A sewage treatment aeration system multi-fan cooperative optimization control method and system

Abstract

The application discloses a kind of sewage treatment aeration system multi-fan collaborative optimization control method, it is related to sewage treatment automatic control and energy-saving optimization technical field.Acquire the aeration demand air volume of current control period, construct the multi-objective optimization model with minimum total power consumption, fan efficiency optimization, maximum stability of output total air volume as target;Improved NSGA II algorithm is solved by using good point set initialization, BLX alpha crossover and dynamic congestion degree calculation, and the Pareto optimal solution set is obtained;Select the solution that satisfies air volume constraint, total power consumption is lowest and fan efficiency variance is minimum as fan control strategy, and issue to controller execution.The application realizes the joint optimization of multiple fan start-stop and frequency, reduces system total power consumption, balances fan load, improves air supply stability under the premise of meeting air demand, Pareto frontier distribution is complete, convergence speed is fast, energy-saving effect is remarkable, and engineering adaptability is strong.

Description

Technical Field

[0001] This invention relates to the field of automatic control and energy-saving optimization technology for wastewater treatment, and in particular to a collaborative optimization control method for multiple blowers in a wastewater treatment aeration system. Background Technology

[0002] In wastewater treatment plant aeration systems, multiple blowers are typically configured to supply air in conjunction with each other to meet the dissolved oxygen requirements of biochemical reactions. In existing technologies, the start-stop combinations and operating frequencies of these blowers largely rely on manual experience or simple logic control (such as fixed-frequency operation and rotating shifts), making it difficult to achieve precise and dynamic airflow matching based on changes in influent load. Existing technologies suffer from the following technical problems: Firstly, the lack of coordinated optimization among multiple blowers often leads to over- or under-supply of air, resulting in energy waste or substandard effluent quality. Secondly, existing optimization methods often focus solely on minimizing total power consumption, neglecting blower load balance and air supply stability, which can easily lead to some blowers operating at high loads for extended periods while others remain idle, affecting equipment lifespan.

[0003] In addition, traditional multi-objective optimization algorithms (such as standard NSGA) II) When dealing with optimization problems involving a mixture of wind turbine start-up and shutdown states (discrete variables) and operating frequency (continuous variables), there are defects such as uneven initial population distribution, insufficient Pareto front diversity, and easy getting trapped in local optima. It is difficult to quickly obtain feasible control strategies that take into account energy saving, stability and load balancing in engineering practice. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a collaborative optimization control method and system for multiple blowers in a wastewater treatment aeration system; On the one hand, a collaborative optimization control method for multiple blowers in a wastewater treatment aeration system is provided, including: Obtain the aeration air volume requirement for the current control cycle; construct a multi-objective optimization model; The improved NSGA-II algorithm is used to solve the multi-objective optimization model to obtain the Pareto optimal solution set; The solution that satisfies the air volume constraint, has the lowest total power consumption, and the smallest variance of the fan efficiency is selected from the Pareto optimal solution set as the fan control strategy. The wind turbine control strategy is sent to the controller for execution, enabling coordinated and optimized control of multiple wind turbines.

[0005] On the other hand, a collaborative optimization control system for multiple blowers in a wastewater treatment aeration system is provided, including: a multi-objective optimization model construction module, used to obtain the aeration demand air volume in the current control cycle and construct a multi-objective optimization model; An improved NSGA-II optimization solution module is used to solve the multi-objective optimization model using the improved NSGA-II algorithm to obtain the Pareto optimal solution set. The optimal solution decision module is used to select from the Pareto optimal solution set the solution that satisfies the air volume constraint, has the lowest total power consumption and the smallest fan efficiency variance, as the fan control strategy. The controller execution module is used to send the wind turbine control strategy to the controller for execution, so as to realize the collaborative optimization control of multiple wind turbines.

