Intelligent control method and control system for sewage treatment based on Internet of Things

Abstract

The invention relates to the technical field of sewage data processing, in particular to a sewage treatment intelligent control method and system based on the Internet of Things. According to the method, a source end calculation load sequence and a plant end actual measurement load sequence are compared to obtain a pipe network transportation attenuation coefficient. And obtaining an optimal time matching relationship by using an optimal distribution algorithm. And taking the ideal factory end load sequence as an ideal sequence, and carrying out deconvolution operation to obtain a load form feature vector representing a pipe network transportation dynamic response result in a rainfall period. And determining a transformation basis of a future rainfall period through a matching method to carry out transformation prediction, and adjusting a dissolved oxygen setting curve according to a predicted plant end load sequence. According to the method, three characteristics which respectively represent pollutant mass attenuation, main body transport delay and time form broadening are sequentially solved from coupled actual measurement signals, so that effective data prediction of a future rainfall period is realized, and a correct dissolved oxygen setting curve adjustment result is fed back based on a prediction result.

Description

Technical Field

[0001] This invention relates to the field of wastewater data processing technology, specifically to an intelligent control method and control system for wastewater treatment based on the Internet of Things. Background Technology

[0002] IoT-based wastewater treatment systems acquire process parameters through sensor networks deployed in pipe networks and treatment plants to achieve automated control. However, when dealing with infiltration events caused by rainfall, large amounts of rainwater rush into sewage pipes, causing complex transport effects on the original wastewater, which traditional control methods struggle to address effectively.

[0003] First, existing technologies typically simplify the impact of rainfall to a dilution of pollutant concentrations. Control systems reduce aeration accordingly upon detecting a decrease in influent chemical oxygen demand (COD) concentration. This understanding overlooks the mass loss of pollutants during transport within the sewer network. In combined sewer systems, excessive flow can trigger overflows, resulting in the direct discharge of untreated wastewater. This means that the total mass of pollutants actually reaching the treatment plant has already decreased. Existing control methods cannot distinguish between this mass loss and concentration dilution, leading to a lack of understanding of the true pollutant load in their control decisions.

[0004] Secondly, existing prediction models mostly focus on the hydraulic level, that is, predicting the time when the rainfall peak arrives at the treatment plant. However, the transport process of pollutants is not completely synchronized with the transport process of water. The transport velocity of the main pollutant load in the pipeline network may differ from the velocity of the water peak, causing the control commands based on flow prediction to have timing deviations and miss the adjustment window.

[0005] Furthermore, existing technologies completely ignore the morphological evolution of pollutant loads during transport. A concentrated pollution pulse generated at the source will become flattened and tailed after flowing through a complex pipeline network due to effects such as velocity gradients and turbulent mixing. Existing control systems predict future pollutant loads as invariant objects, resulting in a mismatch between the adjustment rate of their control response and the dynamic changes in the actual load, which can easily lead to system oscillations or energy waste. Summary of the Invention

[0006] To address the technical problems of existing technologies that fail to effectively consider the load decay caused by rainfall, the time asynchrony between pollutants and water transport, and the dynamic changes in pollutant load caused by complex pipeline networks when predicting pollutant information during rainfall periods, resulting in poor prediction results and the inability of the system to provide an effective dissolved oxygen setting curve based on the prediction results, the present invention aims to provide an intelligent control method and control system for wastewater treatment based on the Internet of Things. The specific technical solution adopted is as follows: This invention proposes an intelligent control method for wastewater treatment based on the Internet of Things, the method comprising: By using IoT devices in the historical sunny day database to count the influent flow and chemical oxygen demand concentration of the sewage treatment pipeline network, the average daily total oxygen demand load on sunny days can be obtained. For historical rainfall periods, based on the total average daily oxygen demand load on sunny days and the daily variation coefficient of standard source load at each time point, the source-end estimated load sequence is obtained; using the chemical oxygen demand concentration and water flow rate at the inlet of the sewage treatment plant collected by IoT devices, the plant-end measured load sequence is obtained; by comparing the source-end estimated load sequence and the plant-end measured load sequence, the pipeline transport attenuation coefficient is obtained. Based on the pipeline transport attenuation coefficient, the source end load sequence is calculated to obtain the quality-adjusted source end sequence; using the adjusted source end sequence as the supplier and the plant end measured load sequence as the demand side, the Hungarian algorithm is used to find the optimal time matching relationship based on the time cost matrix. Based on the optimal time matching relationship, the source end sequence is rearranged to obtain the ideal plant end load sequence; the ideal plant end load sequence is deconvolved to obtain the load shape feature vector; The pipeline transport attenuation coefficient, optimal time matching relationship, and load pattern feature vector are used as the pipeline transport feature set for each rainfall period. For future rainfall periods, the pipeline transport feature set of the matched historical rainfall periods is used as the basis for transformation to obtain the predicted plant load sequence for the future rainfall period. The dissolved oxygen setting curve is adjusted according to the predicted plant load sequence.

