Optimized operation method of contaminative rainwater storage system

Abstract

The application discloses a pollution rainwater regulation and storage system optimization operation method, relates to the rainwater regulation and storage and recycling technical field, and comprises the following steps: based on system operation data, respectively calculating a water quantity situation index, a water quality situation index and an energy situation index, and calculating a comprehensive situation index based on the three indexes; using a preset mapping relationship, converting the comprehensive situation index into a regulation and control priority index; constructing a rolling time domain multi-objective constraint decision problem according to the regulation and control priority index and solving the problem to output an operation strategy; and executing the operation strategy to complete the coordinated operation of a pre-tank, a main regulation and storage tank, a reclaimed water plant and photovoltaic grid connection. The application is used to solve the problem that the existing pollution rainwater regulation and storage system is difficult to consider safety, stability and efficient utilization under multi-dimensional target conflicts.

Description

Technical Field

[0001] This invention relates to the field of rainwater storage and reuse technology, and more specifically, to a method for optimizing the operation of a polluted rainwater storage system. Background Technology

[0002] To alleviate pressure on downstream water environments and urban drainage systems, rainwater storage systems are widely constructed and applied. These systems typically consist of a main storage tank, a reclaimed water plant, and photovoltaic energy units. They must not only perform functions such as rainwater peak shaving and flood control and drainage, but also take into account multiple objectives such as pollutant reduction, reclaimed water utilization, and energy self-sufficiency.

[0003] However, the existing operation and scheduling methods of water storage systems still have significant shortcomings. On the one hand, traditional operation strategies mostly rely on fixed empirical rules or single objectives, lacking comprehensive consideration of changes in inflow water quality and quantity, as well as fluctuations in photovoltaic power generation, making it difficult to cope with complex scenarios such as sudden rainstorms, pollutant impacts, and fluctuations in energy supply and demand. On the other hand, even if some systems introduce optimization models, their parameter settings and objective functions are often singular, failing to simultaneously balance safety, stability, and efficiency.

[0004] Furthermore, existing studies often disconnect water quality control from water volume scheduling, and a unified decision-making framework for reclaimed water utilization and energy optimization is lacking. For example, when water levels and pollutant concentrations in storage tanks are high, relying on experience-based outflow strategies may lead to the risk of downstream receiving water bodies exceeding discharge standards. Conversely, when photovoltaic power generation conditions are favorable, the lack of comprehensive optimization of reclaimed water plant load scheduling and energy consumption results in energy waste or increased operating costs. Overall, existing methods fail to achieve unified quantification and dynamic optimization of multi-source data, leading to a lack of scientific rigor and adaptability in system operation. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an optimized operation method for polluted rainwater storage systems. This method calculates and integrates water quantity, water quality, and energy status indices, and combines rolling time-domain multi-objective constraint optimization with collaborative scheduling execution to address the problem that existing polluted rainwater storage systems struggle to balance safety, stability, and efficient utilization under multi-dimensional objective conflicts.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] An optimized operation method for a polluted rainwater storage system includes the following steps: based on system operation data, calculate the water quantity status index, water quality status index, and energy status index respectively, and calculate a comprehensive status index based on the three indices; using a preset mapping relationship, convert the comprehensive status index into a control priority index; construct and solve a rolling time-domain multi-objective constrained optimization model based on the control priority index, and output an operation strategy; execute the operation strategy to complete the coordinated operation of the pre-treatment reservoir, main storage tank, reclaimed water plant, and photovoltaic grid connection.

[0008] In a preferred embodiment, the step of calculating the water quantity status index, water quality status index, and energy status index respectively, and calculating the comprehensive status index based on the three indices, includes: preprocessing the system operation data; obtaining the water quantity status index, water quality status index, and energy status index respectively through a preset calculation logic based on the preprocessed data; and unifying the dimensions of the water quantity status index, water quality status index, and energy status index to obtain the comprehensive status index.

[0009] In a preferred embodiment, the water quality status index is obtained by deriving a water quality status index through a preset calculation logic as follows: the exceedance of major pollutants is determined based on the predicted inflow water quality and flow rate within a rolling time window; the relative deviation of exceedance and the exceedance load per unit time are calculated according to the pollutant limits; and the exceedance is weighted and summed by combining the removal rate of the pre-treatment tank to obtain the water quality status index.