[0006] The above technical solution has the following advantages or beneficial effects: This invention constructs a multi-objective optimization model and employs an improved NSGA. The II algorithm can significantly reduce the total system power consumption while meeting the required air volume, balance the load of each operating fan, and improve the stability of air supply; at the same time, it adopts optimal point set initialization and BLX... The calculation of α-crossing and dynamic congestion makes the Pareto front distribution more complete and the convergence speed faster. The resulting control strategy has significant energy-saving effect and good engineering adaptability. Attached Figure Description

[0007] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0008] Figure 1 Here is a flowchart of the multi-objective control for wind turbine collaborative optimization based on the improved NSGA-II in Example 1; Figure 2 The flowchart is for the improved NSGA-II algorithm of Example 1. Detailed Implementation

[0009] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0010] All data acquisition in this embodiment is carried out in accordance with laws and regulations and with user consent, and the data is used legally.

[0011] Example 1 To address the shortcomings of existing wastewater treatment aeration systems, such as insufficient multi-fan coordination, load imbalance and air supply fluctuations caused by single-objective optimization, and the tendency of traditional algorithms to get trapped in local optima and lack diversity when handling mixed variable optimization, this embodiment provides a collaborative optimization control method for multi-fan wastewater treatment aeration systems. Figure 1 As shown, it includes the following steps: S1. Obtain the aeration air volume requirement for the current control cycle; construct a multi-objective optimization model; the model is based on the fan start / stop status. Operating frequency Output air volume The decision variables are determined by minimizing total power consumption, optimizing fan efficiency, and maximizing the stability of total output air volume. The constraints are air volume constraints and fan characteristic constraints.

[0012] Among them, the start-stop status of each wind turbine ; Operating frequency of each wind turbine ; in, Indicates the minimum frequency at which the fan is allowed to operate; Indicates the highest frequency at which the fan is allowed to operate; Minimize total power consumption

[0013] in, , , , is the power consumption fitting coefficient of the i-th wind turbine; M represents the total number of wind turbines participating in the collaborative optimization control; To optimize wind turbine performance, the performance index can be quantified as minimizing the variance of wind turbine efficiency.

[0014] Among them, the efficiency of each operating fan , To minimize the variance in efficiency of each operating fan; Maximize the stability of total output air volume, that is, minimize the deviation between the actual total air volume and the required total air volume:

[0015] in This represents the total air volume required for the current cycle as predicted by the MSTP-Net model.

[0016] Furthermore, in order to ensure the safety of wind turbine operation and the executability of control commands in practical engineering applications, the present invention can further set the following engineering constraints: Minimum number of fans to start: The number of fans in operation during the current control cycle shall not be less than the minimum number of fans required to meet the predicted air volume demand; Standby fan constraint: For aeration systems that adopt a multi-use-one-standby or two-use-one-standby mode, at least one fan shall be kept in standby status. Frequency variation amplitude constraint: The frequency variation of the same fan in adjacent control cycles shall not exceed a preset threshold in order to avoid frequent fluctuations in the inverter output; Start-stop frequency constraint: The number of times the same fan is started and stopped per unit time shall not exceed the preset number, so as to reduce mechanical shock and extend the fan life; Minimum continuous operating time constraint: The continuous operating time of the started fan shall not be less than the preset time when no fault or emergency occurs; Air supply safety margin constraint: The total output air volume retains a preset safety margin on the basis of meeting the predicted demand air volume, in order to cope with short-term fluctuations in water inlet load.

[0017] S2. Solve the multi-objective optimization model using the improved NSGA-II algorithm to obtain the Pareto optimal solution set.

[0018] Specifically, such as Figure 2 As shown, step S2 includes: S201, Parameter initialization: Population initialization is performed using the optimal point set to generate the initial population.

[0019] The parameter initialization includes setting the population size. Maximum number of iterations Crossover probability Probability of mutation BLX-α crossover operator parameters wait.

[0020] The core idea of ​​optimal point set theory is to select a uniformly distributed set of points within a unit cube such that the deviation between the points is minimized. Let... Let be a unit cube in s-dimensional Euclidean space, if Then the optimal point set can be defined as:

[0021] In the formula, For the sample size, For best results, Indicates taking the decimal part; Indicates in generation When the nth sample point, the th The optimal parameters used by the dimension; This indicates the index of the best point, which also corresponds to the first point in the initial population. Individual. Best point. Usually taken ,in It is to satisfy The smallest prime number.

