Sludge reflux ratio control method and control system thereof

Abstract

The invention relates to a sludge reflux ratio control method. The method comprises the following steps: calculating and outputting a sludge mixed liquid reflux ratio and a sludge reflux ratio through a trained learning algorithm model; calculating the total reflux ratio through a classical algorithm; setting an upper limit reflux threshold value, a lower limit reflux threshold value and a reflux ratio adjustment strategy of the equipment; judging the size relationship between the sludge mixed liquid reflux ratio and the sum of the sludge reflux ratios, the total reflux ratio, the upper limit reflux threshold value and the lower limit reflux threshold value, executing a reflux ratio adjustment strategy according to a judgment result, and outputting a sludge mixed liquid reflux ratio execution value and a sludge reflux ratio execution value; and converting the sludge mixed liquid reflux ratio execution value and the sludge reflux ratio execution value into a sludge mixed liquid reflux amount and a sludge reflux amount, and outputting the sludge mixed liquid reflux amount and the sludge reflux amount to equipment. The invention further discloses a system for executing the method. The result of the learning algorithm model can better fit the actual operation condition of the equipment, intelligent and efficient sewage treatment reflux ratio control is realized, and the stability of the effluent quality is improved.

Description

Technical Field

[0001] The present invention relates to the field of water treatment, and more particularly, to a method and a control system for controlling the sludge reflux ratio. Background Art

[0002] The state has gradually increased the requirements for the total nitrogen concentration in the effluent of sewage treatment plants to address the nitrogen pollution problem in the water environment. Therefore, controlling the total nitrogen emission has become an important task in the field of water treatment. Sewage treatment plants are facing the following dilemmas:

[0003] 1. Most of the process designs of sewage treatment plants are constructed in accordance with the national first-class A standard, that is, the total nitrogen standard for the effluent is 15 mg / L, but the current requirement for the total nitrogen in the effluent has dropped to 6 - 8 mg / L. Referring to the "Code for Design of Outdoor Wastewater Engineering", most urban domestic sewage treatment plants using the AAO process need to basically reach the limit of nitrogen removal of the AAO process if they want to achieve the above-mentioned total nitrogen target value for the effluent.

[0004] 2. The reflux ratio is a key control technology for reducing the total nitrogen concentration in the effluent. The dispatchers of sewage treatment plants mainly rely on experience for extensive regulation. This traditional dispatching method is prone to problems such as high energy consumption, low nitrogen removal efficiency, and large system fluctuations due to the lack of refined data analysis and decision-making guidance, and it is difficult to meet the increasingly strict water quality requirements.

[0005] Therefore, in order to optimize the reflux ratio to obtain better results, in combination with big data analysis and learning algorithm models, the real-time data collected is input into the learning algorithm model to obtain the reflux ratio and output it for controlling the reflux ratio in the equipment. However, the learning algorithm model is the theoretical optimal value, and the actual equipment operates under the limit state and cannot reach the optimal value output by the learning algorithm model, resulting in the difficulty of applying the optimal value output by the learning algorithm model to actual control in actual operation. Summary of the Invention

[0006] In order to overcome the problem that the reflux ratio value output by the learning algorithm model in the above-mentioned prior art exceeds the actual equipment control limit, the present invention provides a method and a system for controlling the sludge reflux ratio, which adjusts the reflux ratio value output by the learning algorithm to enable it to be applied in practice and improve the accuracy of the reflux ratio.

[0007] To solve the above technical problems, the technical solution adopted by the present invention is: a method for controlling the sludge reflux ratio, comprising the following steps:

[0008] Step 1: Calculate and output the sludge mixture reflux ratio and the sludge reflux ratio through a trained learning algorithm model;

[0009] Step 2: Calculate the total reflux ratio through a classical algorithm;

[0010] Step 3: Set the upper limit reflux threshold, lower limit reflux threshold, and reflux ratio adjustment strategy of the equipment; judge the magnitude relationship between the sum of the sludge mixture reflux ratio and the sludge reflux ratio, the total reflux ratio, the upper limit reflux threshold, and the lower limit reflux threshold, and execute the reflux ratio adjustment strategy according to the judgment result, and output the sludge mixture reflux ratio execution value and the sludge reflux ratio execution value;

[0011] The specific reflux ratio adjustment strategy is as follows:

[0012] If the sum of the sludge mixture reflux ratio and the sludge reflux ratio and the total reflux ratio are both greater than the upper limit reflux threshold, then use the upper limit value of the sludge mixture reflux ratio as the sludge mixture reflux ratio execution value, and use the upper limit value of the sludge reflux ratio as the sludge reflux ratio execution value;

[0013] If the sum of the sludge mixture reflux ratio and the sludge reflux ratio is between the upper limit reflux threshold and the lower limit reflux threshold, then calculate and output the sludge mixture reflux ratio and the sludge reflux ratio using the learning algorithm model as the sludge mixture reflux ratio execution value and the sludge reflux ratio execution value;

[0014] If the sum of the sludge mixture reflux ratio and the sludge reflux ratio is greater than the upper limit reflux threshold and the total reflux ratio is between the upper limit reflux threshold and the lower limit reflux threshold, reduce the sludge mixture reflux ratio and the sludge reflux ratio proportionally until their sum is equal to the total reflux ratio, and use the proportionally reduced sludge mixture reflux ratio and sludge reflux ratio as the sludge mixture reflux ratio execution value and the sludge reflux ratio execution value;

[0015] If the sum of the sludge mixture reflux ratio and the sludge reflux ratio is less than the lower limit reflux threshold and the total reflux ratio is between the upper limit reflux threshold and the lower limit reflux threshold, increase the sludge mixture reflux ratio and the sludge reflux ratio proportionally until their sum is equal to the total reflux ratio, and use the proportionally increased sludge mixture reflux ratio and sludge reflux ratio as the sludge mixture reflux ratio execution value and the sludge reflux ratio execution value;

[0016] If the sum of the sludge mixture reflux ratio and the sludge reflux ratio and the total reflux ratio are both less than the lower limit reflux threshold, then use the lower limit value of the sludge mixture reflux ratio as the sludge mixture reflux ratio execution value, and use the lower limit value of the sludge reflux ratio as the sludge reflux ratio execution value;

[0017] Step 4: Convert the sludge mixture reflux ratio execution value and the sludge reflux ratio execution value into the sludge mixture return flow rate and the sludge return flow rate and output them to the equipment.

[0018] The upper limit reflux threshold and the lower limit reflux threshold of the device are the operating limits of the device, and the device cannot operate beyond these values or operate at these values for a long time or frequently. The result of the total reflux ratio calculated by the classical algorithm will meet the operating conditions of the device, be more in line with the actual situation of the device, and only have a small probability of exceeding the upper limit reflux threshold and the lower limit reflux threshold of the device. However, this algorithm cannot directly obtain the return ratio of the sludge mixture and the sludge return ratio to directly let the device execute. Therefore, the two target values need to be adjusted manually according to the actual situation and cannot reach the optimal or the best value of the reflux ratio control, which affects the final water treatment effect. The return ratio of the sludge mixture and the sludge return ratio calculated and output by the learning algorithm model often only consider obtaining the optimal value, so it often causes the output values to exceed the upper limit reflux threshold and the lower limit reflux threshold of the device.

[0019] In the above technical solution, the return ratio of the sludge mixture and the sludge return ratio calculated and output by the learning algorithm model in the adjustment strategy are between the upper limit reflux threshold and the lower limit reflux threshold, indicating that the optimal value can be executed by the device, so the device can be allowed to execute. When the return ratio of the sludge mixture and the sludge return ratio are not between the upper limit reflux threshold and the lower limit reflux threshold, but the total reflux ratio is between the upper limit reflux threshold and the lower limit reflux threshold, the return ratio of the sludge mixture and the sludge return ratio can be adjusted proportionally based on the total reflux ratio. The basis of this adjustment method is the optimal value output by the learning algorithm model, while taking into account the executability of the device, so that the executed values of the return ratio of the sludge mixture and the sludge return ratio obtained after adjustment are in a better situation and can be executed by the device, improving the accuracy of the executed values of the return ratio of the sludge mixture and the sludge return ratio, and improving the treatment efficiency and the stability of the effluent quality.