[0010] In a preferred embodiment, the step of converting the comprehensive situation index into a control priority index using a preset mapping relationship includes: the control priority index includes control urgency and benefit weight; the water quantity situation index, water quality situation index, and energy situation index are standardized, and the water quantity urgency, water quality urgency, and energy urgency are calculated using a monotonic mapping function, and the three are weighted and synthesized according to preset coefficients to obtain the control urgency; the three standardized situation indices are non-negatively normalized to obtain a benefit weight vector.

[0011] In a preferred embodiment, the step of constructing a rolling time-domain multi-objective constrained optimization model based on the regulation priority index includes: adjusting the safety margin of the constraints according to the urgency of regulation in the rolling prediction time domain; assigning revenue weights to each objective function to construct a joint objective function; the joint objective function consists of minimizing the water quality exceedance, minimizing the net electricity purchase cost, maximizing the supply of qualified reclaimed water, and minimizing the emergency release volume; setting constraints, including water balance constraints, reservoir capacity and water level constraints, equipment operating capacity constraints, effluent water quality constraints, photovoltaic grid-connected capacity constraints, and outflow control constraints of downstream facility capacity and permitted discharge rate, as well as safe reservoir capacity constraints and pre-release window constraints under rainstorm conditions, wherein the equipment operating capacity constraints are pump station capacity and start-stop interval constraints.

[0012] In a preferred embodiment, the rolling time-domain multi-objective constrained optimization model is solved using a constrained Bayesian optimization algorithm.

[0013] In a preferred embodiment, the step of using the constrained Bayesian optimization algorithm to solve the problem includes: constructing a surrogate estimator for the objective function and a surrogate estimator for the constraint feasibility, respectively outputting the objective prediction statistic and the constraint feasibility probability; constructing a constrained acquisition function based on the feasibility probability and in conjunction with a control priority index, generating and evaluating candidate strategy vectors within the current rolling time window, and obtaining an evaluation result on the satisfaction of the objective value and constraints; updating the surrogate estimator for the objective function and the surrogate estimator for the constraint feasibility according to the evaluation result, and outputting the optimal feasible strategy for the current time window, using the optimal feasible strategy as the starting input for solving the problem in the next time window.

[0014] In a preferred embodiment, before constructing the constrained acquisition function based on the feasibility probability and in conjunction with the control priority index, the method further includes: calculating an environmental prediction fluctuation index, which is obtained by weighting the variance of rainfall prediction and the variance of photovoltaic power generation prediction within a rolling time window; dynamically setting a constraint satisfaction probability threshold based on the environmental prediction fluctuation index; and specifically, generating and evaluating candidate strategy vectors within the current rolling time window involves calculating the objective function acquisition value of the strategy vector only when the predicted feasibility probability of the candidate strategy vector on the constraint feasibility surrogate estimator is higher than the constraint satisfaction probability threshold.

[0015] In a preferred embodiment, the step of dynamically setting the probability threshold for constraint satisfaction based on the environmental prediction volatility index specifically involves: mapping the environmental prediction volatility index to an adjustment coefficient using a preset positive correlation mapping function; and using the adjustment coefficient to correct the basic probability threshold.

[0016] This invention provides an optimized operation system for a polluted rainwater storage system, comprising: a comprehensive situation assessment module, used to calculate a water quantity situation index, a water quality situation index, and an energy situation index based on system operation data, and to calculate a comprehensive situation index based on the three indices; a regulation priority mapping module, used to convert the comprehensive situation index into a regulation priority index using a preset mapping relationship; a multi-objective optimization decision module, used to construct and solve a rolling time-domain multi-objective constrained decision problem based on the regulation priority index, and output an operation strategy; and a decision execution module, used to execute the operation strategy and complete the coordinated operation of the pre-storage reservoir, the main storage tank, the reclaimed water plant, and the photovoltaic grid connection.

[0017] An optimized operation device for a polluted rainwater storage system includes a memory and a processor: the memory is used to store a program; the processor is used to execute the program to implement the various steps of the optimized operation method for the polluted rainwater storage system.

[0018] A readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for optimizing the operation of a polluted rainwater storage system.

