Intelligent sewage treatment method and system based on data fusion

Abstract

The invention discloses an intelligent sewage treatment method and system based on data fusion, and relates to the technical field of sewage treatment.The intelligent sewage treatment method and system have the advantages that compared with a traditional monitoring mode depending on a single index, key parameters such as pollutant types, sludge volume indexes and filamentous bacterium biomass are integrated, comprehensive pollution load indexes are generated in combination with a dynamic normalization algorithm, and the comprehensive pollution load indexes are calculated; the coupling effect of organic matter impact and filamentous bacterium activity can be recognized more accurately, misjudgment of a single parameter is avoided, regulation and control can be started at the initial stage of exceeding the standard of the filamentous bacterium activity coefficient through a preset threshold triggering response strategy, compared with a mode that traditional microscopic examination lags for 4-8 hours, the response speed is effectively increased, and the detection efficiency is improved. Meanwhile, floc structure recovery serves as a termination condition, carbon source waste caused by excessive adjustment can be avoided, and the system recovery period is shortened by 30%.

Description

Technical Field

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

[0002] Sludge bulking caused by excessive growth of filamentous bacteria often poses a challenge to the stability of biological wastewater treatment processes. Traditional pretreatment processes often utilize a combination of physical interception, chemical coagulation, and hydrolysis and acidification, with operating parameters typically set based on fixed design values ​​or empirical experience. In recent years, some new technologies have attempted to incorporate online water quality monitoring and activated sludge models for process control of wastewater treatment. However, these technologies still suffer from several significant drawbacks: First, pretreatment unit regulation exhibits significant hysteresis, requiring manual intervention to adjust hydraulic retention time, making it difficult to respond promptly to sudden, high-volume surges, such as those caused by heavy rain. Second, existing filamentous bacteria monitoring relies on offline microscopy, which lacks real-time feedback on microbial community changes, leading to inadequate control of coagulant dosage and dissolved oxygen in biosorption tanks. These technical shortcomings limit the system's ability to withstand shock loads, increasing the pressure on subsequent biochemical units and potentially reducing overall treatment efficiency. Therefore, there is an urgent need to develop a dynamic front-end pretreatment control system based on multi-source data fusion to improve the system's response to sudden changes in water quality. Summary of the Invention

[0003] In response to the defects in the existing technology, the present invention provides a sewage intelligent treatment method and system based on data fusion to solve the above technical problems.

[0004] On the one hand, it provides a smart sewage treatment method based on data fusion, including the following: Real-time collection of influent water quality parameters, including pollutant types, real-time dissolved oxygen concentration, sludge volume index, and instantaneous influent flow rate, as well as filamentous bacteria biomass index; The influent water quality parameters and filamentous bacteria biomass indicators were normalized, and the comprehensive pollution load index and filamentous bacteria activity coefficient were generated based on the normalized data; When it is detected that both the comprehensive pollution load index and the filamentous bacteria activity coefficient exceed the preset threshold, the shock load response strategy is activated; The activated sludge floc structure is monitored in real time, and the shock load response strategy is stopped when the floc diameter recovers to 150-300 μm.

[0005] Preferably, the comprehensive pollution load index is obtained by the following formula: ,

[0006] in, The time node is The comprehensive pollution load index at For the The contribution weight of each pollutant, is the total amount of pollutant types, For the The pollutants at the time point The normalized value of It is a regulatory factor affecting filamentous bacteria. is the exponential moving average smoothing coefficient, Time node the historical moving average of filamentous fungal biomass at the previous time point; The activity coefficient of filamentous fungi is specifically obtained by the following formula: ,

[0007] in, At the time node The activity coefficient of filamentous fungi is is the biomass weight of filamentous fungi, At the time node The normalized value of filamentous fungal biomass at is the dissolved oxygen weight, The time node is The real-time dissolved oxygen concentration at is the upper limit of the dissolved oxygen threshold, The time node is The sludge volume index at is the critical sludge volume index.

