Wastewater and waste liquid treatment self-adaptive adjustment method based on real-time monitoring feedback

Abstract

The invention discloses a wastewater and waste liquid treatment self-adaptive adjustment method based on real-time monitoring feedback, and relates to the technical field of wastewater treatment.The method comprises the steps that a pollutant monitoring index set sequence is obtained; determining a fusion pollutant monitoring index set; and the control system is used for carrying out adaptive regulation and control. The technical problems that an existing wastewater treatment scheme cannot construct an effective self-adaptive closed-loop system, depends on single-index analysis and cannot optimize treatment parameters in real time, so that the system is lack of flexibility and responsiveness are solved, and the purposes of constructing an effective'monitoring-feedback-regulation 'self-adaptive closed-loop system, realizing multi-index comprehensive analysis and improving the treatment efficiency are achieved. And the treatment parameters are optimized in real time, so that the technical effects of flexibility and responsiveness of the wastewater treatment system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of wastewater treatment, and particularly relates to an adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback. Background Art

[0002] In the process of wastewater and liquid waste treatment, some existing solutions have used sensors to monitor indicators such as pH, conductivity, COD, and BOD of influent and effluent. In a stable treatment scenario, these monitoring means can reflect some characteristics of wastewater and liquid waste to a certain extent. However, with the improvement of environmental protection requirements and the diversification of treatment scenarios, obvious defects have emerged in the existing technologies. Existing monitoring data are often only used for recording or alarming, and an effective adaptive closed-loop system has not been constructed, and treatment parameters cannot be optimized according to the real-time changes of wastewater and liquid waste. This results in the lack of flexibility and responsiveness of the treatment system, making it difficult to adapt to the complex and changeable wastewater and liquid waste treatment requirements, unable to efficiently and accurately treat wastewater and liquid waste, and difficult to meet the increasingly strict environmental protection standards. Summary of the Invention

[0003] This application solves the technical problem that existing wastewater treatment solutions fail to construct an effective adaptive closed-loop system, rely on single-index analysis, cannot optimize treatment parameters in real time, and thus lead to the lack of flexibility and responsiveness of the system. This application collects and removes impurities from laboratory wastewater and liquid waste, continuously extracts pollutant indicators using a sensor assembly to obtain a monitoring indicator sequence, determines a set of fused pollutant monitoring indicators through multi-index interaction and fusion, and based on this, feedback-optimizes the treatment adjustment parameters and inputs them into the control system to regulate the treatment equipment, realizing the real-time optimization of treatment parameters, enabling the wastewater and liquid waste treatment system to flexibly and quickly respond to changes in wastewater characteristics, making the wastewater and liquid waste treatment process more efficient and accurate, and meeting different wastewater treatment requirements.

[0004] In view of the above technical problems, this application proposes a technical solution for an adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback. Among them, the method includes: collecting organic waste liquid and inorganic waste liquid in the laboratory through a waste liquid collection device, decomposing the collected organic waste liquid using a high-temperature oxidation decomposition device, adsorbing heavy metals in the inorganic waste liquid using anion and cation resins, obtaining treated wastewater and removing impurities and suspended matters in a buffer sedimentation tank, continuously extracting pollutant indicators from the treated wastewater according to a preset set of pollutant indicators using a sensor assembly to obtain a sequence of pollutant monitoring indicator sets; performing interactive fusion on the sequence of pollutant monitoring indicator topologies in chronological order to determine a set of fused pollutant monitoring indicators; based on the set of fused pollutant monitoring indicators, feedback-optimize the treatment adjustment parameters to determine the target treatment adjustment parameters, input the target treatment adjustment parameters into the control system, and use the control system for adaptive regulation.

[0005] One or more technical solutions are proposed in this application, and at least the following technical effects are achieved:

[0006] In this application, by setting up a waste liquid collection device, a high-temperature oxidation decomposition device, anion and cation resins, a buffer sedimentation tank, a sensor assembly, etc. in the wastewater and waste liquid treatment process, impurities and suspended solids are removed in the buffer sedimentation tank, and the sensor assembly collects data according to a preset set of pollutant indicators to form a sequence of pollutant monitoring indicators. With the help of multi-index interaction and fusion of this sequence, a set of fused pollutant monitoring indicators is determined. Based on this set, the treatment adjustment parameters are feedback-optimized, and the target treatment adjustment parameters are input into the control system to achieve real-time adaptive regulation, complete the dynamic optimization of the wastewater treatment process, and achieve the technical effect of constructing an effective "monitoring - feedback - adjustment" adaptive closed-loop system, realizing comprehensive multi-index analysis, real-time optimization of treatment parameters, and thus improving the flexibility and responsiveness of the wastewater treatment system.

[0007] The above content outlines the self-adaptive adjustment method for wastewater and waste liquid treatment based on real-time monitoring and feedback in this application. The steps of the technical solution will be described in detail in the following specific embodiments to facilitate a clear and complete understanding of this application by those skilled in the art. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0009] Figure 1 FIG. is a schematic flow chart of the self-adaptive adjustment method for wastewater and waste liquid treatment based on real-time monitoring and feedback provided by an embodiment of this application.

[0010] Figure 2 FIG. is a schematic flow chart of determining a set of fused pollutant monitoring indicators in the self-adaptive adjustment method for wastewater and waste liquid treatment based on real-time monitoring and feedback provided by an embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0011] In this application, by setting up a waste liquid collection device, a high-temperature oxidation decomposition device, anion and cation resins, a buffer sedimentation tank, and a sensor assembly, impurities and suspended solids in the wastewater are removed in the buffer sedimentation tank. The sensor assembly obtains a sequence of pollutant monitoring index sets based on a preset set of pollutant indexes, and then through multi-index interaction and fusion, a fused pollutant monitoring index set is obtained. Based on this set, the processing adjustment parameters are feedback-optimized, and the optimized target processing adjustment parameters are input into the control system to achieve real-time regulation of the wastewater treatment equipment, thereby constructing an "monitoring - feedback - regulation" adaptive closed-loop system, achieving the technical effect of constructing an effective "monitoring - feedback - regulation" adaptive closed-loop system, realizing comprehensive multi-index analysis, real-time optimizing the processing parameters, and further enhancing the flexibility and responsiveness of the wastewater treatment system.

[0012] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present application.

[0013] It should be noted that any variations of the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products, or devices.

[0014] As Figure 1 shown, an adaptive adjustment method for wastewater and waste liquid treatment based on real-time monitoring and feedback, wherein the method includes:

[0015] Step A100: Collect organic waste liquid and inorganic waste liquid in the laboratory through a waste liquid collection device, decompose the collected organic waste liquid using a high-temperature oxidation decomposition device, adsorb heavy metals in the inorganic waste liquid using anion and cation resins to obtain treated wastewater, remove impurities and suspended solids from the wastewater in a buffer sedimentation tank, and continuously extract pollutant indexes of the treated wastewater using a sensor assembly according to a preset set of pollutant indexes to obtain a sequence of pollutant monitoring index sets.