[0019] The technical effects and advantages of the optimized operation method for polluted rainwater storage system of the present invention are as follows:

[0020] By calculating and integrating water quantity, water quality, and energy status indices separately, the system's operational status across multiple dimensions can be comprehensively and in real time. A pre-defined mapping relationship is used to convert the integrated status index into a control priority indicator, effectively transforming complex multi-dimensional monitoring data into decision-making weight signals with guiding significance. A rolling time-domain multi-objective constraint optimization model is constructed and solved based on the control priority indicator, enabling dynamically adaptive operational strategies while satisfying multiple constraints. By executing this operational strategy, the coordinated operation of the pre-treatment reservoir, main regulating tank, reclaimed water plant, and photovoltaic grid connection is achieved, effectively ensuring the system's safety, stability, and efficiency under varying rainfall and energy conditions. Attached Figure Description

[0021] Figure 1 A schematic diagram of the optimized operation method of the polluted rainwater storage system provided in the embodiments of the present invention;

[0022] Figure 2 A block diagram illustrating the composition of an optimized operation system for a polluted rainwater storage system provided in this embodiment of the invention;

[0023] Figure 3 A structural block diagram of an exemplary electronic device provided for implementing embodiments of the present disclosure;

[0024] Figure 4 This is a schematic diagram of an exemplary storage medium that can be used to implement embodiments of the present disclosure, as provided in the embodiments of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0026] Example 1, Figure 1The present invention provides an optimized operation method for a polluted rainwater storage system, comprising the following steps:

[0027] S1. Based on system operation data, calculate the water quantity status index, water quality status index, and energy status index respectively, and calculate the comprehensive status index based on the three types of indices.

[0028] It should be noted that the polluted rainwater storage system in this embodiment includes a pre-treatment reservoir, a main storage tank, a reclaimed water plant, a photovoltaic power generation system, and a pumping station and effluent system. The pre-treatment reservoir is used for the initial reduction of pollutants in initial rainwater and highly polluted runoff. The main storage tank is used for rainwater storage. The reclaimed water plant performs advanced treatment on the water in the storage tank. The photovoltaic system provides some energy for operation. System operation data is acquired through online monitoring sensors and a SCADA system, including rainfall, flow rate, water level, pumping station operating status, concentrations of major pollutants (chemical oxygen demand COD, ammonia nitrogen, total phosphorus, etc.), photovoltaic power generation and grid-connected power, and time-of-use electricity price.

[0029] In this embodiment, the calculation of the water quantity status index, water quality status index, and energy status index, and the calculation of the comprehensive status index based on the three indices, includes: preprocessing the system operation data; deriving the water quantity status index, water quality status index, and energy status index respectively through preset calculation logic based on the preprocessed data; and unifying the dimensions of the water quantity status index, water quality status index, and energy status index to obtain the comprehensive status index. Specifically:

[0030] 1. Filter and remove outliers from the monitoring data to form an effective data sequence within the rolling time window;

[0031] 2. The normalization method is used to unify the dimensions of the three types of indices. The formula is as follows:

[0032] (1)

[0033] in, The original index, , These are the historical minimum and maximum values. This is the normalized exponent.

[0034] The water quality status index is obtained through preset calculation logic as follows:

[0035] The regulation pressure is determined based on the predicted inflow and reservoir water level within a rolling time window. Specifically, the predicted inflow is set as follows: Outflow rate is The actual water storage volume of the regulation and storage system is The total designed effective storage capacity of the water storage system is The prediction window is The risk of exceeding storage capacity in the future period. It can be represented as:

[0036] (2)

[0037] in, For the integral time variable within the prediction window, the storage capacity is the sum of the available effective volume of the pre-storage tank and the main storage tank.

[0038] The water volume status index is defined as:

[0039] (3)

[0040] in, Let t be the water volume status index at time t. The weighted coefficients were obtained through least squares fitting. This represents the largest inflow into the water storage system in history. The aforementioned water volume status index... Physically, this characterizes the current hydraulic safety pressure and urgency of water storage faced by the system. Among these, This reflects the risk of overflow due to insufficient storage capacity in the future. This reflects the current instantaneous inflow intensity; The higher the value, the tighter the system's storage capacity and the faster the inflow, resulting in a higher physical risk of overflow.