[0007] Furthermore, the method for obtaining the average daily total oxygen demand load includes: For each sunny day in the historical sunny day database, the cumulative value of the product of the influent flow rate and the chemical oxygen demand concentration at each moment is taken as the daily oxygen demand load for each sunny day; the average daily oxygen demand load of all sunny days is calculated to obtain the total average daily oxygen demand load for sunny days.

[0008] Furthermore, the method for obtaining the source-end estimated load sequence includes: The daily variation coefficient of the standard source load at each time point is normalized and used as the load generation ratio weight value at each time point. The load generation ratio weight value at each time point is multiplied by the total average daily oxygen demand load on sunny days to obtain the source load at each time point. The source load at all times constitutes the source load sequence.

[0009] Furthermore, the method for obtaining the measured load sequence at the plant end includes: The chemical oxygen demand (COD) concentration at the inlet of the wastewater treatment plant is multiplied by the water flow rate to obtain the measured load at the plant end at each time point. The measured loads at the plant end at all times constitute the measured load sequence at the plant end.

[0010] Furthermore, the method for obtaining the pipeline transport attenuation coefficient includes: The pipeline transport attenuation coefficient is obtained by summing the elements of the measured load sequence at the plant end as the numerator and summing the elements of the estimated load sequence at the source end as the denominator.

[0011] Furthermore, the method for obtaining the adjusted source sequence includes: Each element in the source-end estimated load sequence is multiplied by the pipeline transport attenuation coefficient to obtain the quality-adjusted source-end sequence.

[0012] Furthermore, the method for screening historical rainfall periods includes: Using total rainfall, peak rainfall intensity, and rainfall duration as rainfall characteristics, the Euclidean distance between the rainfall characteristics of future rainfall periods and each historical rainfall period is obtained. The historical rainfall period with the smallest Euclidean distance is selected as the matching historical rainfall period.

[0013] Furthermore, the method for obtaining the predicted plant load sequence includes: Each element in the future source-end estimated load sequence is multiplied by the pipeline transport attenuation coefficient in the matching pipeline transport feature set to obtain the predicted quality-adjusted source-end sequence; the predicted quality-adjusted source-end sequence is rearranged according to the optimal time matching relationship in the matching pipeline transport feature set to obtain the predicted ideal plant-end load sequence; the predicted ideal plant-end load sequence is convolved with the load shape feature vector in the matching pipeline transport feature set to obtain the predicted plant-end load sequence.

[0014] Furthermore, the adjustment of the dissolved oxygen setpoint curve based on the predicted plant load sequence includes: During the commissioning phase of the wastewater treatment system, the dissolved oxygen setpoint function is obtained by calibrating the DO value required to maintain the effluent quality meeting the standards under different load conditions and performing linear regression. The independent variable of the dissolved oxygen setpoint function is the load, and the dependent variable is the dissolved oxygen setpoint value. The predicted plant load sequence is substituted into the dissolved oxygen setpoint function to obtain the dissolved oxygen setpoint curve, which is then fed back to the control system for adjustment.

[0015] The present invention also proposes an intelligent control system for wastewater treatment based on the Internet of Things, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the steps of the intelligent control method for wastewater treatment based on the Internet of Things.