[0016] In the embodiments of the present application, the waste liquid collection device is a device for collecting waste water and waste liquid in a laboratory, and is the starting link of the entire waste liquid treatment process. The buffer sedimentation tank is a facility for removing impurities and suspended solids after the waste liquid is collected. The preset pollutant index set includes solution pH, ion concentration, chemical oxygen demand, total dissolved solids, liquid level, etc. The pollutant monitoring index set sequence is a series of data sets obtained by continuously extracting pollutant indexes from the waste liquid after removing impurities and suspended solids.

[0017] Specifically, first, use the waste liquid collection device to collect the waste liquid generated in the laboratory. The collection method is to place the automatic collection buckets by category (organic waste liquid, heavy metal waste liquid, waste acid, etc.) at the waste liquid collection location of the instrument or beside the cleaning pool of the physical and chemical laboratory utensils, and classify and collect the waste liquid generated by the instrument in sequence.

[0018] For the collected organic waste liquid, use a high-temperature oxidation decomposition device for decomposition. The high-temperature oxidation decomposition device of the present invention includes a muffle furnace and an adsorption tank. An adsorption and filtration liquid for adsorbing and filtering harmful gases is filled in the adsorption tank. By installing a plurality of air flow baffles in the inner cavity of the muffle furnace, an air flow channel is provided between each air flow baffle and an inner side surface of the inner cavity of the muffle furnace. The air flow channels at every two adjacent air flow baffles are respectively located at the opposite inner side surfaces of the inner cavity of the muffle furnace. The muffle furnace is provided with an oxygen inlet, which can provide sufficient oxygen when the muffle furnace performs high-temperature oxidation decomposition on the organic waste liquid. The combustion exhaust gas is further treated by an exhaust gas purification device, and the high-temperature decomposition rate is as high as 99.99%. Moreover, the waste heat carried by the generated high-temperature gas can be recovered and reused through a heat exchange device, such as preheating the organic waste liquid to be treated, or providing partial heat energy for other small laboratory equipment that needs heating, reducing the total amount of organic waste liquid from the perspective of reduction, and improving the energy utilization rate from the perspective of resource reuse.

[0019] For inorganic waste liquid, use cation and anion resins to adsorb heavy metals therein. Imported high-performance cation and anion resins are used in the device. The heavy metals are adsorbed thoroughly, with a large adsorption capacity and an efficiency as high as 99.9%. The heavy metal concentration in the adsorbed waste liquid reaches a few tenths of PPM, far lower than the heavy metal discharge standard of urban sewage stipulated by the national standard. The adsorbed and recovered heavy metals can be further purified and then reapplied to industrial production, such as being used as raw materials in fields such as electronic component manufacturing and alloy production, realizing the reuse of heavy metal resources, achieving the dual goals of reduction and resource reuse, and improving the resource utilization degree.

[0020] The treated wastewater is transported to a buffer sedimentation tank, where impurities and suspended solids are removed through two methods: natural sedimentation and flocculant-assisted sedimentation. During natural sedimentation, under the action of gravity, large particle impurities and suspended solids gradually sink to the bottom of the tank, which is the initial separation achieved based on the density difference of substances. To enhance the sedimentation effect, flocculants are also added (in the field of wastewater treatment, common flocculants include inorganic flocculants (such as polyaluminum chloride, aluminum sulfate, ferrous sulfate, etc.), organic flocculants (such as polyacrylamide, etc.), and microbial flocculants). Flocculants can cause the tiny particles in the wastewater to aggregate into larger flocs, accelerating the sedimentation process. After being treated by these two methods, approximately 70%-80% of the large particle impurities and suspended solids can be removed from the wastewater.

[0021] Subsequently, a sensor assembly including a pH sensor, a conductivity sensor, a COD sensor, a TDS sensor, and a liquid level sensor is used. Each sensor measures the solution pH, ion concentration, chemical oxygen demand, total dissolved solids, and liquid level respectively according to its own detection principle. Among them, the pH sensor determines the pH by detecting the potential difference between the detection electrode and the solution; the conductivity sensor measures the ion concentration using the relationship between the solution conductivity and the ion concentration; the COD sensor measures the chemical oxygen demand by the chemical oxidation method; the TDS sensor obtains this index based on the correlation between the conductivity and the total dissolved solids; the liquid level sensor measures the liquid level through methods such as ultrasonic waves and static pressure. These sensors continuously collect data at time intervals of 5-10 minutes. The data of each index collected each time are combined into a set. As time goes by, multiple such sets are arranged in chronological order, and finally a set sequence of pollutant monitoring indicators is obtained.

[0022] By obtaining the set sequence of pollutant monitoring indicators, it provides an accurate data basis for subsequent treatment adjustment.

[0023] Step A200: Interactively fuse the topological network sequence of the pollutant monitoring indicators in chronological order to determine the fused pollutant monitoring indicator set.

[0024] In the embodiment of the present application, the topological network sequence of the pollutant monitoring indicators constructs a pollutant index topological network based on the correlation degree between preset pollutant index sets. The fused pollutant monitoring indicator set is the result obtained by interactively fusing the topological network sequence of the pollutant monitoring indicators in chronological order.

[0025] Optionally, first construct a pollutant index topology network based on the association degree between preset pollutant index sets, where each topology node corresponds to a pollutant index, and each edge represents the association degree between two pollutant indexes with an association relationship; then copy the topology network according to the number of sets in the pollutant monitoring index set sequence, and fill in data based on the sequence to obtain a sequence of pollutant monitoring index topology networks; finally, perform interactive fusion on the sequence of topology networks in chronological order to determine the fused pollutant monitoring index set, and the specific steps are described in detail in A210 - A230.

[0026] By determining the fused pollutant monitoring index set, it provides comprehensive and accurate data support for subsequent optimization of the processing adjustment parameters based on this set.

[0027] Step A300: Based on the fused pollutant monitoring index set, perform feedback optimization on the processing adjustment parameters to determine the target processing adjustment parameters, input the target processing adjustment parameters into the control system, and use the control system for adaptive regulation.

[0028] In the embodiment of the present application, the processing adjustment parameter is a variable that can be adjusted to optimize the processing effect during the wastewater and waste liquid treatment process. The target processing adjustment parameter is the parameter determined through comparative analysis of the fused pollutant monitoring index set and the target pollutant monitoring index threshold set and after feedback optimization for regulating the wastewater and waste liquid treatment equipment. The control system is a system that receives the target processing adjustment parameters and performs adaptive regulation on the wastewater and waste liquid treatment equipment.

[0029] In an embodiment of the present application, first obtain the target pollutant monitoring index threshold set, then calculate the deviation degree between the target set and the fused pollutant monitoring index set to obtain a set of pollutant monitoring index deviation values, and finally perform feedback optimization on the processing adjustment parameters based on this set of deviation values to obtain the target processing adjustment parameters. The specific steps are described in detail in A310 - A330.