[0020] The less common situation is that the total reflux ratio is also not between the upper limit reflux threshold and the lower limit reflux threshold. In this case, the device is allowed to operate according to the upper limit value of the return ratio of the sludge mixture, the upper limit value of the sludge return ratio, the lower limit value of the return ratio of the sludge mixture, or the lower limit value of the sludge return ratio. The upper limit value of the return ratio of the sludge mixture, the upper limit value of the sludge return ratio, the lower limit value of the return ratio of the sludge mixture, and the lower limit value of the sludge return ratio are the limit values at which the device can operate. Although this situation is less common, if the total reflux ratio is also not between the upper limit reflux threshold and the lower limit reflux threshold, it means that to obtain the effluent quality that meets the requirements or to maintain the stability of the effluent quality, the device must operate under the limit state.

[0021] Preferably, the learning algorithm model is a random forest model, and the input features include the treated water volume, the influent COD concentration, the influent total nitrogen concentration, the effluent total nitrogen concentration, and the sludge concentration in the biochemical tank. The output results are the return ratio of the sludge mixture and the sludge return ratio.

[0022] Preferably, the construction of the random forest model includes:

[0023] After standardizing the input features using StandardScaler, set various hyperparameter combinations defining the random forest model, and use the MultiOutputRegressor function to wrap it to enable the random forest model to support multi-objective prediction;

[0024] Use grid search and 5-fold cross-validation to obtain the best hyperparameter combination of the random forest model, and reconstruct the random forest model with the best hyperparameter combination.

[0025] Preferably, the training of the random forest model includes:

[0026] Collect historical data, and use the water treatment volume, influent COD concentration, influent total nitrogen concentration, effluent total nitrogen concentration, biochemical pool sludge concentration, sludge return ratio, and sludge mixed liquor return ratio at the same moment as a set of data to construct a dataset;

[0027] Divide the data samples in the dataset into a training set and a validation set;

[0028] The training set is used to fit the random forest model, and the validation set is used to calculate and evaluate the output results. After meeting the evaluation criteria, the training of the random forest model is completed.

[0029] Preferably, based on the root mean square error and R 2 value to calculate and evaluate the output results.

[0030] Preferably, use the interquartile range (IQR) criterion to detect and remove outliers in the dataset.

[0031] Preferably, in step two, the total reflux ratio is specifically:

[0032]

[0033] In the formula, R 总 is the total reflux ratio, %; η is the total nitrogen removal rate; TN 进 is the influent total nitrogen concentration, mg / L; TN 出 is the effluent total nitrogen concentration, mg / L.

[0034] Preferably, the sludge mixed liquor return flow and the sludge return flow are respectively:

[0035] Q 内 = R 内 × Q

[0036] Q 外 = R 外 × Q

[0037] In the formula, Q is the water treatment volume; Q内 is the return flow rate of the sludge mixture; Q 外 is the sludge return flow rate; R 内 is the execution value of the sludge mixture return ratio; R 外 is the execution value of the sludge return ratio.

[0038] A sludge return ratio control system includes:

[0039] A data management module, which is used to collect data at each stage of sewage treatment and store and analyze it;

[0040] A classical algorithm module, which is used to calculate the total return ratio according to the data collected by the data management module;

[0041] A learning algorithm module, which is used to train according to the data collected by the data management module and output the return flow rate of the sludge mixture and the sludge return flow rate;

[0042] A decision-making module, which judges the magnitude relationship between the sum of the return flow rate of the sludge mixture and the sludge return flow rate, the total return ratio, the upper limit return threshold and the lower limit return threshold, and executes the return ratio adjustment strategy according to the judgment result and outputs the execution value of the sludge mixture return ratio and the execution value of the sludge return ratio; and converts the execution value of the sludge mixture return ratio and the execution value of the sludge return ratio into the return flow rate of the sludge mixture and the sludge return flow rate and outputs them;

[0043] An equipment control module, which is used to receive the return flow rate of the sludge mixture and the sludge return flow rate and execute to turn on the internal return pump and the external return pump.

[0044] Further, the equipment control module includes a manual control mode, an automatic control mode and a semi-automatic control mode.

