-based classification.

[0084] If the machine learning model provides a probability distribution of the classification, these probabilities can be used as confidence indicators, and the final classification result is determined according to the confidence of the machine learning model and the rule-based classification;

[0085] Using a machine learning model to classify the data, obtaining the confidence (probability) of each category, and according to the rule classification, the confidence of each category can be set to a fixed value or dynamically adjusted according to the historical accuracy;

[0086] The confidence of each category is integrated to calculate the total confidence of each category, and the calculation formula is:

[0087] Wherein, c model (c) represents the confidence of the machine learning model for classification c, C ruler (c) represents the confidence of the rule-based classification for classification c;

[0088] According to the weighted voting combined with the confidence evaluation method, the final sewage type of each data sample is determined, and the final classification result of each data sample is output for further sewage treatment equipment control decision;

[0089] For example, the machine learning model predicts that the probability of a data sample being "strong acid / strong alkali sewage" is 0.7, and the probability of being "high organic pollution water" is 0.3;

[0090] According to the rule-based classification, the data sample is classified as "strong acid / strong alkali sewage" (voting 1), and not classified as "high organic pollution water" (voting 0), assuming the weight of the machine learning model is 0.7 and the weight of the rule-based classification is 0.3;

[0091] For "strong acid / strong alkali sewage": 0.7x0.7+0.3x1=0.49+0.3=0.79, for "high organic pollution water": 0.7x0.3+0.3x0=0.21;

[0092] The data sample is classified, and the final classification result is "strong acid / strong alkali sewage";

[0093] For "strong acid / strong alkali sewage": the confidence of the machine learning model is 0.7, the confidence of the rule-based classification is 0.5, and the integrated confidence is 0.7+0.5=1.20, for "high organic pollution water": the confidence of the machine learning model is 0.3, the confidence of the rule-based classification is 0.2, and the integrated confidence is 0.3+0.2=0.5, and the final classification result is "strong acid / strong alkali sewage";

[0094] By integrating the weighted voting and confidence evaluation method, the final classification result of the sewage data sample is obtained, and the classification result is used to guide the operation parameter adjustment of the sewage treatment equipment to ensure the best treatment effect.

[0095] Obtain the operation processing information generated after adjusting the operation parameters of the sewage treatment equipment, analyze the operation processing information, determine the sewage treatment situation after adjusting the operation parameters of the equipment according to the final classification result, and adjust and optimize the sewage equipment control according to the obtained processing result, the operation processing information includes equipment operation information, settlement influence information;

[0096] The device operation information includes an operation device response stability index, and the settlement influence information includes a multi-dimensional settlement deviation index. After collection, the operation device response stability index and the multi-dimensional settlement deviation index are respectively marked as YXS and DWC.

[0097] The operation device response stability index is a comprehensive index for evaluating the stability and reliability of the response of the sewage treatment device during the treatment process. The operation device response stability index reflects whether the device can make adjustments stably and efficiently when facing changes in sewage and maintain the stability of the treatment effect. By monitoring the adjustment time, operation parameter fluctuation, and failure frequency of the device and analyzing the results, problems in device operation can be identified and solved, treatment strategies can be optimized, the reliability and efficiency of the device can be improved, and the stability of the treatment effect can be ensured.

[0098] The operation device response stability index has the following aspects:

[0099] Improving the reliability of the device: By continuously monitoring the adjustment time, operation parameter fluctuation, and failure frequency of the device, potential problems in device operation can be discovered in a timely manner, signs of unstable device response can be identified in advance, device failures can be prevented, unplanned downtime can be reduced, and resources can be concentrated on maintaining unstable devices, thereby improving the reliability of the device;

[0100] Optimizing treatment strategies: Through feedback from the operation device response stability index, device operation parameters (such as aeration time, reagent dosage, stirring speed, etc.) can be dynamically adjusted, sewage treatment schemes can be optimized, and the response speed of the device to changes in sewage can be evaluated and improved, ensuring that the treatment scheme can adapt to changes in the characteristics of the sewage in a timely manner and improving the treatment efficiency;

[0101] Reducing operation costs: By improving the stability and reliability of the device, reducing the cost of frequent maintenance and replacement of parts, optimizing device operation parameters and treatment strategies, improving treatment efficiency, reducing energy consumption and reagent consumption, and saving operation costs.