[0030] Then, input the target processing adjustment parameters into the control system. After receiving the parameters, the control system starts adaptive regulation. Taking the chemical dosing pump and aeration equipment as examples, the control system will send a control signal to the chemical dosing pump according to the chemical dosing adjustment instruction in the target processing adjustment parameters. If it is necessary to increase the dosage of the oxidant, the control system will increase the working frequency or opening time of the chemical dosing pump, thereby increasing the dosage of the oxidant per unit time; otherwise, it will decrease the working frequency of the chemical dosing pump or shorten the opening time. For the aeration equipment, the control system changes the motor speed or valve opening of the aeration equipment according to the aeration volume adjustment instruction to adjust the aeration volume. Through the precise regulation of these devices, the wastewater and liquid waste treatment process can be optimized in real time according to the target processing adjustment parameters, and finally proceed efficiently in the direction of meeting the threshold of the target pollutant monitoring index.

[0031] Through the above steps, the precise control of the wastewater and liquid waste treatment process is realized, and the treatment effect and efficiency are improved.

[0032] Furthermore, step A200 in the method provided by the embodiment of the present application includes:

[0033] A210: Based on the association degree between the preset pollutant index sets, construct a pollutant index topology network, where each topology node in the pollutant index topology network corresponds to a pollutant index, and each edge is the association degree between two pollutant indexes with an association relationship.

[0034] A220: Copy the pollutant index topology network according to the number of sets in the pollutant monitoring index set sequence, and perform data filling based on the pollutant monitoring index set sequence to obtain a pollutant monitoring index topology network sequence.

[0035] A230: Perform interactive fusion on the pollutant monitoring index topology network sequence in chronological order to determine the fused pollutant monitoring index set.

[0036] In the embodiment of the present application, the pollutant index topology network is a network constructed based on the association degree between the preset pollutant index sets.

[0037] Specifically, first, a pollutant index topological network is constructed based on the correlation degree between the preset pollutant index sets. As shown in Table 1, there are various correlations between pH (acidity and alkalinity), conductivity, COD (chemical oxygen demand), and other indicators. For example, when adding acid / alkali for neutralization, the ion concentration increases, and the conductivity increases; too low or too high pH values may affect the reaction efficiency of certain organic substances, thereby affecting the COD removal rate, etc. Based on these relationships, each pollutant index is set as a topological node (such as pH, conductivity, COD, etc.), and the correlation degree between two pollutant indexes with a correlation relationship is used as an edge to build a pollutant index topological network. For example, if pH and conductivity are closely related, an edge reflecting their correlation degree is constructed between the nodes representing the two.

[0038] Next, according to the number of sets in the pollutant monitoring index set sequence, the pollutant index topological network is replicated, and data filling is performed based on the pollutant monitoring index set sequence. Assuming that there are n sets in the pollutant monitoring index set sequence, n pollutant index topological networks are replicated. Each topological network is filled with the pollutant monitoring index set data corresponding to the corresponding moment. For example, at a certain moment, data such as a pH value of 6.5, a conductivity of 200 μS / cm, and a COD of 80 mg / L are measured, and these data are filled into the corresponding nodes of the corresponding topological network to obtain a pollutant monitoring index topological network sequence.

[0039] Finally, the pollutant monitoring index topological network sequence is interactively fused in chronological order. First, a multi-scale anomaly recognition network layer set is pre-constructed, and it is used to extract features from the pollutant monitoring index topological network sequence to obtain a multi-scale pollutant monitoring index feature set; then this set is interactively fused to obtain fused multi-scale pollutant monitoring index features, and then these features are subjected to convolution analysis with the pollutant monitoring index topological network of the last one in the pollutant monitoring index topological network sequence, and finally a fused pollutant monitoring index set is obtained. The specific steps are described in detail in A231 - A232.

[0040] By comprehensively analyzing the index correlation relationships and data changes in the topological networks at different times, determining the fused pollutant monitoring index set can more comprehensively and accurately reflect the pollution status of the wastewater, provide a more reliable basis for optimizing the subsequent treatment adjustment parameters, and significantly improve the accuracy and effectiveness of wastewater treatment.

[0041] Table 1: Physical meaning, correlation, and typical correlation logic table of wastewater pollutant indicators.

[0042]

[0043]

[0044] Furthermore, as Figure 2As shown in the figure, step A230 in the method provided by the embodiments of the present application includes:

[0045] A231: Pre-construct a set of multi-scale anomaly recognition network layers, and use the set of multi-scale anomaly recognition network layers to extract features from the topological network sequence of pollutant monitoring indicators, obtaining a set of multi-scale pollutant monitoring indicator features.

[0046] A232: Perform interactive fusion on the set of multi-scale pollutant monitoring indicator features to obtain fused multi-scale pollutant monitoring indicator features, and perform convolution analysis on the fused multi-scale pollutant monitoring indicator features and the pollutant monitoring indicator topology at the last position in the topological network sequence of pollutant monitoring indicators, obtaining the set of fused pollutant monitoring indicators.

[0047] In the embodiments of the present application, the set of multi-scale anomaly recognition network layers is constructed by traversing a preset set of pollutant indicators to extract historical anomaly logs, analyzing the duration of abnormal data, etc., to obtain a set of screened abnormal data durations, and is a set used for extracting features from the topological network sequence of pollutant monitoring indicators. The fused multi-scale pollutant monitoring indicator features are obtained by randomly fusing any two features in the set of multi-scale pollutant monitoring indicator features multiple times to obtain an initial set, and then performing mean processing, and are the key intermediate results for determining the set of fused pollutant monitoring indicators.

[0048] Optionally, the process of pre-constructing the set of multi-scale anomaly recognition network layers is as follows: Traverse the preset set of pollutant indicators to extract historical anomaly logs, forming a set of log groups corresponding to each indicator; then extract a set of abnormal data duration groups, respectively find the maximum and minimum durations of each group and take the union to obtain a set of screened abnormal data durations, and finally construct the set of multi-scale anomaly recognition network layers based on this set. The specific steps are described in detail in A231-1 to A231-4.

[0049] After constructing the set of multi-scale anomaly recognition network layers, input the topological network sequence of pollutant monitoring indicators into this set. The set of multi-scale anomaly recognition network layers will analyze the data in the topological network sequence from different scales:

[0050] Step A: According to the preset convolutional neural network algorithm in the set of multi-scale anomaly recognition network layers (the specific steps are described in detail in A231-4), perform a preliminary scan on the topological network sequence. Use convolutional kernels of different sizes to simulate multi-scale perspectives from micro to macro. Smaller convolutional kernels focus on the local features of a single or a few nodes and their edges, such as short-term fluctuations of specific pollutant indicators and close associations with adjacent indicators; larger convolutional kernels focus on the comprehensive features of nodes and edges in a wider area, such as long-term trends and global associations between overall indicators.