[0008] Preferably, the shock load response strategy specifically includes the following steps: Using a long-short-term memory network trained on historical data, the water inflow increment for the next two hours is predicted. The minimum volume required for the regulating pond is calculated based on the water inflow increment, and the regulating pond is used to temporarily store excess sewage based on the minimum volume calculation result. Bioinhibitors are added to the rapid mixing tank according to the activity coefficient of filamentous bacteria.

[0009] Generally, sewage treatment steps are carried out in a series of sequentially connected regulating tanks, rapid mixing tanks, primary sedimentation tanks, hydrolysis and acidification tanks, anaerobic tanks, biochemical tanks, and secondary sedimentation tanks. Based on the above steps, the surface load of the primary sedimentation tank is adjusted to 1.2-1.8 m³ / (m²·h) according to the water inflow increment to enhance the retention of suspended solids, and 10-50 mg / L of sodium hypochlorite is added to the rapid mixing tank to selectively inhibit filamentous bacteria; at the same time, the dissolved oxygen in the hydrolysis and acidification tank is controlled at 0.2-0.5 mg / L to inhibit the aerobic metabolism of filamentous bacteria.

[0010] Preferably, the water inflow increment in the next two hours is predicted by a long short-term memory network trained based on historical data, and the minimum volume required for the regulating tank to be occupied is calculated based on the water inflow increment, specifically including the following steps: Based on the predicted water inflow increment, a time series prediction algorithm is used to obtain the predicted instantaneous flow rate change rate; When the instantaneous flow rate change rate exceeds the change threshold, the volume priority occupancy algorithm is used to calculate the minimum volume required by the regulating tank.

[0011] The influent carbon-nitrogen ratio is further predicted based on the predicted influent increment prediction algorithm. When the predicted influent carbon-nitrogen ratio is greater than 8, the bypass pipe is opened to allow 20-30% of the influent to enter the anaerobic tank directly. Compared with the traditional fixed diversion mode, it can accurately match the denitrification carbon source demand of the anaerobic tank. By diverting 20-30% of the high C / N sewage in advance through the bypass pipe, the instantaneous load of the subsequent biochemical unit is reduced, avoiding the drastic fluctuation of DO caused by carbon-nitrogen imbalance in the traditional single process, and ensuring the stability of nitrifying bacteria activity.

[0012] Preferably, the volume priority occupancy algorithm is specifically shown in the following formula: ,

[0013] in, The minimum effective volume that the regulating tank needs to occupy, is the instantaneous maximum water inlet flow rate, is the time window, is the empirical coefficient, is the sludge volume index.

[0014] Preferably, the bioinhibitor is added according to a gradient addition strategy, wherein the gradient addition strategy is specifically as follows: When the activity coefficient of filamentous bacteria reaches the first gradient value, polyaluminium chloride is added; When the activity coefficient of filamentous bacteria reaches the second gradient value and the sludge volume index reaches the addition threshold, free nitrite and anthraquinone derivatives are added together; After addition, the changes in bacterial community structure were detected online by Raman spectroscopy.

[0015] When the first gradient value is 0.3 < filamentous bacteria activity coefficient ≤ 0.5, the dosage of polyaluminium chloride is 15 mg Al³+ / g MLSS. When the second gradient value is 0.5 < filamentous bacteria activity coefficient ≤ 0.7, the dosage threshold is sludge volume index > 180 mL / g. The combined dosage of free nitrite is 2 mg / L and the anthraquinone derivative is 0.5 mg / L. The dosage frequency is adjusted according to the changes in the online detection of bacterial community structure, and the sampling interval is 15 minutes. Compared with traditional empirical dosage (such as fixed dosage of 20 mg Al³+ / g MLSS), this gradient strategy can control the Al³+ residual content below 0.1 mg / L, avoiding the increase of chemical sludge. In addition, the second gradient combined addition of free nitrite and anthraquinone derivatives forms a dual mechanism of "membrane rupture and chain scission" through the cell membrane penetration of FNA and the electron transfer inhibition of AD, which can effectively improve the inactivation rate of filamentous bacteria while avoiding the imbalance of microbial community caused by excessive dosage.