[0102] For example, if an aeration device has large fluctuations in adjustment time when treating high-organic-load sewage, maintenance can be performed in advance according to the operation device response stability index, thereby avoiding downtime caused by device failure, and further adjusting the aeration time and reagent dosage to optimize the treatment effect and ensure that the effluent quality meets the standard;

[0103] The logic for obtaining the operation device response stability index is as follows:

[0104] By installing sensors and monitoring devices, the adjustment time, operation parameters, and failure occurrence of the device are monitored in real time, the start and end times of each adjustment operation are recorded, adjustment time data is obtained, and data sets of device operation, including adjustment time, operation parameters, failure occurrence time, and frequency, are recorded periodically, and actual adjustment time Adjact and the expected adjustment time Adj exp and calculate the standard deviation of the adjustment time and the average value, calculate the adjustment time deviation, the expression is:

[0105] Extract the time point and amplitude ΔParam of parameter change from the data set, record the time point t of parameter reaching steady state stable and the time point t of parameter change change , calculate the parameter change response time: Get the number of failures F of the equipment in the time period and the total running time T in the time period;

[0106] Get the standard deviation σ of the adjustment time Adj and the average value μ Adj , calculate the stable value of the adjustment time, the expression is: Get the standard deviation σ of the running parameter deviation Par and the average value μ Par , calculate the fluctuation value of the running parameter: Calculate the running equipment response stability index, the expression is:

[0107] It should be noted that the amplitude of parameter change is the difference between the initial value of the parameter and the stable value after adjustment. The initial value and the stable value of the parameter are extracted from the data, and the absolute value of the difference between the two is taken as the amplitude of parameter change; The number of failures of the equipment in the time period and the total running time are analyzed by statistics in the specified time period, and the specific time range is determined according to the actual demand.

[0108] The multi-dimensional settling deviation index is used to represent the deviation of the settling effect of sewage in the sewage treatment process. The multi-dimensional settling deviation index reflects the comprehensive performance, deviation and instability of the settling process by comprehensively evaluating the deviation of each key parameter in the settling process, so as to help identify and quantify the problems in the settling process, optimize the treatment process, and improve the overall effect and efficiency of sewage treatment.

[0109] The multi-dimensional settling deviation index has an impact on the following aspects:

[0110] Sewage treatment efficiency: By monitoring and optimizing the multi-dimensional settling deviation index, the deviation in the settling process can be found in time, and measures can be taken to adjust, so as to improve the efficiency of solid-liquid separation, ensure that the effluent water quality meets the standard, optimize the settling process, reduce unnecessary energy and reagent consumption, and improve resource utilization efficiency;

[0111] Environmental impact: by optimizing the sedimentation process, improving the removal efficiency of pollutants, reducing the emission of pollutants, protecting the environment, and improving the water quality, the treated wastewater can be more widely reused, saving water resources.

[0112] The logic of obtaining the multi-dimensional sedimentation deviation index is as follows:

[0113] Obtain the total suspended solid content TS in the sewage, the volume of the sludge after sedimentation TJ, and the interface sedimentation speed SD in the data sample, and obtain the ideal total suspended solid concentration value TS re , the ideal sludge volume index value TJ re , and the ideal interface sedimentation speed value SD re , calculate the total suspended solid deviation value:

[0114] Establish a total suspended solid deviation value set , a sludge volume index value set , and an interface sedimentation speed value set , and obtain the total suspended solid deviation standard value after standardization: n represents the total number of data samples, max represents, similarly, the sludge volume index standard value TJP and the interface sedimentation speed standard value SDP are calculated, the multi-dimensional sedimentation deviation index is calculated, and the calculation expression is:

[0115] It should be noted that the total suspended solid reflects the content of solid particles in the sewage, and the total suspended solid can be obtained by obtaining a certain amount of sewage sample from the sewage treatment facility, filtering and drying, and weighing, calculating the weight difference before and after drying, and the volume ratio of the sewage sample to obtain the total suspended solid content; obtain the sludge sample from the sedimentation tank or aeration tank in the sewage treatment facility, pour the sludge sample into a measuring cylinder, and let it stand for 30 minutes to measure the volume of the sludge after sedimentation; the interface sedimentation speed reflects the sinking speed of the solid-liquid interface in the sedimentation process, which is the sinking speed of the sludge suspension interface per unit time.