[0016] The present invention has the following beneficial effects: To analyze the specific impacts of rainfall periods, this invention first determines the total daily average oxygen demand (COD) load based on sunny day data. Then, based on effective information about rainfall periods, it determines the source-end estimated load sequence to characterize the theoretical COD load in the service area before entering the pipeline network. Further, based on actual plant measurement data, it obtains the plant-end measured load sequence. By comparing the source-end estimated load sequence and the plant-end measured load sequence, the pipeline transport attenuation coefficient can be obtained, thus taking into account the load attenuation caused by rainfall. This invention further uses the pipeline transport attenuation coefficient to inversely deduce the source-end estimated load sequence, obtaining the mathematical premise that the total supply and demand are equal in the optimal allocation algorithm. This determines the supply-demand relationship, and the optimal time matching relationship can be obtained using the optimal allocation algorithm. The optimal time matching relationship characterizes the synchronicity of pollutants and water during transportation. This invention further uses the measured load sequence at the plant end as an ideal sequence and convolves it with an unknown load shape broadening function. Therefore, by using the ideal plant end load sequence as the ideal sequence and performing deconvolution, a load shape feature vector characterizing the dynamic response of pipeline transportation during rainfall periods can be obtained. Then, all pipeline transportation feature sets are statistically analyzed, and the basis for future rainfall periods is determined through matching methods, enabling effective transformation prediction. Based on the predicted plant end load sequence, the dissolved oxygen setpoint curve is adjusted. This invention, through effective extrapolation of source and plant end load information during rainfall periods in historical databases, sequentially solves for three features characterizing pollutant mass decay, main transport delay, and time shape broadening from coupled measured signals, thereby achieving effective data prediction for future rainfall periods and providing correct dissolved oxygen setpoint curve adjustment results based on the prediction results. Attached Figure Description

[0017] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This invention provides an Internet of Things-based intelligent control method and control system for wastewater treatment, as one embodiment of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an IoT-based intelligent control method and control system for wastewater treatment proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] The following description, in conjunction with the accompanying drawings, details the specific scheme of the intelligent control method and control system for wastewater treatment based on the Internet of Things provided by this invention.

[0022] Please see Figure 1 The diagram illustrates a flowchart of an IoT-based intelligent control method for wastewater treatment, according to an embodiment of the present invention. The method includes: Step S1: Utilize IoT devices in the historical sunny day database to statistically analyze the influent flow rate and chemical oxygen demand (COD) concentration of the sewage treatment pipeline network, and obtain the average daily COD load for sunny days.

[0023] This invention aims to determine the pollutant transport characteristics of a pipeline network during a rainfall period by analyzing the coupling relationship between measured and inferred data. Therefore, it first requires collecting various data using IoT devices and storing them in a database for analysis. In this invention, a flow meter and an online chemical oxygen demand (COD) analyzer are installed at the main inlet of the wastewater treatment plant. Additionally, one or more rain gauges are installed within the area served by the pipeline network to collect daily rainfall. This invention sets data collection every 300 seconds, with all data collection frequencies being equal. This allows for the acquisition of influent flow rate sequences, COD concentration sequences, and rainfall sequences on a daily basis.

[0024] It should be noted that, due to the lack of historical data for analysis during the initial system deployment, it was impossible to directly determine what constituted a normal sunny day. Therefore, this embodiment of the invention analyzes rainfall sequences and filters out valid sunny days from the historical database construction phase. Specifically, the system automatically retrieves the collected data and filters out continuous 24-hour data periods that meet the following conditions: at all time points within this period and in the preceding 72 hours, all recorded rainfall intensities are below a preset minimum threshold. This invention sets the threshold to 0.1 mm / h. The filtered 24-hour data period is defined as a sunny day.

[0025] It should be noted that the historical database in this embodiment of the invention has an update function. Whenever new data for a new day is available, the corresponding data is added to the database. If the number of database samples is too large, data samples from more distant times can be removed. The specifics will not be elaborated here.