[0051] Step B: During the process of scanning the topological network sequence, in combination with the Long Short-Term Memory (LSTM) algorithm, the forget gate will determine whether to retain or discard the information in the previous memory cell based on the current input and the previous state. If the abnormal change of a certain indicator lasts for a long time and is stable, the forget gate may retain the historical information related to it; if the change is unstable or no longer important, the forget gate will reduce the memory of it. The input gate is responsible for controlling the input of new information. When a new abnormal change of an indicator or a change in the correlation strength is detected, the input gate will integrate this information into the memory cell. The output gate outputs the prediction or feature representation at the current moment based on the content of the memory cell and the current input, which is used to capture the periodic pattern of the indicator. For example, when a certain pollutant indicator shows periodic high and low changes, the output gate can accurately reflect this pattern. Through the coordinated operation of the forget gate, input gate, and output gate, the LSTM algorithm effectively remembers the important historical information in the topological network sequence.

[0052] Step C: At the same time, the graph attention mechanism algorithm is adopted. First, the algorithm will traverse all the nodes in the topological network sequence and the associated edges between them to obtain the weight information of the degree of association represented by each edge. For those edges with higher weights and obvious changes in the association strength, the algorithm will assign more attention resources to the corresponding nodes. For example, if the weight of the associated edge between pH value and conductivity changes significantly within a certain period of time, the algorithm will focus more computing resources and attention on these two nodes.

[0053] When strengthening the extraction of key features, the algorithm will conduct in-depth analysis on the key nodes of interest and their surrounding local topological structures from multiple dimensions. First, it will analyze the attributes of the nodes themselves. For example, when focusing on the pH value node, it will analyze its numerical change trend, change rate, etc., to determine whether it is continuously rising, falling, or fluctuating. Then, it will study the association relationship between the node and its adjacent nodes. For example, if the pH value and conductivity nodes are closely related, it will analyze whether they are positively or negatively correlated and the stability of this association in different time periods.

[0054] At the same time, observe the overall form of the local topological structure, such as whether there are specific sub-structure patterns, whether it is a chain, ring, or other complex structures. Through comprehensive consideration of these aspects, the key information contained in them is mined. For example, certain indicator combinations frequently show abnormal fluctuations under specific structures, thereby discovering the laws hidden in the wastewater pollutant indicators and providing a more accurate basis for determining the set of integrated pollutant monitoring indicators in the follow-up.

[0055] Step D: Integrate the features obtained after these algorithm processes, and collect various features obtained after the initial scan by the convolutional neural network, the capture of time series features by the long short-term memory network, and the extraction of key features by the graph attention mechanism. Then, classify them according to the pollutant index type, time scale, and importance of the features corresponding to the features. Merge the features of the same type of index at different scales. For example, integrate the features of pH value at the microscopic and macroscopic scales together. For key features with a higher degree of importance, give higher weights. Finally, summarize all the integrated features to form a multi-scale pollutant monitoring index feature set.

[0056] The process of performing interactive fusion on the multi-scale pollutant monitoring index feature set to obtain the fused multi-scale pollutant monitoring index features is as follows: Randomly select any two features from the multi-scale pollutant monitoring index feature set multiple times for fusion to obtain the initial fused multi-scale pollutant monitoring index feature set, and then perform mean processing on this set to obtain the fused multi-scale pollutant monitoring index features. The specific steps are described in detail in A232-1 - A232-2.

[0057] Finally, perform convolutional analysis on the fused multi-scale pollutant monitoring index features and the pollutant monitoring index topology at the last position in the pollutant monitoring index topology sequence. When performing convolutional analysis, regard the fused multi-scale pollutant monitoring index features as a special "convolution kernel", and the pollutant monitoring index topology at the last position in the pollutant monitoring index topology sequence is the data matrix to be analyzed. Starting from the pollutant indicators and their associated relationships represented by each node and edge of the topology, slide the fused multi-scale pollutant monitoring index features point by point on this topology. At each slide, perform a multiplication operation on the fused features and the elements at the corresponding positions of the topology, and then add up all the multiplication results to obtain a new value. By traversing the entire topology and arranging these new values according to their positions, a new matrix is generated. This new matrix is the result of the convolutional analysis, that is, the fused pollutant monitoring index set.

[0058] Through the above steps, the pollution status of the wastewater is comprehensively and accurately reflected, providing key data support for subsequent optimization of treatment adjustment parameters based on this and achieving more efficient wastewater treatment.

[0059] Furthermore, step A231 in the method provided in the embodiment of the present application includes:

[0060] A231-1: Traverse the preset pollutant index set to extract historical wastewater monitoring anomaly logs, and obtain a set of historical wastewater monitoring anomaly log groups, where each historical wastewater monitoring anomaly log group corresponds to a pollutant index.

[0061] A231-2: Extract the duration of abnormal data from the set of historical wastewater monitoring abnormal log groups to obtain a set of abnormal data duration groups.

[0062] A231-3: Respectively extract the maximum abnormal data duration and the minimum abnormal data duration of each abnormal data duration group in the set of abnormal data duration groups, and perform a union operation on the extraction results to obtain a set of screened abnormal data durations.

[0063] A231-4: Construct a set of multi-scale abnormal recognition network layers based on the set of screened abnormal data durations.

[0064] Specifically, first, traverse the preset pollutant index set, which includes indicators such as solution pH, ion concentration, chemical oxygen demand, total dissolved solids, and liquid level. Taking a chemical laboratory as an example, a large amount of historical data on these indicators will be accumulated during long-term wastewater and waste liquid monitoring. By extracting historical wastewater monitoring abnormal logs, a set of historical wastewater monitoring abnormal log groups is obtained, and each group corresponds to a pollutant index. For example, in the recorded historical data, it is found that the chemical oxygen demand index is abnormal after a specific experimental operation. These abnormal data are grouped together to form an abnormal log group corresponding to chemical oxygen demand.

[0065] Next, extract the duration of abnormal data from these sets of abnormal logs. Suppose that within a certain period, the abnormal data of chemical oxygen demand lasted for 5 hours, and the abnormal data of conductivity lasted for 2 hours. In this way, a set of abnormal data duration groups is obtained. Then, respectively extract the maximum and minimum abnormal data durations of each abnormal data duration group, and perform a union operation on the extraction results. For example, the maximum duration in the chemical oxygen demand abnormal duration group is 5 hours, and the minimum duration is 1 hour; the maximum duration in the conductivity abnormal duration group is 2 hours, and the minimum duration is 0.5 hour. These durations are combined to obtain a set of screened abnormal data durations, that is, {0.5, 1, 2, 5}.