[0041] The specific steps for calculating the water quality status index are as follows:

[0042] The exceedance of major pollutants is determined based on the predicted inflow water quality and flow rate within a rolling time window; specifically: [The text then lists the major pollutants and their limits, which are not directly related to the preceding sentence about pollutants.] Its predicted concentration is The inflow rate is The pollutant limit is The overload is:

[0043] (4)

[0044] in, Indicates pollutants The load exceeds the standard at time t.

[0045] The relative deviation from the pollutant limit and the excess load per unit time are calculated, and the excess measures are weighted and summed in conjunction with the pre-treatment tank removal rate to obtain the water quality status index. The formula is:

[0046] (5)

[0047] in, Let t be the water quality status index at time t. pollutants The weighting coefficient of the relative deviation exceeding the standard, pollutants The overload weighting coefficient per unit time was obtained by fitting using the least squares method. For pre-treatment of pollutants The removal rate. The water quality status index Physically, this index characterizes the overall pollution load intensity of inflow stormwater. It comprehensively considers the exceedance multiple of pollutant concentration (relative deviation) and the total pollution amount (exceedance load), combined with the facility's treatment capacity (removal rate). The higher the value, the more severe the pollution of the incoming water body, and the greater the risk of environmental impact on downstream receiving water bodies or treatment facilities.

[0048] The specific steps for calculating the energy state index are as follows:

[0049] Based on photovoltaic power generation, pump station operating power, and time-of-use electricity pricing, the system energy consumption pressure is determined, specifically as follows: Let the pump station operating power be... Photovoltaic power generation is The electricity price is Then the net purchased power is:

[0050] (6)

[0051] Electricity load per unit time is:

[0052] (7)

[0053] The formula for calculating the energy state index is:

[0054] (8)

[0055] in, Let be the energy state index at time t. The weighting coefficients are obtained through least squares fitting. , These represent the historical maximum power and maximum electricity load of the pumping station, respectively. The energy status index... Physically, it characterizes the economic cost pressures of system operation and dependence on the external power grid. This index is correlated with net purchased power and real-time electricity costs. The higher the value, the more insufficient photovoltaic absorption is at present and the period of high electricity price, the higher the energy consumption cost of system operation, and the more urgent the need for energy-saving dispatch.

[0056] This step calculates and integrates the water quantity, water quality, and energy status indices separately, and performs rolling predictions and quantitative calculations on the system operation data. This enables a comprehensive perception of the rainwater storage system in terms of water quantity, water quality, and energy status, solving the problems of difficulty in unifying the dimensions of multi-source heterogeneous operation data and the inability to quantify the system status. This provides basic data support and calculation basis for subsequent priority mapping and optimized operation of regulation.

[0057] S2, specifically, maps the comprehensive situation index to a control priority indicator through a control priority mapping model:

[0058] The regulation priority indicators include regulation urgency and benefit weight;

[0059] After standardization, the water quantity status index, water quality status index, and energy status index are obtained through monotonic mapping, and then weighted according to preset coefficients to synthesize the control urgency; specifically:

[0060] First, the three types of situation indices are standardized, as follows: The standardized situation index is obtained by linear normalization mapping to the interval [0,1]. subscript These correspond to water quantity, water quality, and energy status, respectively. After standardization, the water quantity status index, water quality status index, and energy status index are monotonically mapped to obtain the water quantity urgency, water quality urgency, and energy urgency, respectively. These are then weighted and synthesized according to preset coefficients to determine the control urgency. Specifically, for each type of standardized status index... Apply piecewise linear monotonic mapping function Obtain the corresponding urgency The formula is:

[0061] (9)

[0062] in For a monotonically non-decreasing mapping, and These are the urgency threshold and saturation threshold, respectively, satisfying... The urgency of the adjustment is then weighted and synthesized according to preset coefficients, as shown in the formula:

[0063] (10)

[0064] in In order to regulate the urgency, and , Indicates the weight of water urgency. Indicates the urgency weight of water quality, This represents the energy urgency weight, and the weight values ​​are determined through least-squares calibration or analytic hierarchy process (AHP) based on historical samples. Example values ​​are... ;

[0065] The standardized three types of situation indices are normalized using non-negativity to obtain the payoff weight vector. The steps are as follows:

[0066] Define the payoff weight vector as First calculate the sum of the standardized states. ,in To prevent the stabilization constant from having a denominator of zero, the example uses... The non-negative normalization formula for the return weights is:

[0067] (11)

[0068] in and Thus, the payoff weight vector at time t is obtained. It should be noted that in the above formula... and It can be set according to historical quantiles; in the example, it is taken as... and through annual data calibration The value is chosen to match the project's safety, environmental, and economic preferences.