[0026] This invention requires a baseline parameter, independent of the pipeline transportation process, to characterize the total daily pollutant generation in the service area for calculating the source load during rainfall periods. Therefore, it is necessary to statistically analyze the load on sunny days in a historical sunny-day database for subsequent estimations. The load is positively correlated with the influent flow rate and chemical oxygen demand (COD) concentration; that is, a higher influent flow rate and higher COD indicate a higher load.

[0027] Preferably, in this embodiment of the invention, the method for obtaining the total daily average oxygen demand load includes: For each sunny day in the historical sunny day database, the cumulative value of the product of the influent flow rate and the chemical oxygen demand (COD) concentration at each moment is taken as the daily COD load for each sunny day. The average daily COD load of all sunny days is then calculated to obtain the total average daily COD load for sunny days. In other words, this embodiment of the invention uses a multiplication operation to construct a positive correlation, thereby quantifying the COD load at each sampling moment. It should be noted that the load data calculated in this embodiment of the invention is a quantified value, which can be considered to have only numerical values, ignoring its units.

[0028] Step S2: For historical rainfall periods, based on the total average daily oxygen demand load on sunny days and the daily variation coefficient of the standard source load at each time, obtain the source-end estimated load sequence; use the chemical oxygen demand concentration and water flow rate at the inlet of the sewage treatment plant collected by IoT devices to obtain the plant-end measured load sequence; compare the source-end estimated load sequence and the plant-end measured load sequence to obtain the pipeline transport attenuation coefficient.

[0029] In this embodiment of the invention, each rainfall day is further analyzed, i.e., the historical rainfall period in the historical database. Similar to the steps above, a continuous 24-hour data period that meets the following conditions is selected: at all time points in the period and within the preceding 72 hours, the rainfall intensity of all recorded data is greater than a preset minimum threshold. In this invention, the threshold is set to 0.1 mm / h. The selected 24-hour data period is defined as a rainfall day, i.e. a rainfall period.

[0030] For each historical rainfall period, this embodiment of the invention requires analyzing the data in the database to extract its pipeline transportation characteristics. The analysis method is the same for each historical rainfall period; this embodiment of the invention only uses one historical rainfall period as an example.

[0031] First, the total daily average oxygen demand load for sunny days has been obtained in step S1. This characteristic serves as the baseline data for analyzing load distribution. Therefore, this embodiment of the invention determines the standard source load daily variation coefficient value for each time point by consulting publicly available industry reference materials. The standard source load daily variation coefficient value represents the distribution weight of pollutant generation rates determined by social activities within a day. It is a system parameter determined as a prerequisite before analyzing specific rainfall events. This embodiment of the invention can take values ​​according to the relevant provisions in GB50014 and perform normalization processing. It is necessary to ensure that the sum of the normalization results for each time point is 1. If the number of results consulted is insufficient to cover all times within a rainfall day, a linear interpolation algorithm is needed to obtain the final results for each time point. Since the standard source load daily variation coefficient for each time point is a weight value, combining it with the total daily average oxygen demand load for sunny days yields the source load estimate for each time point, thus obtaining the source load estimate sequence.

[0032] Preferably, in this embodiment of the invention, the method for obtaining the source-end estimated load sequence includes: The daily variation coefficient of the standard source load at each time point is normalized and used as the load generation ratio weight value at each time point. The load generation ratio weight value at each time point is multiplied by the total average daily oxygen demand load on sunny days to obtain the source load at each time point. The source load at all times constitutes the source load sequence.

[0033] In this embodiment of the invention, the normalization method used can be the Softmax function for mapping, which is a technical means well known to those skilled in the art and will not be described in detail here.

[0034] This invention further integrates IoT devices to obtain the measured load sequence at the wastewater treatment plant inlet based on the chemical oxygen demand (COD) concentration and water flow rate. The source-end estimated load sequence represents the theoretical COD load generated in the area served by the wastewater treatment plant before entering the pipe network during a historical rainfall period. The plant-end measured load sequence represents the actual COD load reaching the wastewater treatment plant inlet after undergoing complete pipe network transport during the rainfall period. In other words, one represents the theoretical load generated at the required service end, and the other represents the actual load generated upon arrival at the plant.