[0022] Mapping the set of optimal points to the search space of decision variables yields the initial population of individuals:

[0023] in, Indicates the first The upper limit of the values ​​that a decision variable can take; Indicates the first The lower bound of the values ​​that a decision variable can take; Indicates the optimal point number; Compared with random initialization, optimal point set initialization has the following advantages: First, the population is evenly distributed in the search space, fully covering the potential optimal region; second, there is no clustering phenomenon, effectively avoiding premature convergence caused by uneven initial population distribution; and third, the initialization process has high determinism, and the algorithm performance has good repeatability. Simulation experiments show that the population initialized with optimal point sets has significantly better uniformity in two-dimensional and three-dimensional spaces than that initialized randomly, laying a good foundation for subsequent evolutionary search.

[0024] S202. Assess the fitness of individuals in the population, calculate the objective function value of individuals in the population, and impose a penalty term on individuals that do not meet the air volume constraint.

[0025] The penalty items are as follows:

[0026] in, This represents the penalty coefficient, used to adjust the intensity of the penalty applied to the objective function value when the airflow is insufficient. The larger the value, the stronger the penalty for individuals who do not meet the airflow constraints.

[0027] S203. Perform rapid non-dominated sorting and dynamic crowding calculation on the population.

[0028] The fast non-fast dominance sort is Pareto hierarchical based on three objective functions.

[0029] The core idea of ​​the dynamic crowding calculation is to use an iterative elimination method to successively select the individual with the highest crowding and update the crowding of the remaining individuals. The specific steps are as follows: For all individuals within the same non-dominated hierarchy, the initial crowding distance is calculated using the standard method; Record the individual with the highest current crowding and move it into the candidate set for the next generation of the population as a reserved individual; Remove the individual from the current level and recalculate the crowding distance of the remaining individuals. During the recalculation, the "space" occupied by the removed individual is released, the relative distance between adjacent individuals changes, and potentially high-quality individuals that were originally covered by dense areas may gain greater crowding. Repeat the above steps until the number of individuals selected from that level meets the required quota.

[0030] S204. The parent population is determined by selection operation, and the offspring population is generated by BLX-α crossover and polynomial mutation or position flip mutation.

[0031] Introducing the BLX-α crossover operator, specifically, let two n-dimensional parent individuals be... and The offspring individuals generated by the BLX-α crossover operator are:

[0032] In the formula, Represents the first parent generation individual In the The values ​​that can be taken on the decision variables; Represents the second parent generation individual In the The values ​​that can be taken on the decision variables; This indicates that the first generation individual is in the [number]th generation. The values ​​that can be taken on the decision variables; This indicates that the second generation individual is in the [number]th generation. The values ​​that can be taken on the decision variables; , A uniformly random number within the interval (0,1). As a control parameter, it is usually taken as , The range of values ​​is This means that offspring individuals can be generated outside the range of parent individuals, thus expanding the search scope.

[0033] The core advantage of the BLX-α operator lies in its hybrid properties: when When the value is in the interval [0,1], the offspring are distributed among the parents, achieving local search; when... When the range exceeds this limit, offspring individuals are distributed outside their parents, achieving global exploration. This adaptive adjustment mechanism enables the algorithm to dynamically balance exploration and development during evolution—initially discovering potential high-quality regions through a large-scale search, and later focusing on local fine-grained search to approximate the true Pareto front; under the same parent individual conditions, the BLX-α operator generates offspring individuals with a significantly larger distribution range than the SBX operator, enabling more effective exploration of boundary regions in the decision space, facilitating the discovery of extreme solutions located at the endpoints of the Pareto front, and thus obtaining a more complete front surface distribution.

[0034] The offspring individuals generated by the BLX-α crossover operator are mutated according to the type of decision variable; for continuous running frequency variables in the offspring individuals... Polynomial mutation is used for perturbation; for discrete start-stop state variables The perturbation is performed using bit flip mutation.

