A wastewater treatment method and system based on ammonia nitrogen stripping

Abstract

This invention discloses a wastewater treatment method and system based on ammonia nitrogen stripping, belonging to the field of wastewater stripping treatment technology. The method includes: collecting wastewater ammonia nitrogen concentration, pH value, and temperature to generate a water quality monitoring data sequence; calculating the water quality change coefficient and correcting a preset effluent ammonia nitrogen concentration feedback control cycle to generate an adaptive feedback control cycle; optimizing stripping control parameters based on the water quality monitoring data sequence with the goal of maximizing ammonia nitrogen mass transfer, obtaining the optimal stripping control parameters, and calculating the corresponding stripping control time window; performing stripping control according to the optimal stripping control parameters within the stripping control time window, and iteratively optimizing the stripping control parameters based on the adaptive feedback control cycle and real-time effluent ammonia nitrogen concentration. This invention solves the technical problems of insufficient adaptability of control strategies, poor stripping operation stability, large fluctuations in effluent ammonia nitrogen concentration, and low treatment efficiency in existing wastewater ammonia nitrogen stripping treatment processes.

Description

Technical Field

[0001] This invention relates to the field of wastewater stripping treatment technology, specifically to a wastewater treatment method and system based on ammonia nitrogen stripping. Background Technology

[0002] With the rapid development of industries such as coal chemical, coking, landfill leachate treatment, pharmaceuticals, and fine chemicals, the amount of wastewater containing high concentrations of ammonia nitrogen is constantly increasing. Excessive ammonia nitrogen concentration not only affects the effectiveness of subsequent biological treatment but also easily causes environmental problems such as eutrophication of water bodies. Therefore, it is necessary to effectively remove ammonia nitrogen before wastewater discharge or further treatment. Currently, the stripping method is widely used in the treatment of ammonia nitrogen in various industrial wastewaters due to its advantages such as high treatment efficiency, suitability for high-concentration ammonia nitrogen wastewater, and relatively low operating costs.

[0003] In existing stripping processes, operating parameters such as fan operating frequency and influent flow rate are typically controlled based on operational experience or pre-set control strategies. The effluent ammonia nitrogen concentration is then monitored at fixed time intervals to assess the stripping effect. When the effluent ammonia nitrogen concentration changes, the relevant operating parameters are adjusted to maintain the normal operation of the stripping system.

[0004] However, in actual operation, wastewater quality is frequently affected by factors such as production conditions, raw material composition, and discharge processes, resulting in fluctuations in the stripping treatment effect. Simultaneously, the stripping process is also influenced by various factors including equipment operating status, environmental conditions, and process parameters, leading to significant differences in effluent ammonia nitrogen concentration at different operating stages. When operating conditions change, if control parameters are not adjusted promptly or their effectiveness is unstable, fluctuations in the stripping treatment effect can easily occur, affecting the continuous and stable operation of the system and potentially increasing operating costs.

[0005] Therefore, how to improve the operational stability of the wastewater ammonia nitrogen stripping treatment process, ensure that the effluent ammonia nitrogen concentration meets the treatment requirements, and at the same time take into account treatment efficiency and operating costs, has become a concern for those skilled in the art. Summary of the Invention

[0006] This invention provides a wastewater treatment method and system based on ammonia nitrogen stripping to solve the technical problems in the prior art, such as insufficient adaptability of control strategies, poor stability of stripping operation, large fluctuations in effluent ammonia nitrogen concentration, and low treatment efficiency during the ammonia nitrogen stripping treatment of wastewater.

[0007] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a wastewater treatment method based on ammonia nitrogen stripping, comprising: On the main inlet pipe before the wastewater enters the stripping tower, the ammonia nitrogen concentration, pH value and temperature of the wastewater are continuously collected at a preset sampling period to generate a water quality monitoring data sequence. The water quality change coefficient is calculated based on the water quality monitoring data sequence. The water quality change coefficient is used to correct the preset feedback control cycle of effluent ammonia nitrogen concentration and generate an appropriate feedback control cycle. Based on the water quality monitoring data sequence, the ammonia nitrogen stripping control parameters are optimized with the goal of maximizing the mass transfer quality of ammonia nitrogen. The optimal stripping control parameters are obtained, and the stripping control time window corresponding to the water quality monitoring data sequence is calculated by back-calculation using the optimal stripping control parameters. Within the stripping control time window, ammonia nitrogen stripping is controlled according to the optimal stripping control parameters, and the stripping control parameters are iteratively optimized based on the real-time effluent ammonia nitrogen concentration obtained from periodic monitoring, according to the adaptive feedback control cycle.

[0008] Secondly, the present invention provides a wastewater treatment system based on ammonia nitrogen stripping, comprising: The water quality monitoring module is used to continuously collect the ammonia nitrogen concentration, pH value and temperature of the wastewater on the main inlet pipe before the wastewater enters the stripping tower at a preset sampling period, and generate a water quality monitoring data sequence. The feedback cycle generation module is used to calculate the water quality change coefficient based on the water quality monitoring data sequence, use the water quality change coefficient to correct the preset feedback control cycle of effluent ammonia nitrogen concentration, and generate an adapted feedback control cycle. The stripping parameter optimization module is used to optimize the ammonia nitrogen stripping control parameters based on the water quality monitoring data sequence, with the goal of maximizing the mass transfer quality of ammonia nitrogen, to obtain the optimal stripping control parameters, and to back-calculate the stripping control time window corresponding to the water quality monitoring data sequence based on the optimal stripping control parameters. The stripping control module is used to control ammonia nitrogen stripping according to the optimal stripping control parameters within the stripping control time window, and to iteratively optimize the stripping control parameters based on the real-time effluent ammonia nitrogen concentration obtained by periodic monitoring according to the adaptive feedback regulation cycle.

[0009] One or more technical solutions provided in this invention have at least the following technical effects or advantages: This invention can improve the reliability of wastewater quality monitoring results and enhance the adaptability of the stripping control process to changes in operating conditions; it can improve the rationality of stripping control parameters and increase stripping treatment efficiency; it can continuously optimize the stripping control process according to the operating status, improve the stability of stripping operation, reduce fluctuations in effluent ammonia nitrogen concentration, and take into account system operating energy consumption, thereby improving the overall operating performance of wastewater ammonia nitrogen stripping treatment. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of a wastewater treatment method based on ammonia nitrogen stripping provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the closed-loop optimization principle of the stripping control provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a wastewater treatment system based on ammonia nitrogen stripping provided in an embodiment of the present invention.

[0011] In the attached diagram, the components represented by each number are as follows: Water quality monitoring module 11, feedback cycle generation module 12, stripping parameter optimization module 13, stripping control module 14. Detailed Implementation

[0012] This invention provides a wastewater treatment method and system based on ammonia nitrogen stripping, which specifically solves the technical problems in the prior art of insufficient adaptability of control strategies, poor stability of stripping operation, large fluctuations in effluent ammonia nitrogen concentration, and low treatment efficiency in the ammonia nitrogen stripping treatment process of wastewater.

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

[0014] Example 1, as Figure 1 and Figure 2 As shown, the present invention provides a wastewater treatment method based on ammonia nitrogen stripping, the method comprising: S100: On the main inlet pipe before the wastewater enters the stripping tower, the ammonia nitrogen concentration, pH value and temperature of the wastewater are continuously collected at a preset sampling period to generate a water quality monitoring data sequence. Step S100 in the method provided in this embodiment of the invention includes: The water quality monitoring data sequence includes wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence; The standard deviations of the wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence were calculated respectively to obtain the ammonia nitrogen concentration change index, pH value change index, and temperature change index. The water quality change coefficient is obtained by weighting the ammonia nitrogen concentration change index, pH value change index and temperature change index. The weight values ​​are adaptively determined based on the water quality characteristic state in the current water quality monitoring data sequence, and the sum of the weight values ​​of the three is 1.

[0015] In this step, the ammonia nitrogen concentration, pH value and temperature of the wastewater are continuously collected on the main inlet pipe before the wastewater enters the stripping tower at a preset sampling period to form a water quality monitoring data sequence within the current time window.

[0016] Specifically, an online ammonia nitrogen analyzer, an online pH analyzer, and an online temperature sensor are installed sequentially on the main inlet pipe of the stripping tower. The online ammonia nitrogen analyzer is used to detect the ammonia nitrogen concentration in the wastewater in real time and outputs the ammonia nitrogen concentration value; the online pH analyzer is used to detect the acidity or alkalinity of the wastewater in real time and outputs the pH value; and the online temperature sensor is used to detect the wastewater temperature in real time and outputs the temperature value. All three detection devices are communicatively connected to a PLC controller, which synchronously triggers sampling according to a unified clock. The ammonia nitrogen concentration, pH value, and temperature collected at the same time are combined into a set of water quality monitoring data to ensure that the three monitoring parameters have the same time reference.

[0017] In this embodiment, the sampling period is set to 30 seconds, and the statistical time window is set to 10 minutes. Therefore, a total of 20 sets of monitoring data are acquired within each statistical time window. A sliding window method is adopted, whereby the earliest set of data is deleted and the latest set of data is added after each sampling, thus ensuring that the statistical window always contains 20 sets of monitoring data.

[0018] The sampling period is determined based on the wastewater mixing and renewal time within the stripping tower and the stripping control response time. Testing of multiple stripping systems revealed that after aeration and mixing within the stripping tower, the wastewater's water quality parameters change significantly approximately every 20-40 seconds. When the sampling period is less than 20 seconds, the repetition rate of adjacent sampling data is high, increasing data processing volume while failing to improve the accuracy of water quality change identification. When the sampling period exceeds 40 seconds, short-term fluctuations in wastewater are easily missed, leading to lags in subsequent feedback and control. Therefore, this embodiment selects 30 seconds as the sampling period, ensuring real-time monitoring while also considering computational efficiency.

[0019] The statistical time window is determined based on the average hydraulic retention time of wastewater in the stripping tower. The average hydraulic retention time of wastewater in the stripping tower is typically 8–15 minutes. If the statistical time window is less than 5 minutes, the sample size is too small, and the calculated standard deviation is easily affected by random fluctuations. If the statistical time window is greater than 15 minutes, historical data accounts for too high a proportion and cannot reflect current changes in influent water quality in a timely manner. Therefore, this embodiment selects 10 minutes as the statistical time window.

[0020] For example, within a certain statistical time window, 20 sets of monitoring data were collected. The collected ammonia nitrogen concentrations in the wastewater were: 95, 98, 100, 102, 101, 97, 99, 103, 96, 100, 101, 98, 97, 99, 102, 104, 98, 100, 101, and 99 mg / L; the collected pH values ​​were: 10.42, 10.45, 10.47, 10.49, 10.50, 10.44, 10.46, 10.52, and 10.43. The temperatures collected were 10.48, 10.49, 10.45, 10.44, 10.46, 10.50, 10.53, 10.45, 10.47, 10.48, and 10.46; the wastewater temperatures were 31.2, 31.4, 31.6, 31.5, 31.7, 31.3, 31.4, 31.8, 31.2, 31.5, 31.6, 31.4, 31.3, 31.5, 31.7, 31.9, 31.4, 31.6, 31.5, and 31.4℃.

[0021] First, the average values ​​of the wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence were calculated separately. The average value of each sequence was obtained by dividing the sum of all sampled values ​​in the corresponding sequence by the number of samples.

[0022] Subsequently, in order to reflect the degree of fluctuation of various water quality parameters within the current statistical time window, the standard deviations of the wastewater ammonia nitrogen concentration series, wastewater pH value series, and wastewater temperature series were calculated respectively.

[0023] The ammonia nitrogen concentration change index is defined as the standard deviation of the wastewater ammonia nitrogen concentration sequence. It is calculated as follows: subtract the average ammonia nitrogen concentration from the wastewater ammonia nitrogen concentration obtained from each sampling to obtain the corresponding deviation value; sum the squares of each deviation value, divide by the number of samplings, and finally perform a square root operation on the result to obtain the ammonia nitrogen concentration change index.