[0035] Preferably, in this embodiment of the invention, the method for obtaining the measured load sequence at the plant end includes: The chemical oxygen demand (COD) concentration at the inlet of the wastewater treatment plant is multiplied by the water flow rate to obtain the measured load at the plant end at each time point. The measured loads at the plant end at all times constitute the measured load sequence at the plant end.

[0036] In combined sewer systems, the influx of rainfall runoff can lead to some wastewater being directly discharged through overflow outlets, resulting in a lower total mass of pollutants reaching the plant than generated at the source. Therefore, before analyzing time delays and morphological changes, this overall mass attenuation effect must first be quantified and isolated. This invention provides the sewer system transport attenuation coefficient by comparing the source-estimated load sequence with the plant-measured load sequence.

[0037] Preferably, in this embodiment of the invention, the method for obtaining the pipeline transport attenuation coefficient includes: The pipeline transport attenuation coefficient is obtained by summing the elements of the measured load sequence at the plant end as the numerator and summing the elements of the estimated load sequence at the source end as the denominator. It should be noted that, in reality, the sum of the elements of the estimated load sequence at the source end is always greater than the sum of the elements of the measured load sequence at the plant end, because rainfall leads to a decrease in the total mass of pollutants. Therefore, the obtained pipeline transport attenuation coefficient is a value between 0 and 1, which can intuitively reflect the proportion of mass loss of pollutants during pipeline transport.

[0038] Step S3: Based on the pipeline transport attenuation coefficient, back-calculate the source-end load sequence to obtain the quality-adjusted source-end sequence; using the adjusted source-end sequence as the supplier and the plant-end measured load sequence as the demand side, perform Hungarian algorithm optimization based on the time cost matrix to obtain the optimal time matching relationship.

[0039] Because the transport time of pollutants from dispersed sources to the centralized plant inlet is not a single constant value, but rather a distribution influenced by the hydraulic conditions of the pipeline network, this embodiment of the invention needs to further solve for the time mapping relationship describing the main load at the source to the plant, thereby decoupling the time delay characteristics in the transport process. In step S2 of this embodiment, a supply and demand relationship is constructed, and the Hungarian algorithm, an optimal allocation algorithm, is used to solve the problem. The Hungarian algorithm is a well-known technique for solving assignment problems. This algorithm is used to find the optimal path to allocate the source load, adjusted for quality decay, to the measured load at the plant in a way that minimizes the total time transport cost. Therefore, the supply and demand relationship needs to be determined first, where the demand side is obviously the measured load sequence at the plant. In order to satisfy the mathematical premise that the total supply and demand are equal in the optimal allocation algorithm, the source load sequence needs to be back-calculated using the pipeline transport decay coefficient to obtain the adjusted source sequence after quality adjustment. Then, using the adjusted source sequence as the supplier and the measured load sequence at the plant as the demand side, the Hungarian algorithm is used to find the optimal time matching relationship based on the time cost matrix. The time cost matrix is ​​a 24×24 matrix, where each row and column represents a time point. The elements of the matrix are the absolute values ​​of the differences between the indices of two time points; therefore, this time cost matrix is ​​a symmetric Toplitz matrix. The final output is the optimal path on this time cost matrix, representing the optimal time matching relationship and reflecting the time delay characteristics during transportation.

[0040] It should be noted that the process of using the Hungarian algorithm for data-driven inversion is a technique well-known to those skilled in the art. By constructing a simple time cost matrix, the direction of optimization is defined, and then the actual delay caused by the hydraulic conditions of the pipeline network is found by comparing the supply and demand relationship during the iterative process.

[0041] Preferably, in this embodiment of the invention, the method for obtaining the adjusted source sequence includes: Each element in the source-end estimated load sequence is multiplied by the pipeline transport attenuation coefficient to obtain the quality-adjusted source-end sequence.

[0042] Step S4: Based on the optimal time matching relationship, rearrange the source end sequence to obtain the ideal plant end load sequence; perform deconvolution on the ideal plant end load sequence to obtain the load shape feature vector.