[0035] Specifically, for the first Typhoon generator, if its start / stop state variable If the start-stop mutation probability is satisfied, then let the mutated start-stop state variables be: ;when When, it indicates the first The typhoon generators were shut down or put on standby, and their corresponding operating frequencies were adjusted. ;when When, it indicates the first The typhoon generators are started and put into operation, and their operating frequency is set to meet the requirements. ; For the frequency variable of the fan during operation If the frequency mutation probability is satisfied, then polynomial mutation is used to generate the mutated frequency. and will Corrected to Within the specified range, the frequency variable is ensured to meet the constraints of the wind turbine operating characteristics. If the generated individual after mutation does not meet the air volume constraint, minimum number of start-up units constraint, or standby wind turbine constraint, the start-up and shutdown status and operating frequency are corrected according to the aforementioned engineering constraints to ensure that the generated offspring individuals meet the executable requirements of multi-wind turbine collaborative control. S205. Merge the parent and offspring populations, perform rapid non-dominated sorting and dynamic crowding calculation again, and select populations by level from low to high, and within the same level by dynamic crowding from high to low. Individuals enter the next generation.

[0036] S206. Perform elite selection, iterate to the maximum number of iterations, and output the Pareto optimal solution set.

[0037] The initialization of the optimal point set, the BLX-α crossover operator, and the dynamic congestion calculation are not used independently and superimposed on each other. Instead, they correspond to three key stages in the multi-wind turbine collaborative optimization process and form a progressive combination effect under the condition of joint encoding of start-stop status and operating frequency.

[0038] First, optimal point set initialization is crucial for the initial population generation stage. Since individuals controlling multiple fans simultaneously contain start / stop flags and operating frequencies—where the start / stop flags determine the fan combination and the operating frequency determines the output air volume of each fan—random initialization can easily result in a large number of individuals with insufficient air supply, concentrated frequencies, or a single start / stop combination. Optimal point set initialization ensures that different fan start / stop combinations and their corresponding frequency values ​​are evenly distributed within the feasible region, increasing the proportion of individuals in the initial population that satisfy the total air volume constraint and providing a more complete parent sample for subsequent cross-validation.

[0039] Second, the BLX-α crossover operator operates during the continuous frequency variable search phase. Since the power consumption of the aeration blower typically increases non-linearly with frequency, while the total air volume must meet the predicted air volume constraint, the optimal solution often lies between the low-frequency, high-efficiency operating range and the total air volume constraint boundary. BLX-α crossover allows the offspring frequency to expand its search outside the parent frequency range, enabling the algorithm to break free from the limitations of existing parent frequency combinations and explore operating frequency combinations that still meet the air volume requirement while consuming less power. Simultaneously, single-point or uniform crossover is used for the start / stop flags to ensure that discrete start / stop combinations match the continuous frequency search.

[0040] Third, dynamic congestion calculation is applied to the elite retention stage. Pareto solutions for multi-fan aeration control are typically concentrated near the total airflow constraint boundary. If only traditional congestion calculation is used in a single step, some engineering-significant low-power boundary solutions or load-balancing solutions are easily eliminated. This invention employs a dynamic congestion "selection-elimination-recalculation" approach, progressively updating the distance between solutions within the same non-dominated layer. This ensures that different types of solutions with lower power consumption, smaller airflow deviation, and more balanced loads are retained, thereby preventing the Pareto front from collapsing towards a single target region.

[0041] Through the coordination of the three mechanisms mentioned above, the initialization of the optimal point set provides the algorithm with a uniform and highly feasible initial search foundation. BLX-α crossover expands the effective search range of continuous frequency variables on this basis, and dynamic congestion further maintains the diversity of the distribution of optimization results among the three objectives of power consumption, airflow deviation, and load balancing. The combined effect of these three mechanisms enables the present invention to obtain a more stable and easier-to-execute control strategy under the coupling constraints of multi-fan start-stop-frequency-airflow, rather than simply improving the performance of a single algorithm.

[0042] S3. Select the solution from the Pareto optimal solution set that satisfies the air volume constraint, has the lowest total power consumption, and the smallest fan efficiency variance, and use it as the fan control strategy.

[0043] The decision rule applied to the Pareto optimal solution set is: priority is given to selecting solutions where the airflow deviation is within the allowable range. The solution with the minimum total power consumption is selected; if the total power consumption is similar, the solution with the minimum variance in wind turbine efficiency is selected; the final solution gives the start-stop status and operating frequency of each wind turbine.

[0044] S4. Send the wind turbine control strategy to the controller for execution to achieve multi-wind turbine collaborative optimization control.

[0045] Example 2 Based on the actual parameters of the aeration system of a municipal sewage treatment plant, this embodiment builds a simulation platform to verify the method of Embodiment 1.