[0066] Finally, based on this set of durations for screening abnormal data, a set of multi-scale anomaly recognition network layers is constructed. Each duration in the set is used as a convolution scale. First, convolution kernels are designed separately for different convolution scales. Parameters such as the size and weight of the convolution kernel are adjusted according to the corresponding duration so that it can effectively capture the characteristics of pollutant indicators at different time scales. For example, for a short convolution scale of 0.5 hours, a small-sized convolution kernel is designed to focus on rapidly changing local data; for a long convolution scale of 5 hours, a large-sized convolution kernel is designed to cover a wider time range. Then, these convolution kernels of different scales are combined into a network layer structure and arranged in order from short scale to long scale to ensure that when analyzing the topological network sequence of pollutant monitoring indicators, each convolution kernel can work collaboratively from different time scales. During the operation of the network, each convolution kernel processes the topological network sequence data in parallel, integrates and outputs the features it extracts, thereby constructing a set of multi-scale anomaly recognition network layers that can comprehensively analyze the changes in pollutant indicators, capturing both short-term rapid changes and reflecting long-term trends and patterns.

[0067] Through such a set of network layers, the characteristic information in the topological network sequence of pollutant monitoring indicators can be extracted more comprehensively and deeply, providing strong support for subsequent determination of the integrated pollutant monitoring indicator set, thereby achieving more accurate and efficient wastewater treatment regulation.

[0068] Furthermore, step A232 in the method provided by the embodiments of the present application includes:

[0069] A232-1: Randomly and repeatedly perform feature fusion on any two multi-scale pollutant monitoring indicator features in the multi-scale pollutant monitoring indicator feature set to obtain an initial fused multi-scale pollutant monitoring indicator feature set.

[0070] A232-2: Perform mean processing on the initial fused multi-scale pollutant monitoring indicator feature set to obtain the fused multi-scale pollutant monitoring indicator features.

[0071] In the embodiments of the present application, the initial fused multi-scale pollutant monitoring indicator feature set is a set obtained by randomly and repeatedly performing feature fusion on any two multi-scale pollutant monitoring indicator features in the multi-scale pollutant monitoring indicator feature set.

[0072] Specifically, first, the multi-scale pollutant monitoring indicator feature set contains features obtained by analyzing wastewater pollutant indicators from different time and space scales. For example, through the processing of the previous set of multi-scale anomaly recognition network layers, a set covering different features such as short-term fluctuations and long-term trends has been obtained.

[0073] Next, randomly select any two multi-scale pollutant monitoring index features from this set multiple times for feature fusion. For example, select the feature reflecting the short-term change of chemical oxygen demand and the feature reflecting the long-term trend of conductivity, and use the feature fusion algorithm to integrate these two types of features. The algorithm will assign different weights to the short-term change feature of chemical oxygen demand and the long-term trend feature of conductivity according to the pre-set weight allocation rules. For example, if the current wastewater treatment process is more sensitive to the short-term change of chemical oxygen demand, the weight of the corresponding feature will be higher. Then, perform weighted summation or other specific fusion operations on the data of these two types of features according to the algorithm rules to organically combine the advantageous information of the two. For example, fuse the peak change of chemical oxygen demand within a specific short time with the information contained in the long-term stable upward trend of conductivity, so as to obtain a new feature that contains the advantageous information of the short-term change of chemical oxygen demand and the long-term trend of conductivity. In this process, multiple random selections are made to explore the possibilities of different feature combinations as comprehensively as possible. Suppose 100 different feature combinations are selected and fused, so as to obtain rich and diverse fusion results, forming an initial fusion multi-scale pollutant monitoring index feature set.

[0074] Then, perform mean processing on this initial set. Since there is a certain degree of discreteness in the feature fusion results in the initial set, mean processing can balance these differences and extract more representative features. Taking the fusion feature of chemical oxygen demand and conductivity as an example, in the initial set, the numerical distributions of the fusion features of the two under different combination methods are relatively scattered. Through mean processing, calculate the average value of all fusion features to obtain a comprehensive and more stable feature value, and finally obtain the fusion multi-scale pollutant monitoring index feature.

[0075] This feature can more accurately reflect the comprehensive relationship between pollutant indicators in the wastewater, provide more reliable data support for subsequent determination of the fusion pollutant monitoring index set and optimization of treatment adjustment parameters, enable the wastewater treatment system to more accurately adapt to the changes in wastewater characteristics, and improve the treatment efficiency and quality.

[0076] Furthermore, step A232-3 in the method provided by the embodiment of the present application includes:

[0077] A232-3A: Calculate the feature similarity of any two multi-scale pollutant monitoring index features to obtain a feature similarity set.

[0078] A232-3B: Perform normalization processing on the feature similarity set and construct an interaction fusion matrix based on the processing result.

[0079] A232-3C: Perform convolution operations on the interaction fusion matrix and the corresponding two multi-scale pollutant monitoring index features respectively to obtain two multi-scale pollutant monitoring index features.

[0080] A232-3D: Perform multiple random feature fusions on any two multi-scale pollutant monitoring index features in the multi-scale pollutant monitoring index feature set to obtain any two multi-scale pollutant monitoring index features in the multi-scale pollutant monitoring index feature set.

[0081] In the embodiments of the present application, the feature similarity is the value obtained by calculating the similarity degree between any two multi-scale pollutant monitoring index features, and these values constitute the feature similarity set. The interactive fusion matrix is a matrix constructed based on the processing result after normalizing the feature similarity set. The multi-scale pollutant monitoring index feature is an element in the set obtained after the multi-scale anomaly recognition network layer set extracts features from the pollutant monitoring index topology network sequence. The random feature fusion is an operation of performing multiple random combinations and fusions on any two multi-scale pollutant monitoring index features in the multi-scale pollutant monitoring index feature set.

[0082] Specifically, first, calculate the feature similarity of any two multi-scale pollutant monitoring index features to obtain the feature similarity set. Taking the short-term change feature of chemical oxygen demand and the long-term trend feature of conductivity as an example, the cosine similarity algorithm is used to measure the similarity degree between them. First, convert the short-term change feature of chemical oxygen demand and the long-term trend feature of conductivity into feature vectors respectively. Suppose the short-term change feature vector of chemical oxygen demand is A=(a1,a2,...,a n ), and the long-term trend feature vector of conductivity is B=(b1,b2,...,b n ), where n depends on the dimension of the feature. Then, according to the cosine similarity algorithm formula First, calculate the dot product A·B = a1b1 + a2b2 + … + a n b n , and then calculate the norms of vectors A and B respectively Finally, divide the dot product result by the product of the norms of the two vectors to obtain the cosine similarity value of the short-term change feature of chemical oxygen demand and the long-term trend feature of conductivity. Repeat this process for any two features in the multi-scale pollutant monitoring index feature set, and summarize all the obtained similarity values to obtain the feature similarity set.

[0083] Next, perform normalization to map the values in the feature similarity set to a specific interval, such as [0, 1]. This can eliminate the influence caused by different dimensions of different feature similarity values and make subsequent calculations more reasonable. Assume that after normalization, the similarity values are all within the range of [0, 1]. Then, construct an interaction fusion matrix based on these normalized similarity values. The elements of the matrix are determined according to the similarity of two features. The higher the similarity, the larger the value of the corresponding matrix element. For example, if the normalized similarity between feature A and feature B is 0.8, then the element at the corresponding position in the interaction fusion matrix is set to 0.8.