[0043] As the load flows through the pipeline, its temporal shape broadens and tails due to velocity gradients and turbulent mixing. Therefore, after removing mass decay and the main time delay, this morphological change effect must be further quantified. This embodiment of the invention treats this process as a linear time-invariant system, convolving the measured load sequence at the plant end as an ideal sequence with an unknown load shape broadening function. The corresponding load shape feature vector is then obtained through deconvolution. Furthermore, considering the significant time delay between pollutants and water during transport, ideally, this time delay should be eliminated. This requires rearranging the adjusted source-end sequence based on the optimal time matching relationship to obtain an ideal plant-end load sequence, i.e., the ideal plant-end load sequence represents the ideal state plant-end result. Therefore, deconvolution is performed on the ideal plant-end load sequence to obtain the load shape feature vector. The deconvolution algorithm in this embodiment is the Tikhonov regularized deconvolution algorithm. The algorithm outputs a short time-series vector describing the morphological response of a unit load pulse after transport through the pipeline network.

[0044] It should be noted that sequence rearrangement and deconvolution algorithms are well-known techniques in the field and will not be elaborated upon here.

[0045] Step S5: Use the pipeline transport attenuation coefficient, optimal time matching relationship, and load pattern feature vector as the pipeline transport feature set for each rainfall period; for future rainfall periods, use the pipeline transport feature set of the matched historical rainfall periods as the basis for transformation to obtain the predicted plant load sequence for future rainfall periods, and adjust the dissolved oxygen setting curve according to the predicted plant load sequence.

[0046] Through the feature analysis described above, three transport characteristics of pollutants can be obtained for each historical rainfall period in the historical database. Therefore, the pipeline transport attenuation coefficient, optimal time matching relationship, and load pattern feature vector are used as the pipeline transport feature set for each rainfall period.

[0047] This invention focuses on predicting future rainfall periods and then using the prediction results to feed back effective control commands, thereby achieving proactive regulation of the processing flow. The control system's adjustments need to match the load disturbances in terms of amplitude, time, and rate of change; therefore, the prediction result should also be a time-based plant load sequence. Thus, this invention, for future rainfall periods, first obtains matching historical rainfall periods from the historical database. The pipeline transport characteristics of these matching historical rainfall periods can be used as a set of transformation operators. By transforming the source-end load, the final predicted plant load sequence can be predicted.

[0048] The predicted plant load sequence, presented as a time series, describes the COD quality entering the treatment plant during future rainfall periods. This needs to be further converted into executable control commands for the downstream aeration system. This invention adjusts the dissolved oxygen setpoint curve using the predicted plant load sequence. The final dissolved oxygen setpoint curve can then be sent to the treatment plant's central control system to adjust the aeration equipment in a feedforward manner. The shape of the dissolved oxygen setpoint curve corresponds to that of the predicted plant load sequence. Because the amplitude of the setpoint curve has been corrected for pipeline transport attenuation coefficients, over-aeration due to miscalculation of overflow losses is avoided. Because its temporal variation has been calibrated for optimal time matching, it ensures that the increase or decrease in aeration volume is synchronized with the actual arrival time of the load peak. More importantly, because its variation pattern has been shaped by the load pattern characteristic vector, the control system can predict whether the load arrives in a "peak-like" or "gradual" manner, thus adjusting the aeration rate to a more suitable level and avoiding system oscillations and energy waste caused by a mismatch between the control rate and the load change rate.

[0049] Preferably, in this embodiment of the invention, the method for screening historical rainfall periods includes: Total rainfall, peak rainfall intensity, and rainfall duration are used as rainfall features, which effectively characterize the uniqueness of rainfall periods. The Euclidean distance between future rainfall periods and each historical rainfall period is then obtained, and the historical rainfall period with the smallest Euclidean distance is selected as the matching historical rainfall period. It should be noted that the methods for obtaining the three rainfall features are all statistical methods well-known to those skilled in the art, and the algorithm for obtaining the Euclidean distance is also a prior art well-known to those skilled in the art, and will not be elaborated or limited here.