[0016] Preferably, when the filamentous bacteria activity coefficient reaches the second gradient value and the sludge volume index reaches the addition threshold, the following operations are further performed: Transmit shock load warning signals to the back-end biochemical pool aeration control system; Raise the upper limit of dissolved oxygen threshold; Adjust the sludge return ratio of the secondary sedimentation tank.

[0017] The upper limit of the dissolved oxygen threshold is raised to 3.5-4.0 mg / L, and the sludge return ratio of the secondary sedimentation tank is adjusted to 80-100% to maintain the stability of the microbial community. Compared with the traditional DO control range, raising the DO threshold to 3.5-4.0 mg / L can significantly enhance the substrate competition advantage of aerobic bacteria over filamentous bacteria. At the same time, the high DO environment can promote the activity of nitrifying bacteria and improve the ammonia nitrogen removal rate. By maintaining a high sludge concentration in the biochemical pool and thus shortening the sludge age, it can effectively inhibit the excessive proliferation of filamentous bacteria. The synergistic effect of high DO and high return ratio can cope with the impact of influent COD fluctuations of ±30%, and avoid sedimentation failure caused by entanglement of filamentous bacteria.

[0018] Preferably, the activated sludge floc structure is monitored in real time by online optical coherence tomography, and parameters of the sludge floc structure are optimized at the same time. The parameter optimization steps are specifically as follows: Build a digital twin simulation model, run it synchronously in virtual space, and input water increment to generate floc structure evolution curve; When the real-time monitored activated sludge floc structure and the digital twin simulation model exceed the deviation threshold, the biological reaction kinetic parameters of the online optical coherence tomography are automatically corrected.

[0019] The deviation threshold here is generally set to a diameter error > 15%.

[0020] On the other hand, it provides a smart sewage treatment system based on data fusion, including the following: A multi-source sensor for collecting influent water quality parameters in real time, including pollutant types, real-time dissolved oxygen concentration, sludge volume index, and instantaneous influent flow rate; A microbial monitoring unit, wherein the microbial monitoring unit is used to collect filamentous fungal biomass indicators; a multi-source data processing module, which analyzes and processes the data collected by the multi-source sensors and the microbial monitoring unit to generate a comprehensive pollution load index and a filamentous bacteria activity coefficient; Floc structure collection unit, which is used to monitor the floc structure of activated sludge in real time The response strategy execution unit is used to start or stop the impact load response strategy according to the received comprehensive pollution load index, filamentous bacteria activity coefficient and activated sludge floc structure.

[0021] Preferably, the shock load response strategy specifically includes the following steps: predicting the water inflow increment in the next two hours through a long short-term memory network trained based on historical data, calculating the minimum volume required by the regulating tank based on the water inflow increment, and using the regulating tank to temporarily store excess sewage based on the minimum volume calculation result; Bioinhibitors are added to the rapid mixing tank according to the activity coefficient of filamentous bacteria.

[0022] The beneficial effects of the present invention are: compared with the traditional monitoring method that relies on a single indicator, this scheme integrates key parameters such as pollutant type, sludge volume index, filamentous fungus biomass, and combines it with a dynamic normalization algorithm to generate a comprehensive pollution load index, which can more accurately identify the coupling effect of organic matter impact and filamentous fungus activity, avoid misjudgment of a single parameter, and trigger a response strategy through a preset threshold. It can start regulation at the initial stage when the filamentous fungus activity coefficient exceeds the standard. Compared with the traditional microscopic inspection method with a lag of 4-8 hours, the response speed is effectively improved. At the same time, using floc structure recovery as the termination condition can avoid carbon source waste caused by excessive adjustment and shorten the system recovery cycle by 30%. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1A flow chart of the intelligent sewage treatment method based on data fusion provided by the present invention; Figure 2 This is a structural diagram of the sewage intelligent treatment system based on data fusion provided by the present invention;. DETAILED DESCRIPTION

[0025] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0026] The disclosure below provides many different embodiments or examples to realize different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention.