[0069] This step establishes a control priority mapping model and converts the three types of situation indices into problem inputs that can directly drive decision-making under a unified dimension. This realizes the transformation of the quantitative results of system situation into executable control urgency and benefit weights, solving the problem that multi-dimensional situation is difficult to use directly for operational optimization.

[0070] S3: Construct and solve a rolling time-domain multi-objective constraint optimization model based on the control priority index, and output the operation strategy.

[0071] It should be noted that, in this embodiment, the urgency of regulation obtained in step S2 is used. With the profit weight vector As input, a rolling predictive-control framework is adopted; the starting point of the rolling time window is set to... The predicted time domain length is Each step size is controlled to be [number]. Then the prediction time domain of the k-th time window is ; Parameterize the execution strategy into a candidate strategy vector , It consists of discrete decisions and continuous parameters (such as the number of pump stations in operation, gate diversion mode, pre-discharge threshold, set flow rate of each pump, aeration intensity, lower limit of reclaimed water, photovoltaic self-use / grid switching threshold, etc.), and in The optimal feasible strategy for the k-th time window is obtained by solving a multi-objective constrained optimization model. .

[0072] In this embodiment, constructing a rolling time-domain multi-objective constrained optimization model based on the control priority index includes:

[0073] In the rolling forecast time domain, the safety margin of the constraints is adjusted according to the urgency of regulation; specifically, the safety margin is expressed as a monotonic function of the urgency, and the key constraint thresholds are tightened, as shown in the formula:

[0074] (12)

[0075] in To design the highest operating water level, This is the effective upper limit after the urgency level is applied; The permissible limit for pollutant p. Effective limits; The maximum allowable outflow downstream. The effective upper limit; The tightening coefficients are all calibrated using historical data; a pre-leakage lead time is also set. To determine the pre-leakage time window, in which Based on lead time, This is the adjustment coefficient.

[0076] Assign profit weights to each objective function; specifically: according to the profit weight vector described in S2 Through the mapping matrix Convert to a four-dimensional target weight vector The formula is:

[0077] (13)

[0078] in,

[0079] (14)

[0080] This indicates the proportion of water weight allocated to the renewable supply target. This indicates the proportion of water volume weight used for emergency release targets; in the example, it is taken as... .

[0081] The objective function of the rolling time-domain multi-objective constrained optimization model is a joint objective function consisting of minimizing the amount of water quality exceeding standards, minimizing net electricity purchase costs, maximizing the supply of compliant reclaimed water, and minimizing the emergency release volume; specifically, within the time window... Four sub-objectives are defined above, with the following expressions:

[0082] (15)

[0083] (16)

[0084] (17)

[0085] (18)

[0086] in For time t in the strategy Downstream discharge outlets or external discharge branches for pollutants Concentration prediction, For discharge or outflow from receiving water bodies, Pollutant weights (satisfying) ), For time-of-use electricity pricing, For net power consumption forecast, For pump station power prediction, For photovoltaic power generation prediction, To supply flow for reclaimed water, The emergency pre-discharge flow rate is as follows: The joint objective function expression is as follows:

[0087] (19)

[0088] in To place the corresponding sub-targets in the step size The dimensionless cost term, after being split by time granularity or standardized in the [0,1] interval according to historical statistics, ensures that objectives with different dimensions can be weighted.

[0089] The constraints include water balance constraints, reservoir capacity and water level constraints, equipment operating capacity constraints, effluent water quality constraints, photovoltaic grid-connected capacity constraints, as well as outflow control constraints on downstream facility capacity and permitted discharge rates, and safe reservoir capacity constraints and pre-discharge window constraints under heavy rainfall conditions. The equipment operating capacity constraints are pump station capacity and start-stop interval constraints. Specifically:

[0090] 1) Water balance constraint

[0091] For any The water balance constraint is defined as:

[0092] (20)

[0093] in This refers to losses due to evaporation and seepage.