[0084] Then, perform convolution operations to obtain two multi-scale pollutant monitoring index features. The convolution operation is similar to a weighted summation process, where the elements of the interaction fusion matrix are used to weight the corresponding features. For example, perform convolution operations on the interaction fusion matrix with the short-term change feature vector of chemical oxygen demand and the long-term trend feature vector of conductivity. During the operation, the elements in the matrix are multiplied by the feature values at the corresponding positions, and then the product results are added together to obtain two new multi-scale pollutant monitoring index features. These two new features fuse the information of the original features and the correlation information between them.

[0085] Finally, randomly select two or more features from the multi-scale pollutant monitoring index feature set, including the new features obtained through convolution operations and other steps in the previous steps. Then, operate according to the weighted fusion algorithm:

[0086] Step E: First, determine each feature participating in the fusion and select features from the multi-scale pollutant monitoring index feature set. These features may include the new features obtained through various operations before.

[0087] Step F: Then, evaluate the importance or correlation of each feature. Taking the index features such as chemical oxygen demand, conductivity, and solution pH as examples, if in a specific wastewater treatment scenario, the chemical oxygen demand has a greater impact on the treatment process, then the importance of its corresponding feature is high, and a higher weight is correspondingly assigned; conversely, the feature with a smaller impact has a lower weight.

[0088] Step G: After determining the weights, multiply each feature by its corresponding weight. For example, if the weight of feature A is 0.5 and the feature value is 10, the product is 5; if the weight of feature B is 0.3 and the feature value is 8, the product is 2.4.

[0089] Step H: Finally, add up these weighted eigenvalues, 5 plus 2.4 and the values of other weighted features, thus completing the addition and fusion of features, obtaining a comprehensive feature result for more accurate analysis of the characteristics of wastewater in the subsequent steps. By continuously randomly selecting features and using algorithm fusion in this way, the value of different feature combinations can be more comprehensively explored, and more diverse multi-scale pollutant monitoring index features can be obtained.

[0090] The features obtained through the above steps can more comprehensively reflect the comprehensive situation of pollutants in wastewater, providing more accurate data support for subsequent wastewater treatment adjustment.

[0091] Further, step A300 in the method provided by the embodiments of the present application includes:

[0092] A310: Obtain a set of target pollutant monitoring index thresholds.

[0093] A320: Calculate the deviation degree between the set of target pollutant monitoring index thresholds and the set of fused pollutant monitoring indexes, and obtain a set of pollutant monitoring index deviation values.

[0094] A330: Based on the set of pollutant monitoring index deviation values, perform feedback optimization on the treatment adjustment parameters to obtain the target treatment adjustment parameters.

[0095] In the embodiments of the present application, the set of target pollutant monitoring index thresholds is determined based on the standard specifications of wastewater treatment and actual treatment requirements, and is a set of standard values of a series of pollutant indexes used to measure the effect of wastewater treatment.

[0096] In one embodiment, first, a set of target pollutant monitoring index thresholds needs to be obtained. This set is determined according to relevant environmental protection standards, industry specifications, and the requirements for the treated water quality in actual production. For example, according to national regulations, the upper limit of the chemical oxygen demand (COD) emission in the wastewater of this chemical industrial park is 100 mg / L, and the pH value needs to be controlled between 6 and 9. These standard values constitute a part of the set of target pollutant monitoring index thresholds.

[0097] Next, calculate the degree of deviation between the set of target pollutant monitoring index thresholds and the set of fused pollutant monitoring indices, and then obtain the set of pollutant monitoring index deviation values. In the actual processing, the wastewater is continuously monitored by the sensor assembly, and after a series of processes (the specific steps are described in A231 - A232), the set of fused pollutant monitoring indices is obtained. If, in a certain monitoring, the chemical oxygen demand in the set of fused pollutant monitoring indices is 150 mg / L, compared with the target threshold of 100 mg / L, the deviation value is 50 mg / L; the pH value is 5, compared with the lower limit of the target range of 6, the deviation value is -1. Summing up the deviation values of all pollutant indices forms the set of pollutant monitoring index deviation values.

[0098] Then, based on the set of pollutant monitoring index deviation values, use the neural network model to feedback optimize the treatment adjustment parameters to obtain the target treatment adjustment parameters:

[0099] Step I: Collect a large amount of data related to wastewater treatment, covering pollutant monitoring index data, corresponding treatment adjustment parameters, and final treatment effect data under different working conditions.

[0100] Step J: Clean and preprocess this data (data standardization and normalization operations), remove outliers and noise data, make data with different dimensions comparable, and ensure data quality.

[0101] Step K: Determine the structure of the neural network model. When constructing a neural network model for optimizing wastewater treatment adjustment parameters using a multi - layer perceptron (MLP), first set the number of hidden layers and neurons according to the data characteristics and problem complexity. It is determined to set 3 hidden layers. The first hidden layer is set with 64 neurons. This is because the set of input pollutant monitoring index deviation values includes various indices such as solution acidity - alkalinity and ion concentration, and 64 neurons can initially perform feature extraction and combination on this complex information; the second hidden layer is set with 32 neurons for further refining and integrating information; the third hidden layer is set with 16 neurons to focus on key features and prepare for the output.

[0102] Step L: Divide the data that has been cleaned, with outliers and noise removed, etc., into a training set, a validation set, and a test set in the ratio of 70%, 15%, and 15%. Use the training set to train the model, and continuously adjust the weights and biases of the model through the back - propagation algorithm. For example, in the initial stage of training, the difference between the treatment adjustment parameters predicted by the model and the actual parameters is large. The back - propagation algorithm will calculate the error between the predicted value and the true value, and adjust the weights and biases according to the error. For example, if the predicted dosage of chemical agents is significantly different from the actual demand, the algorithm will adjust the weights of the relevant connections to make subsequent predictions closer to the true value.

[0103] Step M: During the training process, use the validation set to evaluate the performance of the model and prevent overfitting. After each training iteration, input the validation set data into the model and observe the change in the error of the model on the validation set. If the error of the model on the training set continues to decrease, but the error on the validation set begins to increase, it indicates that overfitting may have occurred. At this time, adjust the training strategy in a timely manner, such as reducing the learning rate or increasing the regularization term.

[0104] Step N: After the training is completed, use the test set to test the model and verify the generalization ability of the model. Input the test set data into the trained model and calculate the error between the processing adjustment parameters output by the model and the corresponding true processing adjustment parameters in the test set, such as the mean square error (MSE) or the mean absolute error (MAE). If the error is within an acceptable range, it indicates that the model can effectively learn the complex relationship between the deviation values of the pollutant monitoring indicators and the processing adjustment parameters and has good generalization ability. Finally, a neural network model that can accurately predict and optimize the processing adjustment parameters according to the set of deviation values of the pollutant monitoring indicators is obtained, which is used to obtain the target processing adjustment parameters, thereby realizing the precise control of the wastewater treatment equipment.