[0027] The embodiments of the invention are described in detail below with reference to the accompanying drawings.

[0028] like Figure 1 As shown in FIG, the intelligent sewage treatment method based on data fusion includes the following steps: Real-time collection of influent water quality parameters, including pollutant types, real-time dissolved oxygen concentration, sludge volume index, and instantaneous influent flow rate, as well as filamentous bacteria biomass index; The influent water quality parameters and filamentous bacteria biomass indicators were normalized, and the comprehensive pollution load index and filamentous bacteria activity coefficient were generated based on the normalized data; When it is detected that both the comprehensive pollution load index and the filamentous bacteria activity coefficient exceed the preset threshold, the shock load response strategy is activated; The activated sludge floc structure is monitored in real time, and the shock load response strategy is stopped when the floc diameter recovers to 150-300 μm.

[0029] Compared with the traditional monitoring method that relies on a single indicator, this solution integrates key parameters such as pollutant type, sludge volume index, filamentous fungal biomass, and combines it with a dynamic normalization algorithm to generate a comprehensive pollution load index. It can more accurately identify the coupling effect of organic matter impact and filamentous fungal activity, avoid misjudgment of a single parameter, and trigger the response strategy through a preset threshold. It can start regulation at the initial stage when the filamentous fungal activity coefficient exceeds the standard. Compared with the traditional microscopic inspection method with a lag of 4-8 hours, the response speed is effectively improved. At the same time, the recovery of floc structure is used as the termination condition to avoid the waste of carbon source caused by excessive regulation and shorten the system recovery cycle by 30%. In addition, the filamentous fungal biomass index is linked with dissolved oxygen and sludge index for analysis, and the carbon-nitrogen ratio of the hydrolysis acidification tank can be adjusted in a targeted manner. Compared with the fixed control method, the sludge settling performance has been greatly improved.

[0030] More specifically, the comprehensive pollution load index is obtained by the following formula: ,

[0031] in, The time node is The comprehensive pollution load index at For the The contribution weight of each pollutant, is the total amount of pollutant types, For the The pollutants at the time point The normalized value of It is a regulatory factor affecting filamentous bacteria. is the exponential moving average smoothing coefficient, Time node the historical moving average of filamentous fungal biomass at the previous time point; The activity coefficient of filamentous fungi is specifically obtained by the following formula: ,

[0032] in, At the time node The activity coefficient of filamentous fungi is is the biomass weight of filamentous fungi, At the time node The normalized value of filamentous fungal biomass at is the dissolved oxygen weight, The time node is The real-time dissolved oxygen concentration at is the upper limit of the dissolved oxygen threshold, The time node is The sludge volume index at is the critical sludge volume index.

[0033] More specifically, the shock load response strategy includes the following steps: Using a long-short-term memory network trained on historical data, the water inflow increment for the next two hours is predicted. The minimum volume required for the regulating pond is calculated based on the water inflow increment, and the regulating pond is used to temporarily store excess sewage based on the minimum volume calculation result. Bioinhibitors are added to the rapid mixing tank according to the activity coefficient of filamentous bacteria.

[0034] The use of a long-short-term memory network to predict the water inflow increment over the next two hours can reduce the error in the calculation of the equalization tank volume to less than 8% compared to the traditional prediction based on the moving average method. By dynamically allocating the minimum volume (avoiding the waste of tank volume caused by the traditional "full tank operation" mode), the buffering efficiency of the equalization tank is improved. At the same time, the dosage of bio-inhibitors is controlled based on the real-time filamentous bacteria activity coefficient. Compared with the fixed-dose addition method, it can effectively reduce the consumption of reagents, and the inhibition rate of filamentous bacteria is increased to more than 90%, significantly reducing the risk of accidental injury to nitrifying bacteria. Through the dual strategy of "physical buffering + biological inhibition", the peak of the water inlet shock load can be reduced by 40-60%, avoiding the high-load sewage directly entering the biochemical tank due to insufficient capacity of the traditional single equalization tank, and ensuring the stability of the subsequent treatment units.