[0024] Similarly, the pH change index and temperature change index were calculated.

[0025] For example, in this embodiment, the average ammonia nitrogen concentration is calculated to be (95+98+100+102+101+97+99+103+96+100+101+98+97+99+102+104+98+100+101+99)÷20=99.5mg / L. Similarly, the average pH value is calculated to be 10.47 and the average temperature value is calculated to be 31.50℃ using the same method. Subsequently, the standard deviations of the three monitoring parameters are calculated respectively. Among them, the ammonia nitrogen concentration change index is obtained by subtracting the average value from each group of ammonia nitrogen concentrations, squared each deviation, summing all the squared values, dividing by 20, and then taking the square root to obtain 2.29mg / L. Similarly, the pH value change index is calculated to be 0.029 and the temperature change index is 0.18℃ using the same method.

[0026] Since ammonia nitrogen concentration, pH value, and temperature have different dimensions, in order to eliminate the influence of dimensional differences on the subsequent comprehensive evaluation results, this embodiment further normalizes the above three change indices.

[0027] Specifically, each variation index is divided by the upper limit of the standard deviation reference for the corresponding parameter in the historical normal operation database to obtain the corresponding normalized variation index. The upper limit of the historical standard deviation reference is obtained by statistically analyzing historical monitoring data accumulated from more than six months of continuous and stable operation of the air-stripping system, and the 95th percentile of all historical standard deviation data is selected as the upper limit. Statistical analysis of the historical operation database shows that the 95th percentile can cover the vast majority of normal operating conditions while effectively eliminating occasional outliers; therefore, the 95th percentile is selected as the upper limit.

[0028] For example, in this embodiment, the historical database statistics show that the upper limit of the standard deviation of ammonia nitrogen concentration is 18 mg / L, the upper limit of the standard deviation of pH value is 0.60, and the upper limit of the standard deviation of temperature is 4.50℃.

[0029] Furthermore, in order to enable the water quality variation coefficient to more accurately reflect the influence of various water quality parameters on the stripping process under different operating conditions, this embodiment adopts an adaptive weight calculation method based on the current water quality characteristic state.

[0030] Specifically, the proportions of the normalized change index of ammonia nitrogen concentration, pH, and temperature to the sum of the three normalized change indices are first calculated, and these proportions are used as the contribution rates of each parameter. The larger the contribution rate, the more significant the fluctuation of the corresponding parameter within the current time window, and the greater its impact on the stripping operation.

[0031] For example, in this embodiment, the normalized change index of ammonia nitrogen concentration is calculated to be 2.29 ÷ 18 = 0.127, the normalized change index of pH is 0.029 ÷ 0.60 = 0.048, and the normalized change index of temperature is 0.18 ÷ 4.50 = 0.040. The sum of the normalized change indices is 0.127 + 0.048 + 0.040 = 0.215. Therefore, the contribution rate of ammonia nitrogen concentration change is 0.127 ÷ 0.215 = 0.590; the contribution rate of pH change is 0.048 ÷ 0.215 = 0.223; and the contribution rate of temperature change is 0.040 ÷ 0.215 = 0.187.

[0032] Subsequently, the corresponding weights are determined based on the contribution rate of the changes.

[0033] To ensure that all monitoring parameters can participate in the comprehensive evaluation, and to avoid excessive weight concentration caused by drastic short-term fluctuations in a certain parameter, this embodiment uses a combination of basic weight and dynamic weight to determine the final weight.

[0034] Each monitoring parameter was assigned a base weight of 0.20, derived from sensitivity analysis of historical operating samples. When the base weight was less than 0.10, a parameter was prone to complete failure; when the base weight was greater than 0.30, the dynamic adjustment capability decreased. After comprehensive analysis, 0.20 was selected. The remaining weight of 0.80 was dynamically allocated according to the contribution rate of each parameter's change. To ensure that the sum of the dynamic weights and the base weights still equaled 1, the initial weights were normalized again, ensuring that the sum of the ammonia nitrogen concentration weight, pH value weight, and temperature weight always equaled 1.

[0035] For example, in this embodiment, the initial weights for ammonia nitrogen concentration are calculated to be 0.20 + 0.80 × 0.590 = 0.672; the initial weight for pH is 0.20 + 0.80 × 0.223 = 0.378; and the initial weight for temperature is 0.20 + 0.80 × 0.187 = 0.350. The sum of the initial weights is 0.672 + 0.378 + 0.350 = 1.4. Therefore, the weight for ammonia nitrogen concentration is 0.672 ÷ 1.4 = 0.480, the weight for pH is 0.378 ÷ 1.4 = 0.270, and the weight for temperature is 0.350 ÷ 1.4 = 0.250.

[0036] Furthermore, to avoid frequent weight changes due to sampling noise, this embodiment sets the weight update threshold to 0.05. This threshold was obtained through sensitivity analysis of historical operating data. Experimental results show that when the threshold is less than 0.03, it is easily affected by instantaneous measurement noise; when the threshold is greater than 0.07, the response speed to actual water quality changes decreases. Therefore, selecting 0.05 as the weight update threshold can balance response speed and control stability. When the change amplitude of the weight corresponding to the same parameter within two consecutive statistical time windows is less than 0.05, the current change is considered to be a random measurement error, and the corresponding weight is not updated; when the change amplitude is not less than 0.05, the corresponding weight is updated.

[0037] Finally, the water quality change coefficient corresponding to the current statistical time window is obtained by weighting and summing the normalized change index of ammonia nitrogen concentration, the normalized change index of pH value, and the normalized change index of temperature with their respective weights.

[0038] For example, in this embodiment, the water quality change coefficient corresponding to the current statistical time window is 0.127×0.480+0.048×0.270+0.040×0.250=0.08392.

[0039] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0040] In summary, this step utilizes three indicators—ammonia nitrogen concentration, pH value, and temperature—to jointly characterize the current water quality fluctuation. It also dynamically reflects the importance of different parameters through adaptive weighting. Compared to using a single ammonia nitrogen concentration evaluation method, this approach more accurately characterizes changes in the stripping influent water quality, providing a reliable basis for the dynamic correction of subsequent feedback control cycles, thereby improving the stability of stripping control and the efficiency of ammonia nitrogen removal.

[0041] S200: Calculate the water quality change coefficient based on the water quality monitoring data sequence, and use the water quality change coefficient to correct the preset feedback control cycle of effluent ammonia nitrogen concentration, thereby generating an adaptive feedback control cycle. Step S200 in the method provided in this embodiment of the invention includes: The ratio of the average historical water quality change coefficient recorded within the historical time range to the water quality change coefficient is used as the periodic correction coefficient; The product of the cycle correction coefficient and the preset feedback control cycle for effluent ammonia nitrogen concentration is rounded down to obtain the adaptive feedback control cycle.

[0042] In this step, based on the water quality change coefficient corresponding to the current statistical time window calculated in step S100, the preset feedback control cycle of effluent ammonia nitrogen concentration is corrected to generate an adaptive feedback control cycle.

[0043] Specifically, to enable the stripping control to automatically adjust the feedback detection frequency based on the current fluctuations in wastewater quality, this embodiment first establishes a historical water quality change coefficient database. This database records the historical water quality change coefficients corresponding to each statistical time window during the continuous and stable operation of the stripping system.

[0044] This embodiment uses a sliding historical window method to update the average historical water quality change coefficient. After the current statistical time window is calculated, the current water quality change coefficient is added to the historical database, while the earliest recorded data from more than 30 days ago is deleted. Then, the average historical water quality change coefficient is recalculated so that the average historical water quality change coefficient always reflects the operating status of the most recent 30 days.

[0045] The historical time range is determined based on the operating cycle of the stripping system. Analysis of long-term operational data from multiple stripping systems revealed that when the historical time range is less than 15 days, the statistical sample size is small and easily affected by short-term changes in operating conditions; when the historical time range exceeds 60 days, the historical data includes many failed operating states, causing the statistical results to not accurately reflect the current equipment operating level. Therefore, this embodiment selects the most recent 30 days as the historical time range, ensuring statistical stability while also considering the timeliness of the operating status.

[0046] Furthermore, statistical calculations were performed on all historical water quality change coefficients within the historical time frame to obtain the average historical water quality change coefficient.

[0047] The average historical water quality change coefficient is calculated by dividing the sum of all historical water quality change coefficients within the historical time range by the number of historical statistical time windows.

[0048] Furthermore, the average historical water quality change coefficient recorded within the historical time range is compared with the water quality change coefficient corresponding to the current statistical time window, and the result obtained by dividing the average historical water quality change coefficient by the current water quality change coefficient is used as the periodic correction coefficient.

[0049] When the current water quality variation coefficient is greater than the historical average water quality variation coefficient, it indicates that the current wastewater quality fluctuation is higher than the historical average. In this case, the periodic correction coefficient is less than 1, and the feedback control cycle should be appropriately shortened and the detection frequency of ammonia nitrogen concentration in the effluent should be increased. When the current water quality variation coefficient is less than the historical average water quality variation coefficient, it indicates that the current wastewater quality is relatively stable. In this case, the periodic correction coefficient is greater than 1, and the feedback control cycle can be appropriately extended and the detection frequency reduced, thereby reducing operating energy consumption.

[0050] Furthermore, the cycle correction coefficient is multiplied by the preset feedback control cycle for effluent ammonia nitrogen concentration, and the calculation result is rounded to the nearest integer to obtain the suitable feedback control cycle. Using integer minutes as the feedback control cycle facilitates the PLC controller to execute feedback control at fixed time intervals, while avoiding increased control program complexity due to decimal times.

[0051] In this embodiment, the preset feedback control cycle for effluent ammonia nitrogen concentration is set to 5 minutes.

[0052] The feedback control cycle is determined based on the feedback control response time under normal operating conditions of the stripping system. Statistical analysis of 500 normal operating samples revealed that when the feedback control cycle is less than 3 minutes, the ammonia nitrogen mass transfer process within the stripping tower has not fully responded to the previous round of control parameter adjustments, easily leading to frequent adjustments by the control system. When the feedback control cycle exceeds 8 minutes, the influent water quality has already changed significantly, but the control parameters have not been updated in time, easily reducing stripping efficiency. Therefore, this embodiment selects 5 minutes as the preset feedback control cycle, ensuring both real-time control and operational stability.

[0053] Furthermore, to avoid the feedback control cycle being too short or too long under extreme operating conditions from affecting the stable operation of the stripping system, this embodiment further sets upper and lower limits for the feedback control cycle.

[0054] The minimum feedback control period was set to 2 minutes, and the maximum was set to 10 minutes. These upper and lower limits were obtained through analysis of the stripping system's response capability. Experimental results showed that when the feedback control period was less than 2 minutes, the previous round of control parameters had not yet fully taken effect on the stripping process, easily causing system control oscillations; when the feedback control period was greater than 10 minutes, the effluent ammonia nitrogen concentration might have changed significantly, but the control system had not yet made feedback corrections, easily reducing the stripping control accuracy. Therefore, the adaptive feedback control period was limited to the range of 2–10 minutes.

[0055] If the calculated adaptive feedback control period is less than 2 minutes, the final feedback control period is set to 2 minutes; if the calculated adaptive feedback control period is greater than 10 minutes, the final feedback control period is set to 10 minutes.

[0056] For example, in this embodiment, the historical database records 4320 statistical time windows corresponding to the most recent 30 days of normal operation. The average historical water quality change coefficient is 0.120. Step S100 calculates the water quality change coefficient corresponding to the current statistical time window as 0.08392. Therefore, the period correction coefficient is 0.120 ÷ 0.08392 = 1.43. The preset feedback control period for effluent ammonia nitrogen concentration is 5 minutes. Therefore, the adaptive feedback control period is 1.43 × 5 = 7.15 minutes, which is rounded to 7 minutes.