[0046] Simulation scene settings: The aeration system is equipped with three centrifugal fans of the same model, and adopts a "two-in-one-out" operation mode, that is, under normal working conditions, only two fans are running at the same time, and the other one is used as a hot standby. Suppose that the total air demand predicted by the aeration model within the current control cycle (2 hours) is: ; Single fan characteristics: Frequency adjustment range 30 Hz ~ 50 Hz; Air volume-frequency relationship (That is, the air volume is 30 m³ / min at a frequency of 30 Hz and 50 m³ / min at 50 Hz). Power consumption polynomial (Unit: kW) Algorithm parameters: Population size Maximum number of iterations Crossover probability Probability of mutation BLX-α parameters Allowable airflow deviation .

[0047] The simulation steps include: S1, Population initialization.

[0048] Generate 60 initial individuals, each encoded as a 6-dimensional vector: ,in Furthermore, at least two units must be started (to satisfy the "two-in-one-out" logic), with a frequency range of 30~50 Hz. The optimal point set method is used to uniformly distribute the initial solution within the feasible region.

[0049] S2. Assess the fitness of individuals in the population and calculate the objective function value of each individual. Calculate total output air volume ; like Imposing punishment ; Calculate total power consumption ; computational efficiency Then calculate the variance of the operating fan efficiency. ; Calculate air volume deviation .

[0050] S3. Perform fast non-dominated sorting and dynamic crowding calculation on the population; determine the parent population through selection operation, and generate the offspring population using BLX-α crossover and polynomial mutation or bit flip mutation; merge the parent population and the offspring population, and perform fast non-dominated sorting and dynamic crowding calculation again; perform elite selection, iterate to the maximum number of iterations, and output the Pareto optimal solution set.

[0051] S4. After obtaining the Pareto frontier, filter according to the decision rules: First, filter out... The solution with the lowest total power consumption is selected from the solutions. If there are multiple solutions with similar power consumption (difference <1%), the solution with the lowest efficiency variance is selected.

[0052]

[0053] S5, Ablation Comparison Experiment

[0054] The method of this invention successfully found the optimal solution where both fans have a frequency of 32.5 Hz, a total air volume of 65 m³ / min, and a total power consumption of 96.7 kW.

[0055] Compared with the "two units operating at 50Hz" solution, the method of this invention saves approximately 18.4% of power consumption (118.5 kW → 96.7 kW) and avoids excessive air supply, which is beneficial to dissolved oxygen stability.

[0056] Compared to traditional power frequency operation, the method of this invention provides air supply on demand, resulting in significant energy savings.

[0057] Under the constraint that the sum of the output air volume of three fans (two in use and one in standby) and the two operating fans is 65 m³ / min, the method of this invention can quickly converge to the Pareto optimal frontier, obtain the optimal control strategy with low power consumption and load balance, and verify the effectiveness and engineering applicability of the algorithm.

[0058] Example 3 This embodiment provides a collaborative optimization control system for multiple blowers in a wastewater treatment aeration system; including: The multi-objective optimization model construction module is used to obtain the aeration demand air volume for the current control cycle and construct a multi-objective optimization model. An improved NSGA-II optimization solution module is used to solve the multi-objective optimization model using the improved NSGA-II algorithm to obtain the Pareto optimal solution set. The optimal solution decision module is used to select from the Pareto optimal solution set the solution that satisfies the air volume constraint, has the lowest total power consumption and the smallest fan efficiency variance, as the fan control strategy. The controller execution module is used to send the wind turbine control strategy to the controller for execution, so as to realize the collaborative optimization control of multiple wind turbines.

[0059] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system, characterized in that, include: Obtain the aeration air volume required for the current control cycle; Construct a multi-objective optimization model; The improved NSGA-II algorithm is used to solve the multi-objective optimization model to obtain the Pareto optimal solution set; The solution that satisfies the air volume constraint, has the lowest total power consumption, and the smallest variance of the fan efficiency is selected from the Pareto optimal solution set as the fan control strategy. The wind turbine control strategy is sent to the controller for execution, enabling coordinated and optimized control of multiple wind turbines.