[0035] More specifically, the long short-term memory network trained based on historical data is used to predict the water inflow increment in the next two hours. The minimum volume required for the regulating pond is calculated based on the water inflow increment. The following steps are specifically included: Based on the predicted water inflow increment, a time series prediction algorithm is used to obtain the predicted instantaneous flow rate change rate; When the instantaneous flow rate change rate exceeds the change threshold, the volume priority occupancy algorithm is used to calculate the minimum volume required by the regulating tank.

[0036] Compared with the traditional flow prediction method based on a fixed time window, predicting the instantaneous flow change rate through the long short-term memory network can effectively improve the recognition sensitivity of short-term impact loads such as flow surges caused by heavy rain, and avoid the prediction delay caused by the smoothing effect of the traditional method. The volume priority occupancy algorithm is further adopted. When the flow change rate exceeds the threshold, the core area of ​​the regulating pool is locked first. Compared with the traditional uniform distribution strategy, it effectively improves the buffering efficiency and reduces the overflow risk. At the same time, combined with the prediction results and the real-time pool capacity status, the minimum volume calculation granularity is dynamically adjusted, so that the utilization rate of the regulating pool is increased from 50-70% in the traditional mode to 85-90%, while reducing the frequent start and stop times of the water pump and extending the life of the equipment.

[0037] More specifically, the volume priority occupancy algorithm is shown in the following formula: ,

[0038] in, The minimum effective volume that the regulating tank needs to occupy, is the instantaneous maximum water inlet flow rate, is the time window, is the empirical coefficient, is the sludge volume index.

[0039] More specifically, the bioinhibitor is added according to a gradient addition strategy, wherein the gradient addition strategy is as follows: When the activity coefficient of filamentous bacteria reaches the first gradient value, polyaluminium chloride is added; When the filamentous bacteria activity coefficient reaches the second gradient value and the sludge volume index reaches the addition threshold, free nitrite and anthraquinone derivatives are added together; After addition, the changes in bacterial community structure were detected online by Raman spectroscopy.

[0040] Compared to the traditional single-drug dosing model, the gradient strategy achieves a dynamic match between inhibition intensity and pollution level through dual-threshold linkage control of the filamentous bacteria activity coefficient and the sludge volume index. Experiments have shown that this strategy reduces the total dosage of the drug by 20-40%, while stabilizing the proportion of filamentous bacteria below 15%, avoiding damage to nitrifying bacteria caused by excessive dosing during the low-activity period. The second gradient combines the addition of free nitrite and anthraquinone derivatives, destroying the filamentous bacteria cell membrane through FNA and inhibiting its electron transport chain through AD. The synergistic effect achieves an inactivation efficiency of over 95% for filamentous bacteria, effectively reducing sludge toxicity and minimizing the risk of secondary pollution. Finally, online Raman spectroscopy can identify changes in the proportion of filamentous bacteria in a shorter time than offline microscopy, dynamically modifying the dosing strategy and avoiding the drug accumulation problem caused by traditional blind adjustments.

[0041] More specifically, when the filamentous bacteria activity coefficient reaches the second gradient value and the sludge volume index reaches the addition threshold, the following operations are further performed: Transmit shock load warning signals to the back-end biochemical pool aeration control system; Raise the upper limit of dissolved oxygen threshold; Adjust the sludge return ratio of the secondary sedimentation tank.

[0042] Compared with the traditional single method that relies solely on front-end regulation, the coordinated operation of transmitting early warning signals to the back-end biochemical pool, increasing the DO threshold and dynamically adjusting the sludge return ratio can enhance the competitive inhibition of aerobic environment on filamentous bacteria. Increasing the sludge return ratio shortens the sludge age by maintaining a high sludge concentration and inhibits the excessive proliferation of filamentous bacteria. Under the synergistic effect of the two, the sludge settling performance can be restored from 200mL / g to 120mL / g within 4 hours. Compared with the traditional single DO regulation, the efficiency has been greatly improved.