[0105] Suppose that after the neural network model is trained, for every 10 mg / L increase in the deviation value of the chemical oxygen demand, the model recommends increasing the dosage of a specific chemical agent by 5 kg and extending the aeration time by 20 minutes; for a pH value deviating from the lower limit by 1, it is recommended to add an alkaline substance for neutralization, and the addition amount is calculated according to the model. According to the previously calculated chemical oxygen demand deviation value of 50 mg / L and the pH value deviation value of -1, the model will give a corresponding processing adjustment parameter adjustment plan, such as increasing the chemical agent dosage by 25 kg, extending the aeration time by 100 minutes, and adding an appropriate amount of alkaline substance. These adjusted parameters are the target processing adjustment parameters, which are input into the wastewater treatment equipment through the control system.

[0106] Through the above steps, the precise control of the treatment process is realized, thereby improving the wastewater treatment effect and making the treated wastewater more compliant with the discharge standards.

[0107] Furthermore, step A400 in the method provided in the embodiment of the present application includes:

[0108] A410: The sensor assembly includes a PH sensor, a conductivity sensor, a COD sensor, a TDS sensor, and a liquid level sensor.

[0109] Optionally, after the buffer settlement treatment, the wastewater enters the monitoring link. At this time, the PH sensor starts to work, and it determines the acidity and alkalinity of the solution by measuring the activity of hydrogen ions in the wastewater. For example, if the PH value of the wastewater is 4, it indicates that the wastewater is acidic, while the standard PH value for the discharge of general industrial wastewater is 6 - 9, which means that the acidity and alkalinity of the wastewater need to be adjusted.

[0110] The conductivity sensor is used to measure the conductivity of the wastewater, which reflects the ion concentration in the wastewater. Assuming that the measured conductivity is 5000 μS / cm, exceeding the normal range (the normal range is assumed to be 1000 - 3000 μS / cm), this means that the ion concentration in the wastewater is too high.

[0111] The COD sensor measures the chemical oxygen demand in the wastewater through a chemical method, and this indicator reflects the degree of pollution by reducing substances in the water. If the measured COD value is 200 mg / L while the discharge standard is 100 mg / L, it indicates that there are more reducing pollutants in the wastewater and stronger treatment is needed.

[0112] The TDS sensor is used to measure the total dissolved solids, which can comprehensively reflect the total amount of various inorganic and organic substances dissolved in the water. The liquid level sensor monitors the liquid level of the wastewater in real time, providing important parameters for the operation of subsequent treatment equipment. For example, when the liquid level reaches a certain height, the corresponding treatment process is started.

[0113] Furthermore, step A500 in the method provided by the embodiments of the present application includes:

[0114] A510: The preset pollutant index set includes solution pH, ion concentration, chemical oxygen demand, total dissolved solids, and liquid level.

[0115] Optionally, enter the pollutant index monitoring stage. For the solution pH, a PH sensor is used for measurement. The working principle of the PH sensor is based on the induction of hydrogen ion concentration in the solution, converting it into an electrical signal and outputting the corresponding PH value. For example, if the PH value of the wastewater is 4.5 while the discharge standard requires 6 - 9, this indicates that the solution pH deviates from the normal range.

[0116] The ion concentration is monitored by the conductivity sensor. The conductivity sensor indirectly reflects the ion concentration by measuring the ability of the solution to conduct current. Assuming that the conductivity of normal wastewater should be in the range of 1000 - 2000 μS / cm, when the monitored conductivity is 3000 μS / cm, it indicates that the ion concentration is too high.

[0117] The determination of chemical oxygen demand (COD) is carried out by the COD sensor using a specific chemical method. This indicator reflects the content of reducing substances in the water and is an important indicator for measuring the degree of wastewater pollution. For example, if the COD value in the wastewater is 250 mg / L while the industry discharge standard is 100 mg / L, this means that stronger treatment of this wastewater is needed.

[0118] The total dissolved solids are monitored by the TDS sensor, which comprehensively reflects the total amount of various inorganic and organic substances dissolved in the water.

[0119] The liquid level is monitored in real time by a liquid level sensor, providing key information for the operation of the processing equipment. For example, when the liquid level reaches a certain height, the subsequent processing process is automatically started.

[0120] During the entire processing process, these sensors continuously monitor the wastewater according to a preset set of pollutant indicators. Data is recorded at regular intervals, thus forming a sequence of pollutant monitoring indicator sets. These data provide a basis for subsequently constructing a pollutant indicator topology network and determining a set of integrated pollutant monitoring indicators.

[0121] Furthermore, step A600 in the method provided by the embodiments of the present application includes:

[0122] A610: Preset a regulation feedback window.

[0123] A620: Analyze the pollutant indicators of the wastewater and waste liquid within the preset regulation feedback window to determine whether the adjustment result meets the preset requirements. If not, generate a warning instruction, where the warning instruction is used to remind the staff that the wastewater and waste liquid need to continue to be processed and adjusted.

[0124] In the embodiments of the present application, the preset regulation feedback window is a time window preset in the adaptive adjustment method for wastewater and waste liquid treatment based on real-time monitoring feedback. The warning instruction is an instruction generated when, after analyzing the pollutant indicators of the wastewater and waste liquid within the preset regulation feedback window, it is found that the adjustment result does not meet the preset requirements.

[0125] Optionally, first, those skilled in the art preset the regulation feedback window according to past processing experience, the characteristics of the wastewater, and the treatment standards. Assuming that according to the treatment process and water quality fluctuations of the wastewater from a pharmaceutical factory, a regulation feedback window of every 2 hours is set. During this window time, the wastewater continues to be processed, and at the same time, the sensor assembly continuously collects the pollutant indicator data of the wastewater, and these data correspond to preset pollutant indicators such as the solution pH value, ion concentration, chemical oxygen demand, total dissolved solids, and liquid level.

[0126] At the end of each 2-hour regulation feedback window, the system analyzes the pollutant indicator data collected during this period. For example, according to the national and industry discharge standards for pharmaceutical wastewater, the discharge standard for chemical oxygen demand (COD) is 80 mg / L. If the average value of COD monitored within a certain regulation feedback window is 120 mg / L, this indicates that the treated wastewater does not meet the preset requirements; another example is that the standard range of the solution pH value is 6 - 9, while the actually monitored pH value is 4.5, which also shows that the treatment effect is not good.