[0050] Preferably, in this embodiment of the invention, the method for obtaining the predicted plant load sequence includes: Each element in the future source-end estimated load sequence is multiplied by the pipeline transport attenuation coefficient in the matched pipeline transport feature set to obtain the predicted quality-adjusted source-end sequence. The predicted quality-adjusted source-end sequence is then rearranged according to the optimal time matching relationship in the matched pipeline transport feature set to obtain the predicted ideal plant-end load sequence. Finally, the predicted ideal plant-end load sequence is convolved with the load pattern feature vector in the matched pipeline transport feature set to obtain the predicted plant-end load sequence. It should be noted that the method for obtaining the future source-end estimated load sequence is the same as that for the source-end estimated load sequence in step S2; both are a single estimated result, and details will not be elaborated further.

[0051] Preferably, in this embodiment of the invention, adjusting the dissolved oxygen setpoint curve based on the predicted plant load sequence includes: During the commissioning phase of the wastewater treatment system, the dissolved oxygen (DO) setpoint function is obtained by calibrating the required DO value to maintain effluent quality standards under different load conditions and performing linear regression. The independent variable of the dissolved oxygen setpoint function is the load, and the dependent variable is the dissolved oxygen setpoint value. The predicted plant load sequence is substituted into the dissolved oxygen setpoint function to obtain the dissolved oxygen setpoint curve, which is then fed back to the control system for adjustment. In other words, this embodiment of the invention uses linear regression to more effectively correlate the obtained predicted plant load sequence with the final dissolved oxygen setpoint function in terms of function form and time, ensuring effective control adjustment.

[0052] In summary, in this embodiment of the invention, the total daily average oxygen demand load is determined based on sunny day data. The source-end estimated load sequence is determined based on effective information during rainfall periods, and the plant-end measured load sequence is obtained based on actual plant data. The pipeline transport attenuation coefficient is obtained by comparing the source-end estimated load sequence and the plant-end measured load sequence. The source-end estimated load sequence is then inferred from the pipeline transport attenuation coefficient, and the optimal time matching relationship is obtained using an optimal allocation algorithm. The ideal plant-end load sequence is used as the ideal sequence, and deconvolution is performed to obtain the load pattern feature vector characterizing the dynamic response of pipeline transport during rainfall periods. All pipeline transport feature sets are statistically analyzed, and the basis for future rainfall period changes is determined through matching methods, enabling effective change prediction. The dissolved oxygen setpoint curve is adjusted based on the predicted plant-end load sequence. This invention effectively extrapolates the source and plant load information during rainfall periods in historical databases, and sequentially solves three features from coupled measured signals to characterize pollutant mass decay, main transport delay, and temporal broadening, thereby achieving effective data prediction for future rainfall periods. Based on the prediction results, it provides feedback on the correct dissolved oxygen setpoint curve adjustment results.

[0053] Based on the same invention, this embodiment of the invention also proposes an Internet of Things-based intelligent control system for wastewater treatment, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the steps of the Internet of Things-based intelligent control method for wastewater treatment.

[0054] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0055] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A smart control method for wastewater treatment based on the Internet of Things, characterized in that, The method includes: By using IoT devices in the historical sunny day database to count the influent flow and chemical oxygen demand concentration of the sewage treatment pipeline network, the average daily total oxygen demand load on sunny days can be obtained. For historical rainfall periods, based on the total average daily oxygen demand load on sunny days and the daily variation coefficient of standard source load at each time point, the source-end estimated load sequence is obtained; using the chemical oxygen demand concentration and water flow rate at the inlet of the sewage treatment plant collected by IoT devices, the plant-end measured load sequence is obtained; by comparing the source-end estimated load sequence and the plant-end measured load sequence, the pipeline transport attenuation coefficient is obtained. Based on the pipeline transport attenuation coefficient, the source end load sequence is calculated to obtain the quality-adjusted source end sequence; using the adjusted source end sequence as the supplier and the plant end measured load sequence as the demand side, the Hungarian algorithm is used to find the optimal time matching relationship based on the time cost matrix. Based on the optimal time matching relationship, the source end sequence is rearranged to obtain the ideal plant end load sequence; the ideal plant end load sequence is deconvolved to obtain the load shape feature vector; The pipeline transport attenuation coefficient, optimal time matching relationship, and load pattern feature vector are used as the pipeline transport feature set for each rainfall period. For future rainfall periods, the pipeline transport feature set of the matched historical rainfall periods is used as the basis for transformation to obtain the predicted plant load sequence for the future rainfall period. The dissolved oxygen setting curve is adjusted according to the predicted plant load sequence.