[0043] More specifically, the activated sludge floc structure is monitored in real time by online optical coherence tomography, and parameters of the sludge floc structure are optimized. The parameter optimization steps are specifically as follows: Build a digital twin simulation model, run it synchronously in virtual space, and input water increment to generate floc structure evolution curve; When the real-time monitored activated sludge floc structure and the digital twin simulation model exceed the deviation threshold, the biological reaction kinetic parameters of the online optical coherence tomography are automatically corrected.

[0044] Online optical coherence tomography is used to monitor the floc structure in real time. Compared with traditional offline microscopes, it can capture millisecond-level floc morphological changes. Then, by constructing a digital twin model, the water inlet increment input into the digital twin model can predict the evolution of the floc structure and provide early warning of the risk of structural degradation. When the deviation between the OCT measured data and the twin model prediction exceeds the threshold, the biological reaction kinetic parameters are corrected in real time based on the Kalman filter algorithm, so that the model prediction accuracy is stable within ±5% for a long time, avoiding control failure caused by sludge characteristic drift.

[0045] like Figure 2 As shown in the figure, the intelligent sewage treatment system based on data fusion includes the following contents: A multi-source sensor for collecting influent water quality parameters in real time, including pollutant types, real-time dissolved oxygen concentration, sludge volume index, and instantaneous influent flow rate; A microbial monitoring unit, wherein the microbial monitoring unit is used to collect filamentous fungal biomass indicators; a multi-source data processing module, which analyzes and processes the data collected by the multi-source sensors and the microbial monitoring unit to generate a comprehensive pollution load index and a filamentous bacteria activity coefficient; Floc structure collection unit, which is used to monitor the floc structure of activated sludge in real time The response strategy execution unit is used to start or stop the impact load response strategy according to the received comprehensive pollution load index, filamentous bacteria activity coefficient and activated sludge floc structure.

[0046] More specifically, a long-short-term memory network trained based on historical data is used to predict the water inflow increment over the next two hours. The minimum volume required for the regulating pond is calculated based on the water inflow increment, and the regulating pond is used to temporarily store excess sewage based on the minimum volume calculation result. Bioinhibitors are added to the rapid mixing tank according to the activity coefficient of filamentous bacteria.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and description of the present invention.

Claims

1. The intelligent sewage treatment method based on data fusion is characterized by: The following steps are involved: Real-time collection of influent water quality parameters, including pollutant types, real-time dissolved oxygen concentration, sludge volume index, and instantaneous influent flow rate, as well as filamentous bacteria biomass index; The influent water quality parameters and filamentous bacteria biomass indicators were normalized, and the comprehensive pollution load index and filamentous bacteria activity coefficient were generated based on the normalized data; When it is detected that both the comprehensive pollution load index and the filamentous bacteria activity coefficient exceed the preset threshold, the shock load response strategy is activated; The activated sludge floc structure is monitored in real time, and the shock load response strategy is stopped when the floc diameter recovers to 150-300 μm.

2. The intelligent sewage treatment method based on data fusion according to claim 1 is characterized in that: The comprehensive pollution load index is calculated by the following formula: , in, The time node is The comprehensive pollution load index at For the The contribution weight of each pollutant, is the total amount of pollutant types, For the The pollutants at the time point The normalized value of It is a regulatory factor affecting filamentous bacteria. is the exponential moving average smoothing coefficient, Time node the historical moving average of filamentous fungal biomass at the previous time point; The activity coefficient of filamentous fungi is specifically obtained by the following formula: , in, At the time node The activity coefficient of filamentous fungi is is the biomass weight of filamentous fungi, At the time node The normalized value of filamentous fungal biomass at is the dissolved oxygen weight, The time node is The real-time dissolved oxygen concentration at is the upper limit of the dissolved oxygen threshold, The time node is The sludge volume index at is the critical sludge volume index.