[0057] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0058] In summary, this step uses the ratio between the average historical water quality change coefficient obtained from historical data within a historical timeframe and the water quality change coefficient corresponding to the current statistical time window to dynamically adjust the preset feedback control cycle for effluent ammonia nitrogen concentration. When the current wastewater quality fluctuates significantly, the feedback control cycle can be automatically shortened, increasing the feedback control frequency; when the current wastewater quality is relatively stable, the feedback control cycle can be appropriately extended, reducing unnecessary detection times and control actions. This balances the real-time performance of stripping control, operational stability, and energy consumption, thereby improving the efficiency of wastewater ammonia nitrogen stripping treatment.

[0059] S300: Based on the water quality monitoring data sequence, optimize the ammonia nitrogen stripping control parameters with the goal of maximizing the mass transfer quality of ammonia nitrogen, obtain the optimal stripping control parameters, and calculate the stripping control time window corresponding to the water quality monitoring data sequence in combination with the optimal stripping control parameters. Step S300 in the method provided in this embodiment of the invention includes: Configure the ammonia nitrogen stripping control parameter adjustment range, wherein the ammonia nitrogen stripping control parameter adjustment range includes the fan frequency adjustment threshold and the influent flow rate adjustment threshold; Within the fan frequency control threshold and the inlet water flow control threshold, any parameter is randomly selected and combined to obtain several initial ammonia nitrogen stripping control parameters, which are used as several initial solutions. The initial solutions are combined one by one with the water quality monitoring data sequence to obtain several ammonia nitrogen stripping control schemes. The several ammonia nitrogen stripping control schemes are input into a pre-constructed ammonia nitrogen stripping predictor for prediction, and several ammonia nitrogen stripping prediction results are output. Several ammonia nitrogen mass transfer quality coefficients were obtained based on the aforementioned ammonia nitrogen stripping prediction results. With maximizing the mass transfer quality of ammonia nitrogen as the optimization objective and the adjustment range of the ammonia nitrogen stripping control parameters as the optimization space, the ammonia nitrogen stripping control parameters are optimized based on the several initial solutions and several ammonia nitrogen mass transfer quality coefficients to obtain the optimal stripping control parameters.

[0060] In this step, based on the water quality monitoring data sequence obtained in step S100, the stripping control parameters are optimized with the goal of maximizing ammonia nitrogen mass transfer quality, and the optimal stripping control parameters corresponding to the current statistical time window are obtained.

[0061] Specifically, to ensure that the stripping control parameters are always within the safe operating range of the equipment, this embodiment first configures the ammonia nitrogen stripping control parameter adjustment range.

[0062] The control parameters for ammonia nitrogen stripping include the fan frequency control threshold and the influent flow rate control threshold.

[0063] The fan frequency control threshold is determined based on the fan's rated operating parameters. Long-term operational tests on multiple engineering stripping systems revealed that when the fan operating frequency is below 30Hz, the gas-liquid contact intensity within the stripping tower is significantly insufficient, leading to a rapid decrease in ammonia nitrogen stripping efficiency. When the fan operating frequency is above 50Hz, the fan is already close to its rated operating frequency, and further increasing the fan frequency has limited effect on improving ammonia nitrogen removal efficiency while significantly increasing energy consumption. Therefore, this embodiment sets the fan frequency control range to 30–50Hz.

[0064] The influent flow rate control threshold is determined based on the design capacity of the stripping tower. Experimental results show that when the influent flow rate is less than 60% of the design flow rate, the residence time in the stripping tower is too long, resulting in low capacity utilization; when the influent flow rate exceeds 110% of the design flow rate, the gas-liquid contact time in the tower is insufficient, and the ammonia nitrogen mass transfer efficiency is significantly reduced. The design capacity of the stripping tower is 65 m³ / h, therefore, the control range is determined to be 40–70 m³ / h based on 60%–110% of the design flow rate.

[0065] Furthermore, within the two-dimensional parameter space comprised of the fan frequency control threshold and the inlet water flow control threshold, several sets of initial stripping control parameters are generated using a random uniform sampling method, so that all initial solutions are evenly distributed throughout the entire parameter space.

[0066] Specifically, random sampling spaces are first established according to the fan frequency control range and the inlet water flow control range.

[0067] Subsequently, for each set of initial solutions, a fan operating frequency is randomly generated within the fan frequency control range, and an influent flow rate is randomly generated within the influent flow rate control range. The generated fan operating frequency and influent flow rate are then combined to form a set of initial ammonia nitrogen stripping control parameters.

[0068] The operating frequency of the wind turbine is calculated using the following formula: Fan operating frequency = fan frequency lower limit + random number × (fan frequency upper limit - fan frequency lower limit).

[0069] Similarly, the influent flow rate is calculated using the following formula: Inlet flow rate = Lower limit of inlet flow rate + random number × (Upper limit of inlet flow rate - Lower limit of inlet flow rate).

[0070] The random numbers are also uniformly distributed between 0 and 1.

[0071] Furthermore, to avoid generating duplicate or highly similar initial solutions, this embodiment sets a minimum parameter spacing constraint.

[0072] Specifically, the difference in fan operating frequency between two adjacent initial solutions must be no less than 0.5 Hz, and the difference in influent flow rate must be no less than 0.5 m³ / h. If a newly generated initial solution does not meet the above minimum parameter spacing constraint, the corresponding control parameters will be randomly generated again until the minimum parameter spacing constraint is met.

[0073] The minimum parameter spacing is determined based on both the device control resolution and the optimized search accuracy. Experimental results show that when the minimum parameter spacing is less than 0.2, a large number of initial solutions are highly concentrated, reducing population diversity; when the minimum parameter spacing is greater than 1.0, the parameter coverage density is insufficient, making it easy to miss local optimal solutions. Therefore, this embodiment selects 0.5 Hz and 0.5 m³ / h as the minimum parameter spacing.

[0074] Furthermore, the above random generation process is repeated until a preset number of initial solutions are obtained.

[0075] In this embodiment, the initial number of solutions is set to 50 sets. Extensive experiments have shown that when the number of initial solutions is less than 20 sets, the coverage of control parameters is insufficient, easily overlooking optimal control parameter combinations; when the number of initial solutions exceeds 100 sets, the prediction computation time increases significantly, while the improvement in optimization effect is limited. Therefore, considering both computational efficiency and optimization accuracy, this embodiment preferably uses 50 sets of initial solutions as the initial population for the genetic algorithm.

[0076] Subsequently, each set of initial ammonia nitrogen stripping control parameters is combined with the water quality monitoring data sequence corresponding to the current statistical time window obtained in step S100 to form a corresponding ammonia nitrogen stripping control scheme.

[0077] Each ammonia nitrogen stripping control scheme includes: the wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence within the current statistical time window, as well as a corresponding set of fan operating frequency and influent flow control parameters.

[0078] Furthermore, all ammonia nitrogen stripping control schemes are input into the pre-built ammonia nitrogen stripping predictor.

[0079] All ammonia nitrogen stripping control schemes are input into the trained ammonia nitrogen stripping predictor, which outputs the predicted ammonia nitrogen removal rate, predicted ammonia nitrogen removal efficiency, and predicted energy consumption per ton of water treated under the corresponding control scheme. The three together constitute the ammonia nitrogen stripping prediction result for the corresponding control scheme.

[0080] Subsequently, the corresponding ammonia nitrogen mass transfer mass coefficients were evaluated based on the ammonia nitrogen stripping prediction results.

[0081] The ammonia-nitrogen mass transfer quality coefficient is used to characterize the quality of ammonia-nitrogen mass transfer capacity during the stripping process under the corresponding control scheme. Its value is positively correlated with the predicted stripping operation effect. The larger the ammonia-nitrogen mass transfer quality coefficient, the better the gas-liquid mass transfer capacity of the corresponding control scheme.

[0082] Furthermore, with maximizing the mass transfer quality of ammonia nitrogen as the optimization objective, and using the two-dimensional control parameter space formed by the fan frequency control range and the influent flow control range as the optimization search space, the stripping control parameters are iteratively optimized by combining all initial solutions and the corresponding ammonia nitrogen mass transfer quality coefficients to obtain the optimal stripping control parameters corresponding to the current statistical time window.

[0083] Step S300 in the method provided in this embodiment of the invention further includes: Based on historical ammonia nitrogen stripping control records, multiple sample ammonia nitrogen stripping control scheme sets were collected, and the historical ammonia nitrogen removal rate, historical ammonia nitrogen removal efficiency, and historical energy consumption per ton of water treatment corresponding to different sample ammonia nitrogen stripping control schemes were obtained as sample ammonia nitrogen stripping results, thus obtaining multiple sample ammonia nitrogen stripping results. Using multiple samples of ammonia nitrogen stripping results as supervised data, a deep learning model was trained until convergence, generating an ammonia nitrogen stripping predictor.

[0084] In this step, in order to establish an ammonia nitrogen stripping predictor that can predict the stripping operation effect according to different stripping control schemes, this embodiment first constructs a training sample set based on historical ammonia nitrogen stripping control records, and uses the training samples to supervise the training of the feedforward neural network to obtain the ammonia nitrogen stripping predictor.

[0085] Specifically, historical ammonia nitrogen stripping control records were first collected during the continuous and stable operation of the stripping system. Each historical ammonia nitrogen stripping control record corresponds to a complete stripping operation and includes the stripping control scheme within the corresponding statistical time window and the corresponding actual operating results.

[0086] Furthermore, historical ammonia nitrogen stripping control records are screened. To ensure the quality of training data, this embodiment removes historical records corresponding to equipment alarms, fan failures, online instrument malfunctions, manual interventions, and missing data, retaining only data from the normal operation of the stripping system as training samples.

[0087] In this embodiment, the historical data collection period is set to the most recent twelve months. The historical data time range is determined based on the operating conditions of the stripping system. Statistical analysis of operating data from multiple stripping projects revealed that when the historical data time range is less than six months, the sample coverage of operating conditions is insufficient, resulting in poor model generalization ability; when the historical data time range exceeds twenty-four months, the representativeness of the historical data decreases due to equipment aging, changes in packing condition, and adjustments to operating conditions. Therefore, this embodiment selects the most recent twelve months as the historical sample collection time range.

[0088] Furthermore, all historical stripping control records were divided into multiple sample ammonia nitrogen stripping control scheme sets according to statistical time windows.

[0089] Each sample ammonia nitrogen stripping control scheme set includes the wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, wastewater temperature sequence, water quality change coefficient calculated based on the above water quality monitoring data, blower operating frequency, and influent flow rate within the corresponding statistical time window.

[0090] Furthermore, in order to reduce the dimensionality of the input data and maintain the main water quality characteristics, this embodiment calculates the average values ​​corresponding to the wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence, and combines them with the water quality change coefficient calculated in step S100, which together serve as the model input features.

[0091] Therefore, the input features for each sample include: average wastewater ammonia nitrogen concentration, average wastewater pH, average wastewater temperature, water quality variation coefficient, fan operating frequency, and influent flow rate, for a total of six input features.

[0092] Furthermore, the historical stripping operation results corresponding to the ammonia nitrogen stripping control schemes for each sample were obtained.

[0093] The historical stripping operation results include historical ammonia nitrogen removal rate, historical ammonia nitrogen removal efficiency, and historical energy consumption per ton of water treated.

[0094] Historical ammonia nitrogen removal rate is calculated using the mass of ammonia nitrogen removed per unit time. The calculation method is as follows: Historical ammonia nitrogen removal rate = (influent ammonia nitrogen concentration - effluent ammonia nitrogen concentration) × treatment flow rate ÷ stripping treatment time.

[0095] The historical ammonia nitrogen removal rate is calculated as the proportion of removed ammonia nitrogen mass to the total influent ammonia nitrogen. The calculation method is as follows: Historical ammonia nitrogen removal rate = (influent ammonia nitrogen concentration - effluent ammonia nitrogen concentration) ÷ influent ammonia nitrogen concentration.

[0096] Historical energy consumption per ton of water treated is calculated using the total energy consumption per unit volume of water treated. The calculation method is as follows: Historical energy consumption per ton of water treated = total power consumption of the stripping process ÷ water volume treated.