2. The method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system according to claim 1, characterized in that, The multi-objective optimization model takes the wind turbine start-up and shutdown status as an example. Operating frequency Output air volume The decision variables are determined by minimizing total power consumption, optimizing fan efficiency, and maximizing the stability of total output air volume. The constraints are air volume constraints and fan characteristic constraints.

3. The method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system according to claim 1, characterized in that, The improved NSGA-II algorithm is used to solve the multi-objective optimization model to obtain the Pareto optimal solution set, including: Parameter initialization: Population initialization is performed using a set of optimal points to generate an initial population; Fitness assessment of individuals in the population; calculation of objective function values ​​for individuals in the population. Perform rapid non-dominated sorting and dynamic crowding calculation on the population; The parent population is determined by selection operations, and the offspring population is generated by BLX-α crossover and polynomial mutation or position flip mutation. Merge the parent and offspring populations, and perform fast non-dominated sorting and dynamic crowding calculation again; Perform elite selection, iterate until the maximum number of iterations, and output the Pareto optimal solution set.

4. The method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system according to claim 3, characterized in that, The fitness assessment includes calculating three objective function values: total power consumption, fan efficiency variance, and total air volume deviation, and imposing a penalty term on individuals that do not meet the air volume constraint.

5. The method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system according to claim 3, characterized in that, The dynamic congestion calculation includes: For all individuals within the same non-dominated hierarchy, the initial crowding distance is calculated using the standard method; Record the individual with the highest current crowding and move it into the candidate set for the next generation of the population as a reserved individual; Remove the individual from the current level and recalculate the crowding distance of the remaining individuals. During the recalculation, the "space" occupied by the removed individual is released, the relative distance between adjacent individuals changes, and potentially high-quality individuals that were originally covered by dense areas may gain greater crowding. Repeat the above steps until the number of individuals selected from that level meets the required quota.

6. The method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system according to claim 3, characterized in that, The BLX-α includes, assuming two n-dimensional parent individuals are... and The offspring individuals generated by the BLX-α crossover operator are: In the formula, Represents the first parent individual In the The values ​​that can be taken on the decision variables; Represents the second parent generation individual In the The values ​​that can be taken on the decision variables; This indicates that the first generation individual is in the [number]th generation. The values ​​that can be taken on the decision variables; This indicates that the second generation individual is in the [number]th generation. The values ​​that can be taken on the decision variables; , A uniformly random number within the interval (0,1). As a control parameter, it is usually taken as , The range of values ​​is .

7. The method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system according to claim 3, characterized in that, For continuous operating frequency variables in offspring individuals, polynomial mutation is used for perturbation; for discrete start-stop state variables, bit-flip mutation is used for perturbation.

8. The method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system according to claim 1, characterized in that, The method of initializing the population using a set of optimal points includes: Mapping the set of optimal points to the search space of decision variables yields the initial population of individuals. in, Indicates the first The upper limit of the values ​​that a decision variable can take; Indicates the first The lower bound of the values ​​that a decision variable can take; Indicates the optimal point number; Indicates in generation When the nth sample point, the th The optimal parameters used by the dimensional.

9. The method for coordinated optimization control of multiple blowers in a wastewater treatment aeration system according to claim 1, characterized in that, The solution that satisfies the air volume constraint, has the lowest total power consumption, and the smallest fan efficiency variance is selected from the Pareto optimal solution set. The decision rule is as follows: prioritize the solution with the lowest total power consumption and the lowest air volume deviation; if the total power consumption is similar, select the solution with the smallest fan efficiency variance; the final solution gives the start-stop status and operating frequency of each fan.

10. A collaborative optimization control system for multiple blowers in a wastewater treatment aeration system, characterized in that, include: The multi-objective optimization model construction module is used to obtain the aeration demand air volume for the current control cycle and construct a multi-objective optimization model. An improved NSGA-II optimization solution module is used to solve the multi-objective optimization model using the improved NSGA-II algorithm to obtain the Pareto optimal solution set. The optimal solution decision module is used to select from the Pareto optimal solution set the solution that satisfies the air volume constraint, has the lowest total power consumption and the smallest fan efficiency variance, as the fan control strategy. The controller execution module is used to send the wind turbine control strategy to the controller for execution, so as to realize the collaborative optimization control of multiple wind turbines.

Images