[0045] Compared with the prior art, the beneficial effects of the present invention are: by combining the classical algorithm and the learning algorithm model in the return ratio control, the result of the learning algorithm model can be more in line with the actual operating conditions of the equipment, so that it can be applied in practice, realizing intelligent and efficient sewage treatment return ratio control, improving the treatment efficiency and the stability of the effluent water quality, reducing the process accounting, and reducing the dependence on the personal operation experience of the operator for process regulation. Brief Description of the Drawings

[0046] Figure 1 is a flowchart of a sludge return ratio control method of the present invention. Detailed Embodiments

[0047] The accompanying drawings are only for illustrative purposes and should not be construed as limiting the patent; for better illustration of this embodiment, some components in the drawings may be omitted, enlarged or reduced, which does not represent the size of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted. The positional relationships described in the drawings are only for illustrative purposes and should not be construed as limiting the patent.

[0048] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", "long", "short", etc. indicating the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the positional relationship in the drawings are only for illustrative purposes and should not be construed as limiting the patent. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to specific circumstances.

[0049] The technical solutions of the present invention will be further specifically described below through specific embodiments in conjunction with the accompanying drawings:

[0050] Embodiment 1

[0051] As Figure 1 shown in Embodiment 1 of a method for controlling the sludge return ratio, which includes the following steps:

[0052] Step 1: Calculate and output the sludge mixed liquor return ratio and the sludge return ratio through the trained learning algorithm model; the learning algorithm model can be a neural network model, a random forest model, etc.

[0053] Step 2: Calculate the total return ratio through a classical algorithm, specifically:

[0054]

[0055] In the formula, R 总 is the total return ratio, %; η is the total nitrogen removal rate; TN 进 is the influent total nitrogen concentration, mg / L; TN 出 is the effluent total nitrogen concentration, mg / L.

[0056] Step 3: Set the upper limit return threshold, lower limit return threshold and return ratio adjustment strategy of the device; judge the magnitude relationship between the sum of the sludge mixed liquor return ratio and the sludge return ratio, the total return ratio, the upper limit return threshold and the lower limit return threshold, and execute the return ratio adjustment strategy according to the judgment result and output the sludge mixed liquor return ratio execution value and the sludge return ratio execution value;

[0057] The reflux ratio adjustment strategy is specifically as follows:

[0058] If the sum of the return sludge ratio and the mixed liquor return sludge ratio and the total reflux ratio are all greater than the upper limit reflux threshold, then the upper limit value of the mixed liquor return sludge ratio is used as the execution value of the mixed liquor return sludge ratio, and the upper limit value of the return sludge ratio is used as the execution value of the return sludge ratio;

[0059] If the sum of the mixed liquor return sludge ratio and the return sludge ratio is between the upper limit reflux threshold and the lower limit reflux threshold, then the learning algorithm model is used to calculate and output the mixed liquor return sludge ratio and the return sludge ratio as the execution value of the mixed liquor return sludge ratio and the execution value of the return sludge ratio;

[0060] If the sum of the mixed liquor return sludge ratio and the return sludge ratio is greater than the upper limit reflux threshold and the total reflux ratio is between the upper limit reflux threshold and the lower limit reflux threshold, the mixed liquor return sludge ratio and the return sludge ratio are proportionally reduced until their sum is equal to the total reflux ratio, and the proportionally reduced mixed liquor return sludge ratio and return sludge ratio are used as the execution value of the mixed liquor return sludge ratio and the execution value of the return sludge ratio;

[0061] If the sum of the mixed liquor return sludge ratio and the return sludge ratio is less than the lower limit reflux threshold and the total reflux ratio is between the upper limit reflux threshold and the lower limit reflux threshold, the mixed liquor return sludge ratio and the return sludge ratio are proportionally increased until their sum is equal to the total reflux ratio, and the proportionally increased mixed liquor return sludge ratio and return sludge ratio are used as the execution value of the mixed liquor return sludge ratio and the execution value of the return sludge ratio;

[0062] If the sum of the mixed liquor return sludge ratio and the return sludge ratio and the total reflux ratio are all less than the lower limit reflux threshold, then the lower limit value of the mixed liquor return sludge ratio is used as the execution value of the mixed liquor return sludge ratio, and the lower limit value of the return sludge ratio is used as the execution value of the return sludge ratio;

[0063] Step 4: Convert the execution value of the mixed liquor return sludge ratio and the execution value of the return sludge ratio into the mixed liquor return flow rate and the return sludge flow rate and output them to the equipment.