[0097] Subsequently, the entire set of ammonia nitrogen stripping control schemes was used as the model input data, and the corresponding historical ammonia nitrogen removal rate, historical ammonia nitrogen removal efficiency, and historical energy consumption per ton of water treated were used as supervisory data to jointly construct a supervised learning training sample.

[0098] Since the input features and the supervision data have different dimensions, in order to avoid the difference in dimensions affecting the model training, this embodiment first performs normalization processing on the input data and the supervision data respectively.

[0099] Among them, historical ammonia nitrogen removal rate and historical ammonia nitrogen removal percentage are benefit-type indicators and are treated with forward normalization; historical energy consumption per ton of water treated is a cost-type indicator and is treated with reverse normalization. The calculation method is as follows: Normalized value of historical ammonia nitrogen removal rate = (historical ammonia nitrogen removal rate - historical minimum ammonia nitrogen removal rate) ÷ (historical maximum ammonia nitrogen removal rate - historical minimum ammonia nitrogen removal rate).

[0100] Normalized value of historical ammonia nitrogen removal rate = (historical ammonia nitrogen removal rate - historical minimum ammonia nitrogen removal rate) ÷ (historical maximum ammonia nitrogen removal rate - historical minimum ammonia nitrogen removal rate).

[0101] Normalized value of historical energy consumption per ton of water treated = (maximum historical energy consumption per ton of water treated - historical energy consumption per ton of water treated) ÷ (maximum historical energy consumption per ton of water treated - minimum historical energy consumption per ton of water treated).

[0102] After normalization, all input features and supervision data are mapped to the range of 0 to 1.

[0103] Furthermore, this embodiment employs a multi-layer feedforward neural network as the ammonia nitrogen stripping predictor. The feedforward neural network includes an input layer, three hidden layers, and an output layer.

[0104] The input layer is used to receive six input features, therefore the input layer is configured with six input nodes.

[0105] Hidden layers are used to learn the nonlinear mapping relationship between water quality characteristics and stripping operation results. Cross-validation experiments revealed that when there are fewer than two hidden layers, the model's prediction accuracy is low; when there are more than four hidden layers, the model training time increases significantly, and overfitting is more likely to occur. Therefore, this embodiment adopts a three-layer hidden layer structure.

[0106] The first hidden layer has 64 neurons; the second hidden layer has 32 neurons; and the third hidden layer has 16 neurons. All hidden layers use the ReLU activation function.

[0107] Furthermore, to avoid overfitting during model training, this embodiment includes a Dropout layer after each hidden layer. The Dropout ratio is set to 0.20.

[0108] The Dropout ratio is determined through cross-validation. When the Dropout ratio is less than 0.10, overfitting is more pronounced; when the Dropout ratio exceeds 0.30, the model's learning ability decreases. To maintain the basic accuracy of the validation set predictions, the Dropout ratio corresponding to the minimum mean squared error of the validation set is selected as the final parameter. Therefore, in this embodiment, 0.20 is chosen as the Dropout ratio.

[0109] The three output nodes output the predicted ammonia nitrogen removal rate, the predicted ammonia nitrogen removal percentage, and the predicted energy consumption per ton of water treated, respectively. These three predictions together constitute the ammonia nitrogen stripping prediction results for the corresponding stripping control scheme. Since all three output variables are continuous variables, a linear activation function is used in the output layer.

[0110] Furthermore, all training samples are divided into training set, validation set and test set in an 8:1:1 ratio.

[0111] The training set is used to update network parameters; the validation set is used for parameter adjustment and model convergence determination during training; and the test set is used to evaluate the model's prediction performance.

[0112] Furthermore, this embodiment uses the Adam optimization algorithm to train the feedforward neural network. The initial learning rate is set to 0.001, the batch size is set to 64, and the maximum number of training epochs is set to 500.

[0113] The learning rate was determined based on extensive training experiments. When the learning rate is less than 0.0005, the model convergence speed decreases significantly; when the learning rate exceeds 0.005, the model training is prone to oscillations. Therefore, in this embodiment, 0.001 is selected as the initial learning rate.

[0114] Batch size is determined through cross-validation. When the batch size is less than 32, the model gradient fluctuates greatly, resulting in poor training stability; when the batch size exceeds 128, training efficiency decreases and memory usage increases significantly. Therefore, this embodiment selects 64 as the batch size.

[0115] Furthermore, the model loss function uses the total mean squared error.

[0116] Specifically, the mean square errors corresponding to the predicted ammonia nitrogen removal rate, the predicted ammonia nitrogen removal efficiency, and the predicted energy consumption per ton of water treatment are calculated separately, and the sum of the three mean square errors is used as the total loss function of the model.

[0117] Each mean square error is calculated using the average of the squared deviations between the corresponding predicted and actual values.

[0118] During model training, if the decrease in validation set loss is less than 0.001 for twenty consecutive training rounds, the model is considered to have converged and training is stopped.

[0119] The convergence threshold was determined through extensive experiments. When the convergence threshold is less than 0.0005, the model training time increases significantly while the improvement in prediction accuracy is limited; when the convergence threshold exceeds 0.005, the model has not yet fully converged. Therefore, in this embodiment, 0.001 is selected as the model convergence criterion.

[0120] After training, save the network parameters corresponding to the minimum total loss function of the validation set, and generate an ammonia nitrogen stripping predictor.

[0121] For example, this embodiment collected 8650 sets of historical operation records of the stripping system over the past twelve months. After screening, 412 sets of records with equipment alarms, instrument malfunctions, and missing data were removed, leaving 8238 valid samples. Of these, 6590 sets were used as the training set, 824 sets as the validation set, and 824 sets as the test set. The feedforward neural network input layer had 6 input nodes, and the three hidden layers had 64, 32, and 16 neurons respectively. The output layer had 3 output nodes. The model was trained using the Adam optimization algorithm, with an initial learning rate of 0.001 and a Dropout ratio of 0.20. After 186 rounds of training, the validation set loss decreased by less than 0.001 for twenty consecutive rounds, therefore training was stopped, and the ammonia nitrogen stripping predictor was generated.

[0122] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0123] Step S300 in the method provided in this embodiment of the invention further includes: The ammonia nitrogen stripping prediction results include the predicted ammonia nitrogen removal rate, the predicted ammonia nitrogen removal percentage, and the predicted energy consumption per ton of water treated. After normalizing the predicted ammonia nitrogen removal rate, predicted ammonia nitrogen removal efficiency, and predicted energy consumption per ton of water treated, a weighted evaluation is performed to obtain the ammonia nitrogen mass transfer quality coefficient. The ammonia nitrogen mass transfer quality coefficient is positively correlated with the predicted ammonia nitrogen removal rate and predicted ammonia nitrogen removal efficiency, and negatively correlated with the predicted energy consumption per ton of water treated.

[0124] In this step, based on the ammonia nitrogen stripping prediction results output by the ammonia nitrogen stripping predictor, the stripping operation effect corresponding to different stripping control schemes is comprehensively evaluated to obtain the corresponding ammonia nitrogen mass transfer quality coefficient, which provides an evaluation basis for subsequent optimization of stripping control parameters.

[0125] Specifically, each ammonia nitrogen stripping control scheme in the steps is input into the trained ammonia nitrogen stripping predictor, and the corresponding ammonia nitrogen stripping prediction results are output.

[0126] The ammonia nitrogen stripping prediction results include the predicted ammonia nitrogen removal rate, the predicted ammonia nitrogen removal percentage, and the predicted energy consumption per ton of water treated.

[0127] Since the three prediction indicators have different dimensions, in order to eliminate the impact of the difference in dimensions on the comprehensive evaluation results, this embodiment first performs normalization processing on the three prediction indicators.

[0128] Among them, the predicted ammonia nitrogen removal rate and the predicted ammonia nitrogen removal percentage are both benefit-type indicators; the higher the value, the better the stripping effect. Therefore, positive normalization is used. The calculation method is as follows: Normalized value of predicted ammonia nitrogen removal rate = (predicted ammonia nitrogen removal rate - historical minimum predicted ammonia nitrogen removal rate) ÷ (historical maximum predicted ammonia nitrogen removal rate - historical minimum predicted ammonia nitrogen removal rate).

[0129] Normalized value of predicted ammonia nitrogen removal rate = (predicted ammonia nitrogen removal rate - historical minimum predicted ammonia nitrogen removal rate) ÷ (historical maximum predicted ammonia nitrogen removal rate - historical minimum predicted ammonia nitrogen removal rate).

[0130] The predicted energy consumption per ton of water treated is a cost-related indicator; a smaller value indicates better stripping performance. Therefore, reverse normalization is used. The calculation method is as follows: Normalized value of predicted energy consumption per ton of water treatment = (maximum predicted energy consumption per ton of water treatment in history - predicted energy consumption per ton of water treatment) ÷ (maximum predicted energy consumption per ton of water treatment in history - minimum predicted energy consumption per ton of water treatment in history).

[0131] The historical maximum and minimum values ​​are obtained from the statistical results in the historical sample database corresponding to the training ammonia nitrogen stripping predictor, and are updated synchronously with the historical database updates.

[0132] Furthermore, in order to comprehensively reflect the influence of different prediction indicators on the stripping mass transfer effect, this embodiment uses a weighted evaluation method to calculate the ammonia nitrogen mass transfer quality coefficient.

[0133] Since the ammonia nitrogen removal rate directly reflects the mass transfer capacity of ammonia nitrogen per unit time and is the main indicator affecting the stripping efficiency, it is given the highest evaluation weight; the ammonia nitrogen removal rate is used to reflect the stripping treatment effect, so it is given the second highest evaluation weight; the energy consumption per ton of water treated is mainly used to evaluate the operational economy, so it is given a relatively low evaluation weight.

[0134] In this embodiment, the weight for predicting ammonia nitrogen removal rate is set to 0.45, the weight for predicting ammonia nitrogen removal percentage is set to 0.35, and the weight for predicting energy consumption per ton of water treatment is set to 0.20.

[0135] The aforementioned weights were obtained through sensitivity analysis of historical operating samples. Cross-validation was performed on historical samples using different weight combinations. When the weight for predicting ammonia nitrogen removal rate was greater than 0.50, the model focused too much on treatment efficiency, potentially increasing operational energy consumption. Conversely, when the weight for predicting energy consumption per ton of water treated was greater than 0.30, the model tended to prioritize reducing energy consumption, leading to a decrease in stripping efficiency. Considering both stripping efficiency and operational economics, the final weight combination was determined.

[0136] Subsequently, the mass transfer mass coefficient of ammonia nitrogen was calculated based on three normalized indices. The calculation method is as follows: The mass transfer coefficient of ammonia nitrogen is calculated as follows: (Normalized value of predicted ammonia nitrogen removal rate) × 0.45 + (Normalized value of predicted ammonia nitrogen removal percentage) × 0.35 + (Normalized value of predicted energy consumption per ton of water treated) × 0.20.

[0137] The larger the mass transfer coefficient of ammonia nitrogen, the better the overall stripping effect, indicating that the corresponding stripping control scheme can achieve a high ammonia nitrogen removal capacity while maintaining low operating energy consumption.

[0138] Furthermore, to avoid misjudgment when the mass transfer mass coefficients of ammonia nitrogen corresponding to multiple control schemes are small in difference, this embodiment sets a mass coefficient difference threshold. The mass coefficient difference threshold is set to 0.02.

[0139] This threshold was obtained through analysis of historical operating samples. When the quality coefficient difference threshold is less than 0.01, different control schemes are easily affected by prediction errors; when the quality coefficient difference threshold is greater than 0.05, some control schemes with practical optimization value are easily ignored. Therefore, this embodiment selects 0.02 as the quality coefficient difference threshold.

[0140] When the difference between the ammonia-nitrogen mass transfer mass coefficients of the two control schemes is less than 0.02, the overall performance of the two control schemes is considered to be equivalent; when the difference is not less than 0.02, the control scheme with the larger ammonia-nitrogen mass transfer mass coefficient is preferred for subsequent stripping control parameter optimization.