2. The intelligent control method for wastewater treatment based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the total daily average oxygen demand load includes: For each sunny day in the historical sunny day database, the cumulative value of the product of the influent flow rate and the chemical oxygen demand concentration at each moment is taken as the daily oxygen demand load for each sunny day; the average daily oxygen demand load of all sunny days is calculated to obtain the total average daily oxygen demand load for sunny days.

3. The intelligent control method for wastewater treatment based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the source-end estimated load sequence includes: The daily variation coefficient of the standard source load at each time point is normalized and used as the load generation ratio weight value at each time point. The load generation ratio weight value at each time point is multiplied by the total average daily oxygen demand load on sunny days to obtain the source load at each time point. The source load at all times constitutes the source load sequence.

4. The intelligent control method for wastewater treatment based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the measured load sequence at the plant includes: The chemical oxygen demand (COD) concentration at the inlet of the wastewater treatment plant is multiplied by the water flow rate to obtain the measured load at the plant end at each time point. The measured loads at the plant end at all times constitute the measured load sequence at the plant end.

5. The intelligent control method for wastewater treatment based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the pipeline transport attenuation coefficient includes: The pipeline transport attenuation coefficient is obtained by summing the elements of the measured load sequence at the plant end as the numerator and summing the elements of the estimated load sequence at the source end as the denominator.

6. The intelligent control method for wastewater treatment based on the Internet of Things according to claim 5, characterized in that, The method for obtaining the adjusted source sequence includes: Each element in the source-end estimated load sequence is multiplied by the pipeline transport attenuation coefficient to obtain the quality-adjusted source-end sequence.

7. The intelligent control method for wastewater treatment based on the Internet of Things according to claim 1, characterized in that, The method for matching historical rainfall periods includes: Using total rainfall, peak rainfall intensity, and rainfall duration as rainfall characteristics, the Euclidean distance between the rainfall characteristics of future rainfall periods and each historical rainfall period is obtained. The historical rainfall period with the smallest Euclidean distance is selected as the matching historical rainfall period.

8. The intelligent control method for wastewater treatment based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the predicted plant load sequence includes: Each element in the future source-end estimated load sequence is multiplied by the pipeline transport attenuation coefficient in the matching pipeline transport feature set to obtain the predicted quality-adjusted source-end sequence; the predicted quality-adjusted source-end sequence is rearranged according to the optimal time matching relationship in the matching pipeline transport feature set to obtain the predicted ideal plant-end load sequence; the predicted ideal plant-end load sequence is convolved with the load shape feature vector in the matching pipeline transport feature set to obtain the predicted plant-end load sequence.

9. The intelligent control method for wastewater treatment based on the Internet of Things according to claim 1, characterized in that, The adjustment of the dissolved oxygen setpoint curve based on the predicted plant load sequence includes: During the commissioning phase of the wastewater treatment system, the dissolved oxygen setpoint function is obtained by calibrating the DO value required to maintain the effluent quality meeting the standards under different load conditions and performing linear regression. The independent variable of the dissolved oxygen setpoint function is the load, and the dependent variable is the dissolved oxygen setpoint value. The predicted plant load sequence is substituted into the dissolved oxygen setpoint function to obtain the dissolved oxygen setpoint curve, which is then fed back to the control system for adjustment.

10. An intelligent control system for wastewater treatment based on the Internet of Things, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the Internet of Things-based intelligent control method for wastewater treatment as described in any one of claims 1 to 9.

Images