[0064] The mixed liquor return flow rate and the return sludge flow rate are respectively:

[0065] Q 内 =R 内 ×Q

[0066] Q 外 =R 外 ×Q

[0067] In the formula, Q is the treated water volume; Q 内 is the mixed liquor return flow rate; Q 外 is the return sludge flow rate; R 内 is the execution value of the mixed liquor return sludge ratio; R 外It is the execution value of the sludge return ratio.

[0068] The working principle of this embodiment: The sludge mixed liquor return ratio and the sludge return ratio calculated and output by the learning algorithm model in the adjustment strategy are between the upper limit return threshold and the lower limit return threshold, indicating that the optimal value can be executed by the device, so the device can be allowed to execute. When the sludge mixed liquor return ratio and the sludge return ratio are not between the upper limit return threshold and the lower limit return threshold, but the total return ratio is between the upper limit return threshold and the lower limit return threshold, the sludge mixed liquor return ratio and the sludge return ratio can be adjusted proportionally based on the total return ratio. The basis of such an adjustment method is the optimal value output by the learning algorithm model, while taking into account the executability of the device, so that the obtained sludge mixed liquor return ratio execution value and sludge return ratio execution value are in a relatively good situation and can be executed by the device, improving the accuracy of the sludge mixed liquor return ratio execution value and the sludge return ratio execution value, and improving the treatment efficiency and the stability of the effluent quality.

[0069] The less common situation is that the total return ratio is also not between the upper limit return threshold and the lower limit return threshold. At this time, the device is allowed to operate according to the upper limit value of the sludge mixed liquor return ratio, the upper limit value of the sludge return ratio, the lower limit value of the sludge mixed liquor return ratio, or the lower limit value of the sludge return ratio. The upper limit value of the sludge mixed liquor return ratio, the upper limit value of the sludge return ratio, the lower limit value of the sludge mixed liquor return ratio, and the lower limit value of the sludge return ratio are the limit values at which the device can operate. Although this situation is less common, if the total return ratio is also not between the upper limit return threshold and the lower limit return threshold, it means that to obtain qualified effluent quality or to maintain the stability of the effluent quality, the device must operate under its limit state.

[0070] The beneficial effects of this embodiment: By combining the classical algorithm and the learning algorithm model in the control of the return ratio, the results of the learning algorithm model can be more in line with the actual operating conditions of the device, so that it can be applied in practice, realizing intelligent and efficient sewage treatment return ratio control, improving the treatment efficiency and the stability of the effluent quality, reducing the process accounting, and reducing the dependence of the process regulation on the personal operation experience of the operator on duty.

[0071] Embodiment 2

[0072] An embodiment 2 of a sludge return ratio control system, based on Embodiment 1, the difference from Embodiment 1 is that

[0073] The learning algorithm model is a random forest model, and the input features include the treated water volume, the influent COD concentration, the influent total nitrogen concentration, the effluent total nitrogen concentration, and the biochemical tank sludge concentration, and the output results are the sludge mixed liquor return ratio and the sludge return ratio.

[0074] The construction of the random forest model includes:

[0075] After standardizing the input features using StandardScaler, various hyperparameter combinations for defining the random forest model are set, and the MultiOutputRegressor function is used to wrap the random forest model to support multi-objective prediction;

[0076] The best hyperparameter combination of the random forest model is obtained using grid search and 5-fold cross-validation, and the random forest model is reconstructed with the best hyperparameter combination.

[0077] The training of the random forest model includes:

[0078] Collect historical data, and use the water treatment volume, influent COD concentration, influent total nitrogen concentration, effluent total nitrogen concentration, biochemical pool sludge concentration, sludge mixture return ratio, and sludge return ratio at the same moment as a set of data to construct a dataset; use the interquartile range (IQR) criterion to detect and remove outliers in the dataset.

[0079] Divide the data samples in the dataset into a training set and a validation set;

[0080] The training set is used to fit the random forest model, and the validation set and the output results are used to calculate and evaluate based on the root mean square error and R 2 value. After meeting the evaluation criteria, the training of the random forest model is completed.

[0081] The beneficial effects of this embodiment compared to Embodiment 1: Using the random forest model has better accuracy in calculating the return ratio compared to other learning algorithm models. Under the same database, the evaluation results of the trained learning algorithm models can be seen in the following table, and it can be seen that the random forest model has better accuracy.

[0082]

[0083] The other working principles and effects of this embodiment are the same as those of Embodiment 1.