[0141] For example, in this embodiment, a certain stripping control scheme is predicted by an ammonia nitrogen stripping predictor to have a predicted ammonia nitrogen removal rate of 18.5 kg / h, a predicted ammonia nitrogen removal rate of 92%, and a predicted energy consumption per ton of water treated of 2.30 kW·h / m³. Historical sample database statistics show that the predicted ammonia nitrogen removal rate varies from 10 to 22 kg / h, the predicted ammonia nitrogen removal rate varies from 75% to 96%, and the predicted energy consumption per ton of water treated varies from 1.80 to 3.50 kW·h / m³. Therefore, the normalized value of the predicted ammonia nitrogen removal rate is (18.5-10) ÷ (22-10) = 0.708, the normalized value of the predicted ammonia nitrogen removal rate is (92-75) ÷ (96-75) = 0.810, and the normalized value of the predicted energy consumption per ton of water treated is (3.50-2.30) ÷ (3.50-1.80) = 0.706. Further calculations yielded an ammonia-nitrogen mass transfer quality coefficient of 0.708×0.45+0.810×0.35+0.706×0.20=0.744. Since the ammonia-nitrogen mass transfer quality coefficient is 0.744, which is higher than the ammonia-nitrogen mass transfer quality coefficient of 0.718 corresponding to another control scheme, and the difference between the two is 0.026, which is greater than the quality coefficient difference threshold of 0.02, the current stripping control scheme is preferentially retained for subsequent control parameter optimization.

[0142] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0143] Step S300 in the method provided in this embodiment of the invention further includes: Using the aforementioned initial solutions as the initial population, and based on the aforementioned ammonia-nitrogen mass transfer quality coefficients, the roulette wheel selection method is used to select multiple initial solutions with larger ammonia-nitrogen mass transfer quality coefficients from the initial population as parent individuals. For the parent individuals, a crossover operation is performed with a preset crossover probability to generate offspring individuals; For the offspring individuals, a random perturbation is applied to the coding value of some offspring individuals with a preset mutation probability, and the mutated coding value is constrained within the adjustment range of the ammonia nitrogen stripping control parameter; The offspring individuals are used as a new generation population. The ammonia nitrogen mass transfer quality coefficient corresponding to each individual in the new generation population is calculated. The new generation population replaces the initial population, and the selection, crossover, mutation, and population update steps are repeated. Each time the process is repeated, it is determined whether the current iteration count has reached the preset maximum number of iterations. When the preset maximum number of iterations is reached, the iteration is stopped, and the individual with the largest ammonia nitrogen mass transfer mass coefficient in the current population is analyzed as the optimal stripping control parameter output. The optimal stripping control parameter includes the optimal fan frequency and the optimal influent flow rate.

[0144] In this step, in order to quickly search for the combination of control parameters that can obtain the maximum ammonia nitrogen mass transfer quality within the control parameter adjustment range of ammonia nitrogen stripping, this embodiment uses a genetic algorithm to globally optimize the stripping control parameters and uses the ammonia nitrogen mass transfer quality coefficient as the fitness function.

[0145] Specifically, the initial ammonia nitrogen stripping control parameters generated in step S300 are used as the initial population for the genetic algorithm.

[0146] Each initial solution corresponds to an individual in the genetic algorithm, and each individual includes two control variables: the fan operating frequency and the inlet water flow rate.

[0147] This embodiment uses real-number encoding for individual code. The first coded value represents the fan operating frequency, and the second coded value represents the influent flow rate. Using real-number encoding avoids quantization errors caused by binary encoding and improves the optimization accuracy of continuous variables; therefore, this embodiment employs real-number encoding.

[0148] In this embodiment, the initial population size is set to 50 individuals.

[0149] The population size was determined based on both optimization accuracy and computation time. Experimental results showed that when the population size was less than 30, individual diversity was insufficient, making it prone to getting trapped in local optima; when the population size exceeded 80, the improvement in optimization accuracy was limited, while the prediction computation time increased significantly. Therefore, this embodiment selected 50 as the initial population size.

[0150] Furthermore, the ammonia nitrogen mass transfer mass coefficients corresponding to each initial solution calculated in step S300 are used as the fitness values ​​of the corresponding individuals.

[0151] The larger the ammonia nitrogen mass transfer mass coefficient, the better the stripping operation effect can be obtained by the corresponding control parameter combination, and therefore the higher the probability that the corresponding individual will be retained to the next generation.

[0152] Subsequently, the roulette wheel selection method was used to complete the selection of the parent generation.

[0153] Specifically, first, the sum of the fitness of all individuals is calculated. Then, the proportion of each individual's fitness to the total fitness is calculated, and this proportion is used as the selection probability of the corresponding individual. The selection probability of each individual is calculated as follows: The probability of an individual's selection = the mass transfer mass coefficient of ammonia nitrogen for that individual ÷ the sum of the mass transfer mass coefficients of ammonia nitrogen for all individuals.

[0154] Subsequently, a cumulative probability interval is constructed based on the cumulative selection probability of each individual, and a random number between 0 and 1 is randomly generated. When the random number falls within a certain cumulative probability interval, the corresponding individual is selected as the parent individual.

[0155] Furthermore, a crossover operation is performed on the parent individuals. This embodiment uses a simulated binary crossover method to complete the real-number encoded crossover. The crossover probability is set to 0.85.

[0156] The crossover probability was determined based on extensive optimization experiments. When the crossover probability is less than 0.70, the number of new individuals generated is insufficient, resulting in low search efficiency; when the crossover probability exceeds 0.95, the structure of excellent individuals is easily disrupted. Therefore, this embodiment selects 0.85 as the crossover probability.

[0157] Specifically, when the randomly generated probability value is less than 0.85, new real-number codes are generated for the wind turbine operating frequency and water inflow rate corresponding to the two parent individuals according to the simulated binary crossover rule, thereby generating two child individuals; otherwise, the corresponding parent individuals are directly retained to enter the next generation.

[0158] Furthermore, a mutation operation is performed on the generated offspring individuals. In this embodiment, a Gaussian random perturbation method is used to perform real-number encoding mutation. The mutation probability is set to 0.05.

[0159] The mutation probability was determined based on population diversity experiments. When the mutation probability is less than 0.02, the population is prone to premature convergence; when the mutation probability exceeds 0.10, the random perturbation is too large, and the stability of the algorithm decreases. Therefore, this embodiment selects 0.05 as the mutation probability.

[0160] When the randomly generated probability value is less than 0.05, the operating frequency of the fan or the influent flow rate in the offspring individuals will be increased or decreased by a random disturbance value.

[0161] In this embodiment, the standard deviation of the fan operating frequency disturbance is set to 1 Hz, and the standard deviation of the influent flow rate disturbance is set to 2 m³ / h. These disturbance standard deviations are obtained based on sensitivity analysis of the control parameters. When the disturbance amplitude is too small, the algorithm's ability to escape local optima is insufficient; when the disturbance amplitude is too large, it easily destroys the already obtained optimal individuals. Therefore, 1 Hz and 2 m³ / h are selected as the random disturbance standard deviations, respectively.

[0162] Subsequently, boundary constraints are applied to the mutated control parameters.

[0163] Specifically, when the fan operating frequency is less than 30Hz, it is corrected to 30Hz; when the fan operating frequency is greater than 50Hz, it is corrected to 50Hz; when the inlet water flow rate is less than 40m³ / h, it is corrected to 40m³ / h; and when the inlet water flow rate is greater than 70m³ / h, it is corrected to 70m³ / h.

[0164] By using boundary constraints, we ensure that all control parameters remain within the control range of ammonia nitrogen stripping.

[0165] Furthermore, all offspring individuals will be grouped into a new generation of population.

[0166] The size of the new generation population remains consistent with the initial population size, meaning that each generation contains 50 individuals.

[0167] Then, the previous generation of population is replaced with a new generation of population, and the process of selection, crossover, mutation and population renewal is repeated.

[0168] Furthermore, after each round of population update, the current iteration number is determined.

[0169] In this embodiment, the maximum number of iterations is set to 80.

[0170] The maximum number of iterations was determined based on optimization convergence experiments. Experimental results show that when the maximum number of iterations is less than 50 generations, the algorithm has not yet fully converged; when it exceeds 100 generations, the fitness improvement is less than 1%, and continuing iteration is not very meaningful. Therefore, this embodiment selects 80 generations as the maximum number of iterations. When the current iteration count reaches 80 generations, the genetic algorithm iteration is stopped.

[0171] Subsequently, the individual with the largest ammonia nitrogen mass transfer mass coefficient is selected from the current population as the optimal individual, and the corresponding fan operating frequency and influent flow rate are obtained from the analysis of the optimal individual, which are used as the optimal stripping control parameters for the current statistical time window.

[0172] For example, in this embodiment, the initial population size is set to 50 individuals, and the sum of the ammonia-nitrogen mass transfer coefficients of all individuals is 40.68. Specifically, the 18th individual corresponds to a fan operating frequency of 42Hz and an influent flow rate of 56 m³ / h, with an ammonia-nitrogen mass transfer coefficient of 0.892; the 31st individual corresponds to a fan operating frequency of 45Hz and an influent flow rate of 54 m³ / h, with an ammonia-nitrogen mass transfer coefficient of 0.915. Therefore, the probability of the 31st individual being selected is 0.915 ÷ 40.68 = 0.0225. Subsequently, the 18th and 31st individuals undergo simulated binary crossover with a crossover probability of 0.85, generating offspring individuals with a fan operating frequency of 43Hz and an influent flow rate of 55 m³ / h. Then, a Gaussian random perturbation is applied to the fan operating frequency with a mutation probability of 0.05, resulting in new individuals with a fan operating frequency of 44Hz and an influent flow rate of 55 m³ / h. The pre-trained ammonia nitrogen stripping predictor predicted that the ammonia nitrogen mass transfer coefficient corresponding to this individual would increase to 0.936. After further iterative optimization, the optimal ammonia nitrogen mass transfer coefficient increased to 0.951 in the 40th generation, 0.960 in the 60th generation, and reached 0.964 in the 80th generation. Finally, the individual with the highest fitness in the current population corresponds to a blower operating frequency of 44Hz and an influent flow rate of 55m³ / h, and is therefore determined as the optimal stripping control parameters for the stripping system to execute within the current feedback control cycle.

[0173] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0174] Step S300 in the method provided in this embodiment of the invention further includes: Obtain the optimal influent flow rate from the optimal stripping control parameters, and obtain the preset effective liquid holding volume of the stripping tower; Based on the optimal influent flow rate and the effective liquid holding volume of the stripping tower, the expected hydraulic residence time of the wastewater in the stripping tower under the optimal influent flow rate is calculated, and the expected hydraulic residence time is used as the stripping control time window corresponding to the water quality monitoring data sequence.

[0175] In this step, based on the optimized stripping control parameters, the stripping control time window corresponding to the current stripping control parameters is further determined to ensure that the control parameters can cover the duration of the main gas-liquid mass transfer process of the wastewater in the stripping tower.

[0176] Specifically, the optimal influent flow rate is obtained from the optimal stripping control parameters determined in step S300, and the effective liquid holding volume is obtained according to the stripping tower design parameters.

[0177] The effective liquid holding volume of the stripping tower refers to the actual liquid volume that can participate in the gas-liquid mass transfer reaction under normal operating conditions, excluding the liquid volume in the water distribution area, liquid collection tank, pipeline dead zone, and packing voids where effective flow cannot be formed.

[0178] In this embodiment, the effective liquid holding volume of the stripping tower is determined using equipment design parameters.

[0179] Specifically, the effective height and effective cross-sectional area of ​​the stripping tower are obtained from the design drawings of the stripping tower equipment. The effective liquid holding volume of the stripping tower is then calculated after deducting the volume occupied by the packing and the volume occupied by the internal components.