[0116] The obtained running equipment response stability index and multi-dimensional sedimentation deviation index are normalized to generate the equipment control coefficient, and the equipment control coefficient is calibrated as S x , and the formula is:

[0117] In the formula, γ and β are respectively the preset proportion coefficients of the running equipment response stability index and the multi-dimensional sedimentation deviation index, and γ and β are both greater than 0;

[0118] It should be noted that the formula is obtained by collecting a plurality of sets of data sample data by a person skilled in the art and setting a corresponding preset proportion coefficient for each set of data sample data. The specific calculation process is: the obtained running equipment response stability index and multi-dimensional sedimentation deviation index are substituted into the formula to form an equation group, and the calculated preset proportion coefficient is screened and the mean value is taken, so as to obtain the preset proportion coefficient. Since each set of data sample data is different, the specific value of the preset proportion coefficient under different data sample data is also different.

[0119] As can be seen from the formula, the smaller the running equipment response stability index and the larger the multi-dimensional sedimentation deviation index, that is, the larger the value of the equipment control coefficient, indicating that the equipment is unstable in response during operation, the adjustment is not timely, the running parameter fluctuation is large, the deviation in the sedimentation process is large, and there is a significant difference between the actual treatment effect and the expected target value. The device operating parameters and process conditions need to be optimized immediately to improve the stability of the device response and the sedimentation treatment effect, and to ensure that the effluent water quality meets the standard;

[0120] The larger the running equipment response stability index and the smaller the multi-dimensional sedimentation deviation index, that is, the smaller the performance value of the equipment control coefficient, indicating that the equipment shows high stability and reliability when responding to changes in sewage characteristics, and the adjustment is rapid and accurate. The running parameter fluctuation is extremely small, indicating that the equipment response is very stable, the adjustment is timely, the sedimentation process deviation is small, the treatment effect is excellent, the running state is good, the resource and energy consumption utilization efficiency is high, and the existing strategy can be continued and optimized to ensure the continuous efficiency and stability of the sewage treatment process.

[0121] The generated equipment control coefficient is compared with the intelligent control threshold value, and different signals are generated according to the comparison result, that is, a sewage treatment stable signal and a sewage treatment abnormal signal;

[0122] After obtaining the generated equipment control coefficient, the generated equipment control coefficient is compared with the intelligent control threshold value. If the equipment control coefficient is greater than the intelligent control threshold value, an abnormal sewage treatment signal is generated, indicating that the sewage treatment process operating state is abnormal, prompting the need to optimize the device operating parameters and process conditions immediately to improve the stability of the device response and the sedimentation treatment effect, and to ensure that the effluent water quality meets the standard. This mechanism can help to discover and solve problems in operation in time to ensure the efficiency and stability of the sewage treatment process.

[0123] If the equipment control coefficient is less than or equal to the intelligent control threshold value, a sewage treatment stable signal is generated, indicating that the sewage treatment equipment is stable in response under the current conditions, the treatment effect is ideal, the running state is good, and the effluent water quality meets the standard. In this case, the sewage treatment system is running normally and does not need to be adjusted urgently, but should continue to be monitored and maintained to ensure continuous efficient operation.

[0124] When the sewage treatment abnormal signal is generated, it indicates that the process operation state of the sewage treatment equipment is abnormal, and the following specific measures need to be taken immediately to optimize the equipment operation parameters and process conditions.

[0125] According to the analysis of the sewage situation treated during the time period when the sewage treatment abnormal signal is generated, according to the specific time and duration of the sewage treatment abnormal signal, the time trend graph of the key parameters during the abnormal signal period is drawn, the change of each parameter is observed, the correlation analysis is carried out, the correlation between the key parameters is calculated, which parameter change may cause the generation of abnormal signal is identified, the data during the abnormal signal period is compared with the data during the normal operation period, the abnormal points during the abnormal signal period are found out;

[0126] The time sequence data before and after the generation of the abnormal signal is analyzed to find out the possible triggering events or conditions, the running state of the equipment during the abnormal signal period is analyzed, whether the equipment has faults or abnormal operation is checked, whether the process parameter setting is reasonable is checked, whether it needs to be adjusted and optimized is checked;

[0127] According to the analysis results, the operation parameters such as aeration amount, reagent dosage, sedimentation time, sludge return flow are optimized, the specific measures include adjusting the operation parameters of aeration equipment, optimizing the aeration amount and time, ensuring the microbial activity, optimizing the chemical dosing, adjusting the reagent dosage and type, ensuring the best treatment effect; Monitor and adjust the sludge return, according to the sludge concentration and settling performance, adjust the sludge return flow, ensure the appropriate sludge load, optimize the sludge return ratio, prevent the sludge load from being too high or too low to affect the treatment effect;

[0128] According to the characteristics and treatment effect of sewage, the process flow is optimized to improve the treatment efficiency, the equipment with problems is maintained and upgraded to ensure the normal operation of the equipment, and the real-time monitoring and early warning system is strengthened to discover and handle potential problems in time.