3. The intelligent sewage treatment method based on data fusion according to claim 1 is characterized in that: The shock load response strategy specifically includes the following steps: Using a long-short-term memory network trained on historical data, the water inflow increment for the next two hours is predicted. The minimum volume required for the regulating pond is calculated based on the water inflow increment, and the regulating pond is used to temporarily store excess sewage based on the minimum volume calculation result. Bioinhibitors are added to the rapid mixing tank according to the activity coefficient of filamentous bacteria.

4. The intelligent sewage treatment method based on data fusion according to claim 3 is characterized in that: Using a long-term short-term memory network trained based on historical data, the water inflow increment in the next two hours is predicted. The minimum volume required for the regulating pond is calculated based on the water inflow increment. The specific steps include: Based on the predicted water inflow increment, a time series prediction algorithm is used to obtain the predicted instantaneous flow rate change rate; When the instantaneous flow rate change rate exceeds the change threshold, the volume priority occupancy algorithm is used to calculate the minimum volume required by the regulating tank.

5. The intelligent sewage treatment method based on data fusion according to claim 4 is characterized in that: The volume priority occupancy algorithm is specifically shown in the following formula: , in, The minimum effective volume that the regulating tank needs to occupy, is the instantaneous maximum water inlet flow rate, is the time window, is the empirical coefficient, is the sludge volume index.

6. The intelligent sewage treatment method based on data fusion according to claim 3 is characterized in that: The bioinhibitor was added according to a gradient addition strategy, wherein the gradient addition strategy is as follows: When the activity coefficient of filamentous bacteria reaches the first gradient value, polyaluminium chloride is added; When the activity coefficient of filamentous bacteria reaches the second gradient value and the sludge volume index reaches the addition threshold, free nitrite and anthraquinone derivatives are added together; After addition, the changes in bacterial community structure were detected online by Raman spectroscopy.

7. The intelligent sewage treatment method based on data fusion according to claim 6 is characterized in that: When the filamentous bacteria activity coefficient reaches the second gradient value and the sludge volume index reaches the addition threshold, the following operations are further performed: Transmit shock load warning signals to the back-end biochemical pool aeration control system; Raise the upper limit of dissolved oxygen threshold; Adjust the sludge return ratio of the secondary sedimentation tank.

8. The intelligent sewage treatment method based on data fusion according to claim 1 is characterized in that: The activated sludge floc structure is monitored in real time by online optical coherence tomography, and parameters of the sludge floc structure are optimized at the same time. The parameter optimization steps are specifically as follows: Build a digital twin simulation model, run it synchronously in virtual space, and input water increment to generate floc structure evolution curve; When the real-time monitored activated sludge floc structure and the digital twin simulation model exceed the deviation threshold, the biological reaction kinetic parameters of the online optical coherence tomography are automatically corrected.

9. The intelligent sewage treatment system based on data fusion is characterized by: Includes the following: A multi-source sensor for collecting influent water quality parameters in real time, including pollutant types, real-time dissolved oxygen concentration, sludge volume index, and instantaneous influent flow rate; A microbial monitoring unit, wherein the microbial monitoring unit is used to collect filamentous fungal biomass indicators; a multi-source data processing module, which analyzes and processes the data collected by the multi-source sensors and the microbial monitoring unit to generate a comprehensive pollution load index and a filamentous bacteria activity coefficient; Floc structure collection unit, which is used to monitor the floc structure of activated sludge in real time The response strategy execution unit is used to start or stop the impact load response strategy according to the received comprehensive pollution load index, filamentous bacteria activity coefficient and activated sludge floc structure.

10. The intelligent sewage treatment system based on data fusion according to claim 9 is characterized in that: The shock load response strategy specifically includes the following steps: Using a long-short-term memory network trained on historical data, the water inflow increment for the next two hours is predicted. The minimum volume required for the regulating pond is calculated based on the water inflow increment, and the regulating pond is used to temporarily store excess sewage based on the minimum volume calculation result. Bioinhibitors are added to the rapid mixing tank according to the activity coefficient of filamentous bacteria.

Images