[0180] The effective liquid holding volume of the stripping tower is calculated as follows: Effective liquid holding volume of stripping tower = Effective cross-sectional area of ​​stripping tower × Effective liquid height - Volume occupied by packing - Volume occupied by internal components.

[0181] Using the effective liquid holding volume as the basis for calculation can more accurately reflect the actual liquid volume involved in the stripping mass transfer, thereby improving the accuracy of subsequent control time window calculations.

[0182] Furthermore, the expected hydraulic residence time of wastewater in the stripping tower is calculated based on the optimal influent flow rate and the effective liquid holding volume of the stripping tower.

[0183] The expected hydraulic retention time is calculated by dividing the effective liquid holding volume of the stripping tower by the optimal influent flow rate. The calculation method is as follows: Expected hydraulic residence time = effective liquid holding volume of the stripping tower ÷ optimal influent flow rate.

[0184] The effective liquid holding capacity of the stripping tower is expressed in m³; the optimal influent flow rate is expressed in m³ / h; and the calculated expected hydraulic residence time is expressed in h. When used for PLC control, the expected hydraulic residence time is converted to min.

[0185] The expected hydraulic retention time represents the theoretical average residence time required for wastewater to complete the main stripping mass transfer process from the time it enters the stripping tower.

[0186] Furthermore, the expected hydraulic retention time is used as the stripping control time window corresponding to the current water quality monitoring data sequence.

[0187] The stripping control time window represents the duration during which the current optimal stripping control parameters remain unchanged.

[0188] Before the end of the stripping control time window, the stripping control parameter optimization will not be re-executed. Instead, after the time window ends, the latest water quality monitoring data will be collected again, and steps S100 to S300 will be re-executed to update the stripping control parameters corresponding to the next time window.

[0189] By adopting the above method, we can avoid frequently modifying control parameters before the wastewater stripping process is completed, thus improving the stability of stripping control.

[0190] Furthermore, to avoid extreme operating conditions causing the stripping control time window to be too short or too long, this embodiment further sets upper and lower limits for the stripping control time window.

[0191] The minimum value of the stripping control time window is set to 6 minutes, and the maximum value is set to 18 minutes.

[0192] The above upper and lower limits were obtained based on statistical data from the actual operation of the stripping tower. Experimental results show that when the stripping control time window is less than 6 minutes, most of the wastewater has not yet completed the main gas-liquid mass transfer process, and frequent adjustments to the control parameters can easily cause system oscillations. When the stripping control time window exceeds 18 minutes, the influent water quality may have changed significantly, but the control parameters have not been updated, which can easily reduce the stripping control accuracy. Therefore, this embodiment limits the stripping control time window to the range of 6 to 18 minutes.

[0193] When the calculated expected hydraulic residence time is less than 6 minutes, the stripping control time window is set to 6 minutes; when the calculated expected hydraulic residence time is greater than 18 minutes, the stripping control time window is set to 18 minutes; when the calculation result is between 6 and 18 minutes, the calculated expected hydraulic residence time is directly used as the stripping control time window.

[0194] For example, in this embodiment, the optimal influent flow rate obtained in step S300 is 55 m³ / h. The design parameters of the stripping tower equipment show that the effective cross-sectional area of ​​the stripping tower is 3.80 m², the effective liquid height is 2.60 m, the packing volume is 0.90 m³, and the internal component volume is 0.38 m³. Therefore, the effective liquid holding capacity of the stripping tower = 3.80 × 2.60 - 0.90 - 0.38 = 8.60 m³, and the expected hydraulic residence time = 8.60 ÷ 55 = 0.156 h. Converting hours to minutes, the expected hydraulic residence time = 0.156 × 60 = 9.36 min. Since 9.36 min falls between 6 and 18 min, the current stripping control time window is determined to be 9.36 min. Within the 9.36 min period, the optimal blower operating frequency of 44 Hz and the optimal influent flow rate of 55 m³ / h obtained in step S300 were continuously used to perform stripping control; after the time window ended, the latest water quality monitoring data were collected again, and the stripping control parameter optimization process was repeated.

[0195] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0196] In summary, this step utilizes an ammonia nitrogen stripping predictor to predict and evaluate different stripping control schemes, and optimizes the stripping control parameters with the goal of maximizing the ammonia nitrogen mass transfer coefficient to obtain the optimal stripping control parameters; furthermore, it determines the stripping control time window based on the optimal influent flow rate to achieve dynamic optimization control of the stripping process, thereby improving ammonia nitrogen removal efficiency and operational stability.

[0197] S400: Within the stripping control time window, ammonia nitrogen stripping control is performed according to the optimal stripping control parameters, and the stripping control parameters are iteratively optimized based on the real-time effluent ammonia nitrogen concentration obtained from periodic monitoring according to the adaptive feedback control cycle.

[0198] Step S400 in the method provided in this embodiment of the invention includes: Within the stripping control time window corresponding to the current batch of wastewater, according to the adaptive feedback control cycle, the real-time effluent ammonia nitrogen concentration of the stripping tower is periodically obtained, and the real-time effluent ammonia nitrogen concentration is compared with the preset expected effluent ammonia nitrogen concentration to calculate the effluent ammonia nitrogen concentration deviation. Determine whether the deviation of the effluent ammonia nitrogen concentration exceeds the preset tolerance range: If the deviation of the effluent ammonia nitrogen concentration exceeds the preset tolerance range, then the currently executed optimal stripping control parameters are locally fine-tuned according to the direction and magnitude of the deviation of the effluent ammonia nitrogen concentration, and the corrected current stripping control parameters are generated to replace the optimal stripping control parameters for ammonia nitrogen stripping control. If the deviation of the effluent ammonia nitrogen concentration does not exceed the preset tolerance range, the current stripping control parameters shall remain unchanged. The steps for local fine-tuning and correction are as follows: The ratio of the deviation of the effluent ammonia nitrogen concentration to the expected effluent ammonia nitrogen concentration is calculated as the relative deviation rate; When the real-time effluent ammonia nitrogen concentration is greater than the expected effluent ammonia nitrogen concentration, it is determined that the current stripping force is insufficient. The current fan frequency is increased proportionally according to the relative deviation rate, or the current influent flow rate is decreased proportionally according to the relative deviation rate. The absolute value of the adjustment range of the fan frequency is positively correlated with the relative deviation rate, and the absolute value of the adjustment range of the influent flow rate is positively correlated with the relative deviation rate. When the real-time effluent ammonia nitrogen concentration is less than the expected effluent ammonia nitrogen concentration, it is determined that the current stripping force is excessive. The current fan frequency is reduced proportionally according to the relative deviation rate, or the current influent flow rate is increased proportionally according to the relative deviation rate. The absolute value of the adjustment range of the fan frequency is positively correlated with the relative deviation rate, and the absolute value of the adjustment range of the influent flow rate is positively correlated with the relative deviation rate.

[0199] In this step, based on the optimal stripping control parameters determined in step S300, stripping control is performed within the stripping control time window corresponding to the current batch of wastewater. In accordance with the adaptive feedback control cycle determined in step S200, the ammonia nitrogen concentration in the effluent is detected in real time. Based on the detection results, the optimal stripping control parameters are locally fine-tuned to achieve closed-loop feedback control of the stripping process.

[0200] Specifically, within the stripping control time window corresponding to the current batch of wastewater, the PLC controller periodically reads the real-time effluent ammonia nitrogen concentration output by the online ammonia nitrogen analyzer installed on the effluent pipeline of the stripping tower, according to the adaptive feedback control cycle.

[0201] The adaptive feedback control cycle is the adaptive feedback control cycle calculated in step S200, and the stripping control time window is the stripping control time window calculated in step S300 based on the optimal influent flow rate.

[0202] Furthermore, the real-time effluent ammonia nitrogen concentration is compared with the preset expected effluent ammonia nitrogen concentration to calculate the effluent ammonia nitrogen concentration deviation.

[0203] The deviation of effluent ammonia nitrogen concentration is calculated as the difference between the real-time effluent ammonia nitrogen concentration and the expected effluent ammonia nitrogen concentration. The calculation method is as follows: Effluent ammonia nitrogen concentration deviation = Real-time effluent ammonia nitrogen concentration - Expected effluent ammonia nitrogen concentration.

[0204] When the deviation of the effluent ammonia nitrogen concentration is positive, it indicates that the current effluent ammonia nitrogen concentration is higher than the target value and the stripping intensity is insufficient; when the deviation of the effluent ammonia nitrogen concentration is negative, it indicates that the current effluent ammonia nitrogen concentration is lower than the target value and the stripping intensity is redundant.

[0205] Furthermore, to avoid frequent adjustments to control parameters due to online instrument measurement errors or instantaneous fluctuations, this embodiment sets a tolerance range for the effluent ammonia nitrogen concentration. In this embodiment, the tolerance range is set to ±2 mg / L.

[0206] The tolerance range is determined based on the measurement accuracy of the online ammonia nitrogen analyzer and the operational fluctuations of the stripping system. Statistical analysis of long-term operating data from multiple stripping systems revealed that when the tolerance range is less than ±1 mg / L, the control system is easily affected by online detection noise, leading to frequent fluctuations in control parameters; when the tolerance range is greater than ±3 mg / L, the actual effluent ammonia nitrogen concentration has already deviated significantly, but the control system has not yet corrected for this, affecting the stripping control accuracy. Therefore, this embodiment preferably uses ±2 mg / L as the tolerance range.

[0207] Subsequently, it is determined whether the deviation of the effluent ammonia nitrogen concentration exceeds the tolerance range. When the deviation is within the tolerance range, the current stripping operation is considered to meet the control requirements, and the current stripping control parameters are maintained unchanged. When the deviation exceeds the tolerance range, the current stripping control parameters are finely adjusted locally.

[0208] Furthermore, in order to obtain a reasonable adjustment range under different deviation conditions, this embodiment calculates the relative deviation rate.

[0209] The relative deviation rate is calculated as the ratio between the deviation of the effluent ammonia nitrogen concentration and the expected effluent ammonia nitrogen concentration. The calculation method is as follows: Relative deviation rate = deviation of effluent ammonia nitrogen concentration ÷ expected effluent ammonia nitrogen concentration.

[0210] The larger the relative deviation rate, the greater the deviation of the current stripping effect from the target, and the adjustment range of the control parameters should be increased accordingly.

[0211] Furthermore, if the real-time effluent ammonia nitrogen concentration is greater than the expected effluent ammonia nitrogen concentration, the stripping force is deemed insufficient.

[0212] At this point, the fan operating frequency is increased proportionally based on the relative deviation rate, or the inlet water flow rate is decreased proportionally.

[0213] The frequency adjustment amount for the fan is calculated using the following formula: Fan frequency adjustment amount = Fan frequency adjustment coefficient × Relative deviation rate × Current fan operating frequency.

[0214] The adjustment amount of the inlet flow rate is calculated using the following formula: Inlet flow rate adjustment = Inlet flow rate adjustment coefficient × Relative deviation rate × Current inlet flow rate.

[0215] Subsequently, the operating frequency of the blower is increased by the corresponding adjustment amount, or the inlet water flow is decreased by the corresponding adjustment amount.

[0216] Furthermore, when the real-time effluent ammonia nitrogen concentration is less than the expected effluent ammonia nitrogen concentration, the current stripping force is determined to be excessive.

[0217] At this point, the fan operating frequency is reduced proportionally based on the relative deviation rate, or the inlet water flow rate is increased proportionally.

[0218] The adjustment amount corresponds to the decrease in the fan operating frequency and the adjustment amount corresponds to the increase in the inlet water flow. The calculation method for the adjustment amount is the same as that described above.

[0219] Furthermore, in this embodiment, the fan frequency adjustment coefficient is set to 0.60, and the inlet water flow rate adjustment coefficient is set to 0.40.