[0084] Embodiment 3

[0085] An embodiment of a sludge return ratio control system includes:

[0086] A data management module for collecting data at each stage of sewage treatment and storing and analyzing it;

[0087] A classical algorithm module for calculating the total return ratio based on the data collected by the data management module;

[0088] A learning algorithm module for training based on the data collected by the data management module and outputting the sludge mixture return ratio and the sludge return ratio;

[0089] A decision-making module, configured to determine the magnitude relationship among the sum of the sludge mixture return ratio and the sludge return ratio, the total return ratio, the upper limit return threshold, and the lower limit return threshold, and execute a return ratio adjustment strategy according to the determination result, and output the sludge mixture return ratio execution value and the sludge return ratio execution value; and convert the sludge mixture return ratio execution value and the sludge return ratio execution value into the sludge mixture return flow rate and the sludge return flow rate and output them.

[0090] An equipment control module, configured to receive the sludge mixture return flow rate and the sludge return flow rate and execute to turn on the internal return pump and the external return pump until the instantaneous sludge mixture return flow rate detected by the sludge mixture flowmeter reaches the sludge mixture return flow rate value and the instantaneous sludge return flow rate detected by the sludge flowmeter reaches the sludge return flow rate value.

[0091] The sludge return ratio control system of this embodiment can be used to implement the sludge return ratio control method of Embodiment 1 or 2. Among them, the upper limit value and the lower limit value of the sludge mixture return ratio correspond to the operation limits of the internal return pump, and the upper limit value and the lower limit value of the sludge return ratio correspond to the operation limits of the external return pump.

[0092] In this embodiment, the equipment control module includes a manual control mode, an automatic control mode, and a semi-automatic control mode. The automatic control mode among them is that the equipment control module executes the corresponding equipment according to the sludge mixture return flow rate and the sludge return flow rate output by the decision-making module. Manual control is that the staff manually operates the equipment according to experience. The semi-automatic control mode is to display the sludge mixture return flow rate and the sludge return flow rate through a display device such as a display screen, and the staff directly operates the equipment according to the displayed value or after fine-tuning the value.

[0093] Obviously, the above-mentioned embodiments of the present invention are merely examples for clearly explaining the present invention, and are not intended to limit the implementation manners of the present invention. For those of ordinary skill in the art, other different forms of changes or modifications can be made on the basis of the above description. It is not necessary and impossible to list all the implementation manners here. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the claims of the present invention.

Claims

1. A sludge return ratio control method, characterized in that: The steps include: Step 1: Calculate and output the sludge mixed liquor return ratio and the sludge return ratio through the trained learning algorithm model; Step 2: Calculate the total reflux ratio by classical algorithm; Step 3: Set the upper limit reflux threshold, lower limit reflux threshold and reflux ratio adjustment strategy of the equipment; judge the relationship between the sludge mixed liquor reflux ratio and the sum of the sludge reflux ratio, the total reflux ratio, the upper limit reflux threshold and the lower limit reflux threshold, and execute the reflux ratio adjustment strategy according to the judgment result and output the sludge mixed liquor reflux ratio execution value and the sludge reflux ratio execution value; The reflux ratio adjustment strategy is specifically as follows: If the sum of the sludge mixed liquor return ratio and the sludge return ratio and the total return ratio are all greater than the upper limit return threshold, the upper limit value of the sludge mixed liquor return ratio is used as the sludge mixed liquor return ratio execution value, and the upper limit value of the sludge return ratio is used as the sludge return ratio execution value; When the sum of the sludge mixed liquor return ratio and the sludge return ratio is between the upper limit return threshold and the lower limit return threshold, the sludge mixed liquor return ratio and the sludge return ratio are calculated and outputted by the learning algorithm model as the sludge mixed liquor return ratio execution value and the sludge return ratio execution value; The sum of the sludge mixed liquor return ratio and the sludge return ratio is greater than the upper limit return threshold, and the total return ratio is between the upper limit return threshold and the lower limit return threshold. The sludge mixed liquor return ratio and the sludge return ratio are proportionally reduced until the sum of the two is equal to the total return ratio. The sludge mixed liquor return ratio and the sludge return ratio after the proportional reduction are used as the sludge mixed liquor return ratio execution value and the sludge return ratio execution value; The sum of the sludge mixed liquor return ratio and the sludge return ratio is less than the lower limit return threshold, and the total return ratio is between the upper limit return threshold and the lower limit return threshold. The sludge mixed liquor return ratio and the sludge return ratio are increased in proportion until their sum is equal to the total return ratio. The sludge mixed liquor return ratio and the sludge return ratio after the increase in proportion are used as the sludge mixed liquor return ratio execution value and the sludge return ratio execution value; If the sum of the sludge mixed liquor return ratio and the sludge return ratio and the total return ratio are all less than the lower limit return threshold, the lower limit value of the sludge mixed liquor return ratio is taken as the sludge mixed liquor return ratio execution value, and the lower limit value of the sludge return ratio is taken as the sludge return ratio execution value; Step 4: Convert the sludge mixed liquor return ratio execution value and the sludge return ratio execution value into the sludge mixed liquor return volume and the sludge return volume and output them to the device.