[0127] Once it is determined that the adjustment result does not meet the preset requirements, the system will generate a warning instruction. This instruction will notify the staff in various ways, such as popping up a prominent reminder message on the display screen in the central control room and sending a text message reminder to the mobile phones of relevant staff. In this way, the staff is promptly informed that the wastewater and liquid waste need to continue to be treated and adjusted, so that the staff can quickly take corresponding measures according to the warning information, such as adjusting the operating parameters of the treatment equipment, adding specific treatment chemicals, etc., so as to ensure that the wastewater can finally meet the discharge standards, improve the quality and efficiency of wastewater treatment, and avoid environmental pollution problems caused by improper treatment.

[0128] In summary, the adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback provided by the embodiments of the present application has the following technical effects:

[0129] In this application, by setting a variety of sensor components in the wastewater and liquid waste treatment system, the wastewater is monitored according to the preset pollutant index set, and a sequence of pollutant monitoring index sets is obtained. After operations such as constructing a pollutant index topology network and a multi-scale anomaly recognition network layer, the integrated pollutant monitoring index set is determined and stored in a specific data set space for real-time update, and a processing strategy structure body is stored in the system database. By calculating the deviation degree between the target pollutant monitoring index threshold set and the integrated pollutant monitoring index set, combined with mechanisms such as data set management and processing strategy invocation, the processing adjustment parameters are optimized and the equipment is controlled based on the interaction of different data set intervals, ensuring the accurate realization of the adaptive adjustment of wastewater treatment, achieving the technical effect of constructing an effective "monitoring - feedback - adjustment" adaptive closed-loop system, realizing comprehensive analysis of multiple indicators, real-time optimization of processing parameters, and thus improving the flexibility and responsiveness of the wastewater treatment system.

[0130] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but rather will be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0131] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is also intended to include these changes and modifications.

Claims

1. An adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback, characterized in that, The method includes: Collecting organic waste liquid and inorganic waste liquid in the laboratory through a waste liquid collection device, decomposing the collected organic waste liquid by a high-temperature oxidation decomposition device, adsorbing heavy metals in the inorganic waste liquid by anion and cation resins to obtain treated wastewater, removing impurities and suspended substances in the wastewater in a buffer sedimentation tank, and continuously extracting pollutant index of the treated wastewater according to a preset pollutant index set by a sensor assembly to obtain a pollutant monitoring index set sequence; Performing interactive fusion on the pollutant monitoring index topology network sequence in chronological order to determine a fused pollutant monitoring index set; Based on the fused pollutant monitoring index set, performing feedback optimization on the treatment adjustment parameters to determine target treatment adjustment parameters, inputting the target treatment adjustment parameters into a control system, and using the control system for adaptive regulation.

2. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 1, characterized in that Performing multi-index interactive fusion on the pollutant monitoring index set sequence in chronological order to determine fused pollutant monitoring index features, including: Constructing a pollutant index topology network based on the correlation degree between the preset pollutant index sets, where each topology node in the pollutant index topology network corresponds to a pollutant index, and each edge is the correlation degree between two pollutant indexes with a correlation relationship; Copying the pollutant index topology network according to the number of sets in the pollutant monitoring index set sequence, and performing data filling based on the pollutant monitoring index set sequence to obtain a pollutant monitoring index topology network sequence; Performing interactive fusion on the pollutant monitoring index topology network sequence in chronological order to determine the fused pollutant monitoring index set.

3. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 2, characterized in that, Performing interactive fusion on the pollutant monitoring index topology network sequence in chronological order to determine the fused pollutant monitoring index set, including: Pre-constructing a multi-scale anomaly recognition network layer set, and using the multi-scale anomaly recognition network layer set to extract features from the pollutant monitoring index topology network sequence to obtain a multi-scale pollutant monitoring index feature set; Performing interactive fusion on the multi-scale pollutant monitoring index feature set to obtain a fused multi-scale pollutant monitoring index feature, and performing convolution analysis on the fused multi-scale pollutant monitoring index feature and the pollutant monitoring index topology network located at the last position in the pollutant monitoring index topology network sequence to obtain the fused pollutant monitoring index set.

4. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 3, characterized in that, Pre-constructing a multi-scale anomaly recognition network layer set, including: Traversing the preset pollutant index set to extract historical wastewater monitoring anomaly logs to obtain a set of historical wastewater monitoring anomaly log groups, where each historical wastewater monitoring anomaly log group corresponds to a pollutant index; Extracting the duration of abnormal data for the set of historical wastewater monitoring anomaly log groups to obtain a set of abnormal data duration groups; Respectively extracting the maximum abnormal data duration and the minimum abnormal data duration of each abnormal data duration group in the set of abnormal data duration groups, and performing a union operation on the extraction results to obtain a set of screened abnormal data durations; Construct a multi-scale anomaly recognition network layer set based on the set of screening abnormal data durations.

5. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 3, characterized in that, Perform interactive fusion on the multi-scale pollutant monitoring index feature set to obtain fused multi-scale pollutant monitoring index features, including: Randomly perform feature fusion on any two multi-scale pollutant monitoring index features in the multi-scale pollutant monitoring index feature set multiple times to obtain an initial fused multi-scale pollutant monitoring index feature set; Perform mean processing on the initial fused multi-scale pollutant monitoring index feature set to obtain the fused multi-scale pollutant monitoring index features.

6. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 5, characterized in that, Including: Calculate the feature similarity of any two multi-scale pollutant monitoring index features to obtain a feature similarity set; Perform normalization processing on the feature similarity set and construct an interactive fusion matrix based on the processing results; Perform convolution operations on the interactive fusion matrix and the corresponding two multi-scale pollutant monitoring index features respectively to obtain two multi-scale pollutant monitoring index features; Perform multiple random feature fusions on any two multi-scale pollutant monitoring index features in the multi-scale pollutant monitoring index feature set to obtain any two multi-scale pollutant monitoring index features in the multi-scale pollutant monitoring index feature set.

7. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 1, characterized in that, Based on the fused pollutant monitoring index set, perform feedback optimization on the processing adjustment parameters to determine the target processing adjustment parameters, including: Obtain a set of target pollutant monitoring index thresholds; Calculate the deviation degree between the set of target pollutant monitoring index thresholds and the fused pollutant monitoring index set to obtain a set of pollutant monitoring index deviation values; Perform feedback optimization on the processing adjustment parameters based on the set of pollutant monitoring index deviation values to obtain the target processing adjustment parameters.

8. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 1, wherein, The sensor assembly includes a pH sensor, a conductivity sensor, a COD sensor, a TDS sensor, and a liquid level sensor.

9. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 1, characterized in that, The preset pollutant index set includes solution acidity and alkalinity, ion concentration, chemical oxygen demand, total dissolved solids, and liquid level.

10. The adaptive adjustment method for wastewater and liquid waste treatment based on real-time monitoring feedback according to claim 1, wherein, Including: Preset a regulation feedback window; Analyze the pollutant indexes of the wastewater and waste liquid within the preset regulation feedback window to determine whether the adjustment result meets the preset requirements. If not, generate a warning instruction, where the warning instruction is used to remind the staff that the wastewater and waste liquid need to continue to be processed and adjusted.

Images