[0220] The two adjustment coefficients were determined based on the sensitivity analysis of the stripping system. Extensive operational experiments revealed that changes in fan frequency have a direct impact on ammonia nitrogen mass transfer capacity, while changes in influent flow rate affect both residence time and processing capacity. Therefore, the weight of fan frequency adjustment is appropriately higher than that of influent flow rate adjustment. When the fan frequency adjustment coefficient is less than 0.50, the stripping enhancement effect is not significant; when it exceeds 0.80, it easily leads to frequent and significant fan adjustments; when the influent flow rate adjustment coefficient exceeds 0.50, it easily affects the stability of the processing load. Therefore, this embodiment preferably uses a fan frequency adjustment coefficient of 0.60 and an influent flow rate adjustment coefficient of 0.40. When the fan operating frequency is at least 5Hz away from the upper limit, the fan operating frequency is adjusted first; when the fan operating frequency reaches or approaches the upper limit, the influent flow rate is adjusted next.

[0221] Furthermore, to prevent the control parameters from exceeding the safe operating range of the equipment after local fine-tuning, boundary constraints are applied to the corrected control parameters.

[0222] Specifically, when the corrected fan operating frequency is below 30Hz, it is corrected to 30Hz; when it is above 50Hz, it is corrected to 50Hz; when the corrected inlet flow rate is below 40m³ / h, it is corrected to 40m³ / h; when it is above 70m³ / h, it is corrected to 70m³ / h.

[0223] For example, in this embodiment, the optimal stripping control parameters output in step S300 are: fan operating frequency 44Hz, influent flow rate 55m³ / h; the adaptation feedback control cycle determined in step S200 is 7min; and the current stripping control time window is 11min. The desired effluent ammonia nitrogen concentration is set to 15mg / L. After 7min, the real-time effluent ammonia nitrogen concentration is detected to be 18mg / L, so the effluent ammonia nitrogen concentration deviation is 18-15=3mg / L. Since 3mg / L is greater than the ±2mg / L tolerance range, a local fine-tuning is performed. The relative deviation rate is: 3÷15=0.20. If the fan operating frequency is adjusted, the fan frequency adjustment amount is: 0.60×0.20×44=5.28Hz. The corrected fan operating frequency is: 44+5.28=49.28Hz. Since it does not exceed 50Hz, 49.28Hz is ultimately used for stripping control. If the inlet flow rate is adjusted, the adjustment amount is: 0.40 × 0.20 × 55 = 4.40 m³ / h. The corrected inlet flow rate is: 55 - 4.40 = 50.60 m³ / h. The PLC controller selects one of the correction methods according to the preset control strategy to update the current blow-off control parameters and continues to execute subsequent feedback control.

[0224] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0225] In summary, this step, based on obtaining the optimal stripping control parameters, constructs a closed-loop feedback regulation mechanism by combining the adaptive feedback control cycle and real-time effluent ammonia nitrogen concentration. When the effluent ammonia nitrogen concentration deviates from the target value, the stripping control parameters are proportionally and locally fine-tuned, enabling the stripping control to respond promptly to changes in operating status. This ensures that the effluent ammonia nitrogen concentration remains stable and meets the standards while reducing large fluctuations in control parameters, thereby improving the operational stability, control accuracy, and economy of the stripping system.

[0226] Example 2, as Figure 3 As shown, the present invention provides a wastewater treatment system based on ammonia nitrogen stripping, the system comprising: Water quality monitoring module 11 is used to continuously collect the ammonia nitrogen concentration, pH value and temperature of wastewater on the main inlet pipe before the wastewater enters the stripping tower at a preset sampling period, and generate a water quality monitoring data sequence. The feedback cycle generation module 12 is used to calculate the water quality change coefficient based on the water quality monitoring data sequence, use the water quality change coefficient to correct the preset feedback control cycle of effluent ammonia nitrogen concentration, and generate an adapted feedback control cycle. The stripping parameter optimization module 13 is used to optimize the ammonia nitrogen stripping control parameters based on the water quality monitoring data sequence with the goal of maximizing the mass transfer quality of ammonia nitrogen, to obtain the optimal stripping control parameters, and to back-calculate the stripping control time window corresponding to the water quality monitoring data sequence in combination with the optimal stripping control parameters. The stripping control module 14 is used to perform ammonia nitrogen stripping control according to the optimal stripping control parameters within the stripping control time window, and to iteratively optimize the stripping control parameters based on the real-time effluent ammonia nitrogen concentration obtained by periodic monitoring according to the adaptive feedback regulation cycle.

[0227] In this embodiment of the invention, the water quality monitoring module 11 is further used for: The water quality monitoring data sequence includes wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence; The standard deviations of the wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence were calculated respectively to obtain the ammonia nitrogen concentration change index, pH value change index, and temperature change index. The water quality change coefficient is obtained by weighting the ammonia nitrogen concentration change index, pH value change index and temperature change index. The weight values ​​are adaptively determined based on the water quality characteristic state in the current water quality monitoring data sequence, and the sum of the weight values ​​of the three is 1.

[0228] In this embodiment of the invention, the feedback cycle generation module 12 is further configured to: The ratio of the average historical water quality change coefficient recorded within the historical time range to the water quality change coefficient is used as the periodic correction coefficient; The product of the cycle correction coefficient and the preset feedback control cycle for effluent ammonia nitrogen concentration is rounded down to obtain the adaptive feedback control cycle.

[0229] In this embodiment of the invention, the stripping parameter optimization module 13 is further used for: Configure the ammonia nitrogen stripping control parameter adjustment range, wherein the ammonia nitrogen stripping control parameter adjustment range includes the fan frequency adjustment threshold and the influent flow rate adjustment threshold; Within the fan frequency control threshold and the inlet water flow control threshold, any parameter is randomly selected and combined to obtain several initial ammonia nitrogen stripping control parameters, which are used as several initial solutions. The initial solutions are combined one by one with the water quality monitoring data sequence to obtain several ammonia nitrogen stripping control schemes. The several ammonia nitrogen stripping control schemes are input into a pre-constructed ammonia nitrogen stripping predictor for prediction, and several ammonia nitrogen stripping prediction results are output. Several ammonia nitrogen mass transfer quality coefficients were obtained based on the aforementioned ammonia nitrogen stripping prediction results. With the goal of maximizing ammonia nitrogen mass transfer quality, and using the adjustment range of the ammonia nitrogen stripping control parameters as the optimization space, the ammonia nitrogen stripping control parameters are optimized based on the several initial solutions and several ammonia nitrogen mass transfer quality coefficients to obtain the optimal stripping control parameters.

[0230] The construction steps of the ammonia nitrogen stripping predictor include: Based on historical ammonia nitrogen stripping control records, multiple sample ammonia nitrogen stripping control scheme sets were collected, and the historical ammonia nitrogen removal rate, historical ammonia nitrogen removal efficiency, and historical energy consumption per ton of water treatment corresponding to different sample ammonia nitrogen stripping control schemes were obtained as sample ammonia nitrogen stripping results, thus obtaining multiple sample ammonia nitrogen stripping results. Using the set of multiple sample ammonia nitrogen stripping control schemes as input data and the results of multiple sample ammonia nitrogen stripping as supervision data, a deep learning model is trained until convergence to generate an ammonia nitrogen stripping predictor.

[0231] Among them, several ammonia nitrogen mass transfer quality coefficients are obtained based on the aforementioned ammonia nitrogen stripping prediction results, including: The ammonia nitrogen stripping prediction results include the predicted ammonia nitrogen removal rate, the predicted ammonia nitrogen removal percentage, and the predicted energy consumption per ton of water treated. After normalizing the predicted ammonia nitrogen removal rate, predicted ammonia nitrogen removal efficiency, and predicted energy consumption per ton of water treated, a weighted evaluation is performed to obtain the ammonia nitrogen mass transfer quality coefficient. The ammonia nitrogen mass transfer quality coefficient is positively correlated with the predicted ammonia nitrogen removal rate and predicted ammonia nitrogen removal efficiency, and negatively correlated with the predicted energy consumption per ton of water treated.

[0232] The optimization objective is to maximize the mass transfer quality of ammonia nitrogen. The optimization space is defined by the control range of the ammonia nitrogen stripping control parameters. Based on several initial solutions and several ammonia nitrogen mass transfer coefficients, the optimal stripping control parameters are obtained, including: Using the aforementioned initial solutions as the initial population, and based on the aforementioned ammonia-nitrogen mass transfer quality coefficients, the roulette wheel selection method is used to select multiple initial solutions with larger ammonia-nitrogen mass transfer quality coefficients from the initial population as parent individuals. For the parent individuals, a crossover operation is performed with a preset crossover probability to generate offspring individuals; For the offspring individuals, a random perturbation is applied to the coding value of some offspring individuals with a preset mutation probability, and the mutated coding value is constrained within the adjustment range of the ammonia nitrogen stripping control parameter; The offspring individuals are used as a new generation population. The ammonia nitrogen mass transfer quality coefficient corresponding to each individual in the new generation population is calculated. The new generation population replaces the initial population, and the selection, crossover, mutation, and population update steps are repeated. Each time the process is repeated, it is determined whether the current iteration count has reached the preset maximum number of iterations. When the preset maximum number of iterations is reached, the iteration is stopped, and the individual with the largest ammonia nitrogen mass transfer mass coefficient in the current population is analyzed as the optimal stripping control parameter output. The optimal stripping control parameter includes the optimal fan frequency and the optimal influent flow rate.

[0233] The stripping control time window corresponding to the water quality monitoring data sequence is calculated by back-calculating the optimal stripping control parameters, including: Obtain the optimal influent flow rate from the optimal stripping control parameters, and obtain the preset effective liquid holding volume of the stripping tower; Based on the optimal influent flow rate and the effective liquid holding volume of the stripping tower, the expected hydraulic residence time of the wastewater in the stripping tower under the optimal influent flow rate is calculated, and the expected hydraulic residence time is used as the stripping control time window corresponding to the water quality monitoring data sequence.

[0234] In this embodiment of the invention, the blow-off control module 14 is further configured to: Within the stripping control time window corresponding to the current batch of wastewater, according to the adaptive feedback control cycle, the real-time effluent ammonia nitrogen concentration of the stripping tower is periodically obtained, and the real-time effluent ammonia nitrogen concentration is compared with the preset expected effluent ammonia nitrogen concentration to calculate the effluent ammonia nitrogen concentration deviation. Determine whether the deviation of the effluent ammonia nitrogen concentration exceeds the preset tolerance range: If the deviation of the effluent ammonia nitrogen concentration exceeds the preset tolerance range, then the currently executed optimal stripping control parameters are locally fine-tuned according to the direction and magnitude of the deviation of the effluent ammonia nitrogen concentration, and the corrected current stripping control parameters are generated to replace the optimal stripping control parameters for ammonia nitrogen stripping control. If the deviation of the effluent ammonia nitrogen concentration does not exceed the preset tolerance range, the current stripping control parameters shall remain unchanged. The steps for local fine-tuning and correction are as follows: The ratio of the deviation of the effluent ammonia nitrogen concentration to the expected effluent ammonia nitrogen concentration is calculated as the relative deviation rate; When the real-time effluent ammonia nitrogen concentration is greater than the expected effluent ammonia nitrogen concentration, it is determined that the current stripping force is insufficient. The current fan frequency is increased proportionally according to the relative deviation rate, or the current influent flow rate is decreased proportionally according to the relative deviation rate. The absolute value of the adjustment range of the fan frequency is positively correlated with the relative deviation rate, and the absolute value of the adjustment range of the influent flow rate is positively correlated with the relative deviation rate. When the real-time effluent ammonia nitrogen concentration is less than the expected effluent ammonia nitrogen concentration, it is determined that the current stripping force is excessive. The current fan frequency is reduced proportionally according to the relative deviation rate, or the current influent flow rate is increased proportionally according to the relative deviation rate. The absolute value of the adjustment range of the fan frequency is positively correlated with the relative deviation rate, and the absolute value of the adjustment range of the influent flow rate is positively correlated with the relative deviation rate.