2. A sludge return ratio control method according to claim 1, characterized in that: The learning algorithm model is a random forest model. The input features include treated water volume, influent COD concentration, influent total nitrogen concentration, effluent total nitrogen concentration and biochemical pool sludge concentration. The output results are sludge mixed liquor return ratio and sludge return ratio.

3. A sludge return ratio control method according to claim 2, characterized in that: The random forest model construction includes: After using StandardScaler to standardize the input features, set various hyperparameter combinations that define the random forest model, and wrap it with the MultiOutputRegressor function to enable the random forest model to support multi-target prediction; Use grid search and 5-fold cross validation to obtain the best hyperparameter combination of the random forest model, and rebuild the random forest model with the best hyperparameter combination.

4. A sludge return ratio control method according to claim 3, characterized in that: The random forest model training includes: Collect historical data and construct a data set with the treated water volume, influent COD concentration, influent total nitrogen concentration, effluent total nitrogen concentration, biochemical pool sludge concentration, sludge mixed liquor return ratio and sludge return ratio at the same time as a set of data; Divide the data samples in the dataset into training set and validation set; The training set is used to fit the random forest model, and the validation set and output results are used for computational evaluation. Once the evaluation criteria are met, the random forest model training is completed.

5. A sludge return ratio control method according to claim 4, characterized in that: Based on the root mean square error and R 2 The output result is evaluated.

6. A sludge return ratio control method according to claim 4, characterized in that: The interquartile range (IQR) criterion is used to detect and remove outliers in the data set.

7. A sludge return ratio control method according to any one of claims 1 to 6, characterized in that: In step 2, the total reflux ratio is specifically: In the formula, R 总 is the total reflux ratio, %; η is the total nitrogen removal rate; TN 进 is the total nitrogen concentration in the influent, mg / L; TN 出 is the total nitrogen concentration in effluent, mg / L.

8. A sludge return ratio control method according to claim 1, characterized in that: The sludge mixed liquor return volume and sludge return volume are: Q 内 =R 内 ×Q Q 外 =R 外 ×Q In the formula, Q is the amount of water treated; Q 内 is the sludge mixed liquor return volume; Q 外 is the sludge return flow; R 内 R is the execution value of sludge mixed liquor return ratio; 外 It is the execution value of sludge return ratio.

9. A sludge return ratio control system, characterized in that: The method for controlling the sludge return ratio according to any one of claims 1 to 8 comprises: Data management module, used to collect data from each stage of sewage treatment and store and analyze it; The classic algorithm module is used to calculate the total reflux ratio based on the data collected by the data management module; A learning algorithm module, used for training and outputting a sludge mixed liquor return ratio and a sludge return ratio according to data collected by the data management module; A decision module determines the relationship between the sludge mixed liquor return ratio and the sum of the sludge return ratios, the total return ratio, the upper limit return threshold and the lower limit return threshold, and executes the return ratio adjustment strategy according to the judgment result and outputs the sludge mixed liquor return ratio execution value and the sludge return ratio execution value; and converts the sludge mixed liquor return ratio execution value and the sludge return ratio execution value into the sludge mixed liquor return volume and the sludge return volume and outputs them; The equipment control module is used to receive the sludge mixed liquor return flow and the sludge return flow and to start the internal return pump and the external return pump.

10. A sludge return ratio control system according to claim 9, characterized in that: The equipment control module includes a manual control mode, an automatic control mode and a semi-automatic control mode.

Images