[0129] Through the detailed analysis of the sewage situation during the time period when the sewage treatment abnormal signal is generated, the cause of the abnormal signal can be found out and corresponding improvement measures can be developed, the equipment operation parameters and process conditions can be optimized to ensure the efficiency and stability of the sewage treatment process and ensure that the effluent quality meets the standard.

[0130] It should be noted that the threshold information related in this embodiment is set by professionals in advance and will not be explained too much here.

[0131] The present application realizes efficient sewage treatment equipment operation parameter dynamic adjustment by installing multiple types of sensors at key positions of the dry bulk cargo terminal, using edge computing nodes for real-time data acquisition and preprocessing, combining historical characteristics and time characteristics, using a machine learning model for sewage type classification, and through weighted voting and confidence evaluation, comprehensive multiple classification results are obtained to ensure the best treatment effect, and through analysis of the actual sewage treatment situation, the equipment control is continuously optimized to prevent faults, prolong the service life of the equipment and improve the sewage treatment effect.

[0132] The sewage treatment equipment intelligent control method significantly improves the sewage treatment efficiency and effect, reduces the operation cost, enhances the stability and reliability of the equipment operation, realizes efficient, stable and intelligent management of sewage treatment through intelligent monitoring, data analysis and dynamic adjustment, and ensures that the effluent quality meets the standard and protects the environment.

[0133] The above formulas are all dimensionless values, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the latest real situation, and the preset parameters in the formula are set by the person skilled in the art according to the actual situation.

[0134] The above only describes some exemplary embodiments of the present application by way of illustration, and it is self-evident that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above figures and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present application.

[0135] It should be noted that in this paper, if there are relationship terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or equipment including the element.

[0136] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An intelligent control method for a dry bulk terminal sewage treatment plant, characterized in that, The method comprises the following steps; Install multiple types of sensors at different locations of the dry bulk cargo terminal and collect sewage data in real time, perform preliminary data preprocessing and analysis on the sewage treatment, and obtain the preliminary classified sewage types; Extract historical features and time features from the historical data, combine the extracted historical features and time features with the preprocessed data to form a complete feature vector for model input; The model classifies according to the input feature vector, outputs the prediction probability or classification label of each sewage type, combines the rule-based classification and machine learning classification results, combines the weighted voting and confidence evaluation to obtain the final classification result, and dynamically adjusts the operation parameters of the sewage treatment equipment according to the final classification result; Obtain the operation processing information generated after adjusting the operation parameters of the sewage treatment equipment, determine the sewage treatment situation after adjusting the operation parameters of the equipment according to the final classification result, and adjust and optimize the sewage equipment control according to the obtained processing result; Obtain the operation processing information generated after adjusting the operation parameters of the sewage treatment equipment, and the operation processing information includes equipment operation information and sedimentation influence information; The equipment operation information includes an operation equipment response stability index, and the sedimentation influence information includes a multi-dimensional sedimentation deviation index: The operation equipment response stability index and the multi-dimensional sedimentation deviation index are used to generate an equipment control coefficient; The operation equipment response stability index is inversely proportional to the equipment control coefficient, and the multi-dimensional sedimentation deviation index is proportional to the equipment control coefficient; The multi-dimensional sedimentation deviation index is used to represent the deviation of the sewage sedimentation effect in the sewage treatment process; The operation equipment response stability index is used to represent the stability and reliability of the sewage treatment equipment in the treatment process; The logic for obtaining the operation equipment response stability index is as follows: Obtain the number of equipment failures F in a time period and the total operation time T in the time period; By the installed sensors and monitoring devices, real-time monitoring of equipment adjustment time, operating parameters and fault occurrence, recording the start and end time of each adjustment operation, obtaining adjustment time data, periodically recording data sets of equipment operation, including adjustment time, operating parameters, fault occurrence time and frequency, obtaining actual adjustment time Adj act and expected adjustment time Adj exp , and calculating the standard deviation and average value of the adjustment time, calculating the adjustment time deviation, calculating the expression: Extracting the time point of parameter change and the amplitude ΔParam from the data set, recording the time point t when the parameter reaches the steady state stable and the time point t when the parameter changes change , calculating the parameter change response time: The logic for obtaining the multi-dimensional sedimentation deviation index is as follows: standard deviation σ of the adjustment times Adj and the mean value μ Adj a stability value of the adjustment times, calculated as: obtaining a standard deviation σ of the deviation of the operating parameter Par and the average value μ Par calculating a fluctuation value of the operating parameter: The computing device response stability index is calculated, the expression is as follows: Perform preliminary data preprocessing and analysis on the sewage treatment to obtain the preliminary classified sewage types, including: acquiring total suspended solid content TS in sewage, volume TJ of sludge after precipitation, and interface settling velocity SD in the data sample, and acquiring ideal total suspended solid concentration value TS re , ideal sludge volume index value TJ re , ideal interface settling velocity value SD re , calculating total suspended solid deviation value: establishing a set of total suspended solids bias values Set of sludge volume index values interface settling velocity value set The standardization is performed to obtain the total suspended solids deviation standard value: n represents the total number of data samples, max represents the maximum value function, min represents the minimum value function, and similarly, the standard value of the sludge volume index TJP and the standard value of the interface settling velocity SDP are calculated, the multi-dimensional settling deviation index is calculated, and the calculation expression is:

2. The intelligent control method of the dry bulk cargo wharf sewage treatment equipment according to claim 1, characterized in that, Connect the sensors to the edge computing node, set the data collection frequency and data format through the edge computing node, and use mean filtering to denoise the data; Standardize the data to convert the data to a unified unit and scale; Determine the preset water quality parameter threshold according to the real-time collected water quality parameters of the sewage, and use the water quality parameter threshold as the classification condition; According to the preset water quality parameter threshold, perform preliminary classification, traverse each collected data sample, compare each water quality parameter of each data sample with the water quality parameter threshold, and perform preliminary classification according to the comparison result to mark the classification result of each data sample. Extract historical features and time features from the historical data, combine the extracted historical features and time features with the preprocessed data to form a complete feature vector for model input, and the specific steps are as follows:

3. The intelligent control method of a dry bulk cargo wharf sewage treatment device according to claim 2, characterized in that, ​ According to the time characteristics, the change of the sewage characteristics in different time periods is determined, a time label is added to each data sample, and the data sample is divided into different time periods according to the collection time; The moving average and moving standard deviation in the time window in the historical data are calculated, and the historical feature extraction is performed according to the calculation result; After extracting and constructing the time characteristics and the historical characteristics, the preprocessed real-time data, the extracted time characteristics and the historical characteristics are combined and input into the model; The model includes random forest, support vector machine and neural network.

4. The intelligent control method of a dry bulk cargo wharf sewage treatment device according to claim 3, characterized in that, The final classification result is obtained by combining the weighted voting and the confidence evaluation of the rule-based classification and the machine learning classification result, and the specific steps are as follows: According to the preset water quality parameter threshold, the data samples are classified according to the rules, and one or more sewage types are obtained for each data sample according to the rule classification; For each data sample, the classification result of the machine learning model and the rule-based classification result are combined for weighted voting, and different weights are determined for each classification method according to the accuracy; The data samples are classified by using the machine learning model, the confidence evaluation of each category is obtained according to the rule classification, and the total confidence of each category is obtained. According to the weighted voting combined with the confidence evaluation method, the final sewage type of each data sample is determined, and the final classification result of each data sample is output.

5. The intelligent control method of a dry bulk cargo wharf sewage treatment device according to claim 4, characterized in that, According to the obtained processing result, the adjustment and optimization of the sewage equipment control are performed, and the specific steps include: The generated equipment control coefficient is compared with the intelligent control threshold value; If the equipment control coefficient is greater than the intelligent control threshold value, a sewage treatment abnormal signal is generated, and the sewage treatment control scheme is adjusted; If the equipment control coefficient is less than or equal to the intelligent control threshold value, a sewage treatment stable signal is generated, and the sewage treatment control is not required, and the monitoring and maintenance are continued.

Images