[0235] In summary, the wastewater treatment system based on ammonia nitrogen stripping provided in this embodiment of the invention acquires wastewater quality monitoring data sequences in real time through a water quality monitoring module, adaptively generates a feedback control cycle based on water quality changes through a feedback cycle generation module, optimizes stripping control parameters in conjunction with an ammonia nitrogen stripping predictor, determines the stripping control time window based on the optimal stripping control parameters, and then performs closed-loop correction of the stripping control parameters based on the real-time effluent ammonia nitrogen concentration. This achieves dynamic optimization and continuous adjustment of the stripping control parameters, improves the adaptability and operational stability of the stripping control, helps stabilize the effluent ammonia nitrogen concentration, and balances stripping treatment efficiency with operational energy consumption.

[0236] It should be noted that the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0237] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0238] This specification and accompanying drawings are merely illustrative examples of the invention and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its scope. Therefore, if such modifications and modifications fall within the scope of the invention and its equivalents, the invention is intended to include these modifications and modifications.

Claims

1. A wastewater treatment method based on ammonia nitrogen stripping, characterized in that, The methods include: On the main inlet pipe before the wastewater enters the stripping tower, the ammonia nitrogen concentration, pH value and temperature of the wastewater are continuously collected at a preset sampling period to generate a water quality monitoring data sequence. The water quality change coefficient is calculated based on the water quality monitoring data sequence. The water quality change coefficient is used to correct the preset feedback control cycle of effluent ammonia nitrogen concentration and generate an appropriate feedback control cycle. Based on the water quality monitoring data sequence, the ammonia nitrogen stripping control parameters are optimized with the goal of maximizing the mass transfer quality of ammonia nitrogen. The optimal stripping control parameters are obtained, and the stripping control time window corresponding to the water quality monitoring data sequence is calculated by back-calculation using the optimal stripping control parameters. Within the stripping control time window, ammonia nitrogen stripping is controlled according to the optimal stripping control parameters, and the stripping control parameters are iteratively optimized based on the real-time effluent ammonia nitrogen concentration obtained from periodic monitoring, according to the adaptive feedback control cycle.

2. The wastewater treatment method based on ammonia nitrogen stripping according to claim 1, characterized in that, The water quality change coefficient is calculated based on the aforementioned water quality monitoring data sequence, including: The water quality monitoring data sequence includes wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence; The standard deviations of the wastewater ammonia nitrogen concentration sequence, wastewater pH value sequence, and wastewater temperature sequence were calculated respectively to obtain the ammonia nitrogen concentration change index, pH value change index, and temperature change index. The water quality change coefficient is obtained by weighting the ammonia nitrogen concentration change index, pH value change index and temperature change index. The weight values ​​are adaptively determined based on the water quality characteristic state in the current water quality monitoring data sequence, and the sum of the weight values ​​of the three is 1.

3. The wastewater treatment method based on ammonia nitrogen stripping according to claim 1, characterized in that, The preset feedback control cycle for effluent ammonia nitrogen concentration is corrected using the water quality change coefficient to generate an adapted feedback control cycle, including: The ratio of the average historical water quality change coefficient recorded within the historical time range to the water quality change coefficient is used as the periodic correction coefficient; The product of the cycle correction coefficient and the preset feedback control cycle for effluent ammonia nitrogen concentration is rounded down to obtain the adaptive feedback control cycle.

4. The wastewater treatment method based on ammonia nitrogen stripping according to claim 1, characterized in that, Based on the aforementioned water quality monitoring data sequence, the optimal stripping control parameters for ammonia nitrogen stripping are optimized with the goal of maximizing ammonia nitrogen mass transfer quality. These optimal stripping control parameters include: Configure the ammonia nitrogen stripping control parameter adjustment range, wherein the ammonia nitrogen stripping control parameter adjustment range includes the fan frequency adjustment threshold and the influent flow rate adjustment threshold; Within the fan frequency control threshold and the inlet water flow control threshold, any parameter is randomly selected and combined to obtain several initial ammonia nitrogen stripping control parameters, which are used as several initial solutions. The initial solutions are combined one by one with the water quality monitoring data sequence to obtain several ammonia nitrogen stripping control schemes. The several ammonia nitrogen stripping control schemes are input into a pre-constructed ammonia nitrogen stripping predictor for prediction, and several ammonia nitrogen stripping prediction results are output. Several ammonia nitrogen mass transfer quality coefficients were obtained based on the aforementioned ammonia nitrogen stripping prediction results. With the goal of maximizing ammonia nitrogen mass transfer quality, and using the adjustment range of the ammonia nitrogen stripping control parameters as the optimization space, the ammonia nitrogen stripping control parameters are optimized based on the several initial solutions and several ammonia nitrogen mass transfer quality coefficients to obtain the optimal stripping control parameters.

5. The wastewater treatment method based on ammonia nitrogen stripping according to claim 4, characterized in that, The construction steps of the ammonia nitrogen stripping predictor include: Based on historical ammonia nitrogen stripping control records, multiple sample ammonia nitrogen stripping control scheme sets were collected, and the historical ammonia nitrogen removal rate, historical ammonia nitrogen removal efficiency, and historical energy consumption per ton of water treatment corresponding to different sample ammonia nitrogen stripping control schemes were obtained as sample ammonia nitrogen stripping results, thus obtaining multiple sample ammonia nitrogen stripping results. Using the set of multiple sample ammonia nitrogen stripping control schemes as input data and the results of multiple sample ammonia nitrogen stripping as supervision data, a deep learning model is trained until convergence to generate an ammonia nitrogen stripping predictor.

6. The wastewater treatment method based on ammonia nitrogen stripping according to claim 4, characterized in that, Several ammonia nitrogen mass transfer quality coefficients were obtained based on the aforementioned ammonia nitrogen stripping prediction results, including: The ammonia nitrogen stripping prediction results include the predicted ammonia nitrogen removal rate, the predicted ammonia nitrogen removal percentage, and the predicted energy consumption per ton of water treated. After normalizing the predicted ammonia nitrogen removal rate, predicted ammonia nitrogen removal efficiency, and predicted energy consumption per ton of water treated, a weighted evaluation is performed to obtain the ammonia nitrogen mass transfer quality coefficient. The ammonia nitrogen mass transfer quality coefficient is positively correlated with the predicted ammonia nitrogen removal rate and predicted ammonia nitrogen removal efficiency, and negatively correlated with the predicted energy consumption per ton of water treated.

7. The wastewater treatment method based on ammonia nitrogen stripping according to claim 4, characterized in that, With maximizing ammonia nitrogen mass transfer quality as the optimization objective, and using the adjustment range of the ammonia nitrogen stripping control parameters as the optimization space, the ammonia nitrogen stripping control parameters are optimized based on several initial solutions and several ammonia nitrogen mass transfer quality coefficients to obtain the optimal stripping control parameters, including: Using the aforementioned initial solutions as the initial population, and based on the aforementioned ammonia-nitrogen mass transfer quality coefficients, the roulette wheel selection method is used to select multiple initial solutions with larger ammonia-nitrogen mass transfer quality coefficients from the initial population as parent individuals. For the parent individuals, a crossover operation is performed with a preset crossover probability to generate offspring individuals; For the offspring individuals, a random perturbation is applied to the coding values ​​of some offspring individuals with a preset mutation probability, and the mutated coding values ​​are constrained within the adjustment range of the ammonia nitrogen stripping control parameters; The offspring individuals are used as a new generation population. The ammonia nitrogen mass transfer quality coefficient corresponding to each individual in the new generation population is calculated. The new generation population replaces the initial population, and the selection, crossover, mutation, and population update steps are repeated. Each time the process is repeated, it is determined whether the current iteration count has reached the preset maximum number of iterations. When the preset maximum number of iterations is reached, the iteration is stopped, and the individual with the largest ammonia nitrogen mass transfer mass coefficient in the current population is analyzed as the optimal stripping control parameter output. The optimal stripping control parameter includes the optimal fan frequency and the optimal influent flow rate.

8. The wastewater treatment method based on ammonia nitrogen stripping according to claim 1, characterized in that, The stripping control time window corresponding to the water quality monitoring data sequence is calculated by back-calculating the optimal stripping control parameters, including: Obtain the optimal influent flow rate from the optimal stripping control parameters, and obtain the preset effective liquid holding volume of the stripping tower; Based on the optimal influent flow rate and the effective liquid holding volume of the stripping tower, the expected hydraulic residence time of the wastewater in the stripping tower under the optimal influent flow rate is calculated, and the expected hydraulic residence time is used as the stripping control time window corresponding to the water quality monitoring data sequence.

9. A wastewater treatment method based on ammonia nitrogen stripping according to claim 1, characterized in that, According to the aforementioned adaptive feedback control cycle, the stripping control parameters are iteratively optimized based on the real-time effluent ammonia nitrogen concentration obtained from periodic monitoring, including: Within the stripping control time window corresponding to the current batch of wastewater, according to the adaptive feedback control cycle, the real-time effluent ammonia nitrogen concentration of the stripping tower is periodically obtained, and the real-time effluent ammonia nitrogen concentration is compared with the preset expected effluent ammonia nitrogen concentration to calculate the effluent ammonia nitrogen concentration deviation. Determine whether the deviation of the effluent ammonia nitrogen concentration exceeds the preset tolerance range: If the deviation of the effluent ammonia nitrogen concentration exceeds the preset tolerance range, then the currently executed optimal stripping control parameters are locally fine-tuned according to the direction and magnitude of the deviation of the effluent ammonia nitrogen concentration, and the corrected current stripping control parameters are generated to replace the optimal stripping control parameters for ammonia nitrogen stripping control. If the deviation of the effluent ammonia nitrogen concentration does not exceed the preset tolerance range, the current stripping control parameters shall remain unchanged. The steps for local fine-tuning and correction are as follows: The ratio of the deviation of the effluent ammonia nitrogen concentration to the expected effluent ammonia nitrogen concentration is calculated as the relative deviation rate; When the real-time effluent ammonia nitrogen concentration is greater than the expected effluent ammonia nitrogen concentration, it is determined that the current stripping force is insufficient. The current fan frequency is increased proportionally according to the relative deviation rate, or the current influent flow rate is decreased proportionally according to the relative deviation rate. The absolute value of the adjustment range of the fan frequency is positively correlated with the relative deviation rate, and the absolute value of the adjustment range of the influent flow rate is positively correlated with the relative deviation rate. When the real-time effluent ammonia nitrogen concentration is less than the expected effluent ammonia nitrogen concentration, it is determined that the current stripping force is excessive. The current fan frequency is reduced proportionally according to the relative deviation rate, or the current influent flow rate is increased proportionally according to the relative deviation rate. The absolute value of the adjustment range of the fan frequency is positively correlated with the relative deviation rate, and the absolute value of the adjustment range of the influent flow rate is positively correlated with the relative deviation rate.

10. A wastewater treatment system based on ammonia nitrogen stripping, characterized in that, The wastewater treatment method based on ammonia nitrogen stripping according to any one of claims 1 to 9 includes: The water quality monitoring module is used to continuously collect the ammonia nitrogen concentration, pH value and temperature of the wastewater on the main inlet pipe before the wastewater enters the stripping tower at a preset sampling period, and generate a water quality monitoring data sequence. The feedback cycle generation module is used to calculate the water quality change coefficient based on the water quality monitoring data sequence, use the water quality change coefficient to correct the preset feedback control cycle of effluent ammonia nitrogen concentration, and generate an adapted feedback control cycle. The stripping parameter optimization module is used to optimize the ammonia nitrogen stripping control parameters based on the water quality monitoring data sequence, with the goal of maximizing the mass transfer quality of ammonia nitrogen, to obtain the optimal stripping control parameters, and to back-calculate the stripping control time window corresponding to the water quality monitoring data sequence based on the optimal stripping control parameters. The stripping control module is used to control ammonia nitrogen stripping according to the optimal stripping control parameters within the stripping control time window, and to iteratively optimize the stripping control parameters based on the real-time effluent ammonia nitrogen concentration obtained by periodic monitoring according to the adaptive feedback regulation cycle.

Images