[0094] 2) Reservoir capacity and water level constraints

[0095] (twenty one)

[0096] (twenty two)

[0097] (twenty three)

[0098] in This is the reservoir capacity-water level curve.

[0099] 3) Pump station capacity and start / stop interval constraints

[0100] (twenty four)

[0101] And meet the minimum start-stop interval That is, binary operational variables satisfy Make subsequent continuous It must not be flipped again.

[0102] 4) Effluent water quality constraints

[0103] (25)

[0104] (26)

[0105] If the wastewater is discharged separately from the pre-treatment tank, the discharge limit of the pre-treatment tank shall apply.

[0106] 5) Grid-connected capacity constraints for photovoltaic power

[0107] (27)

[0108] (28)

[0109] in, This represents the photovoltaic power generation absorbed by the system within time t. This represents the grid-connected power output of the system at time t. This indicates the maximum grid-connected capacity allowed by the power grid.

[0110] 6) Outflow control constraints and permitted emission rates

[0111] (29)

[0112] 7) Safety storage capacity and pre-discharge time window constraints

[0113] Pre-discharge window before the arrival of heavy rain forecast Internal permission:

[0114] (30)

[0115] And guarantee:

[0116] (31)

[0117] in, For the k-th rolling prediction time, For at any time The predicted lead time for heavy rain, Available storage capacity Predicting the q-quantile of the inbound quantity (e.g.) ), This is for safety margin.

[0118] Furthermore, the rolling time-domain multi-objective constrained optimization model is solved using a constrained Bayesian optimization method. Specifically, it uses the state context... (Constructed by concatenating features such as situation indices, predictions, and equipment status obtained in steps S1 and S2) and candidate policy vectors Using this as input, a surrogate model is constructed to perform a black-box approximation of the objective and constraints, and... Solve within the internal system.

[0119] Solving rolling time-domain multi-objective constrained optimization models using constrained Bayesian optimization methods also includes:

[0120] Construct a surrogate estimator for the objective function and a surrogate estimator for constraint feasibility, and output the objective prediction statistic and constraint feasibility probability, respectively. Specifically:

[0121] For each sub-objective ( Establish a Gaussian process agent For each constraint Establish a feasibility probability model ;

[0122] Based on this, to address the uncertainties in weather and energy forecasting, this embodiment introduces an environmental forecasting fluctuation index to dynamically adjust the probability threshold for constraint satisfaction, thereby pre-screening candidate strategies before constructing the acquisition function. The specific steps are as follows:

[0123] 1. Calculate the environmental forecast volatility index

[0124] The environmental forecast volatility index is calculated using the variance of rainfall forecast and the variance of photovoltaic power generation forecast within a rolling time window to quantify the uncertainty of the external environment at the current moment. The formula is as follows:

[0125] (32)

[0126] in, and These are the average variances of the rainfall and photovoltaic power prediction sequences within the time window (provided by the weather forecasting system). and The historical maximum variance normalization factor. For weighting coefficients (e.g., take...) ).

[0127] 2. Dynamically set the probability threshold for constraint satisfaction

[0128] Based on the environmental prediction volatility index, a minimum feasible probability that the optimization process must adhere to is dynamically set through a nonlinear mapping. When the predicted volatility is large, the algorithm will automatically increase the requirement for the probability of satisfying the constraints, as defined by the formula:

[0129] (33)

[0130] in, To ensure that dynamic constraints satisfy probability thresholds, The base threshold is set to 0.85 in this embodiment. This is the sensitivity coefficient (set to 2.0 in this embodiment). This formula accounts for environmental fluctuations. The higher the threshold The closer it gets to 1, the more robust the system is forced to choose a more robust strategy.

[0131] 3. Construct a data acquisition function with a filtering mechanism.

[0132] A constrained acquisition function is constructed based on feasibility probability and a control priority index. When generating candidate policy vectors within the current rolling time window, the joint feasibility probability of the policy is first determined. Is it greater than If the value is less than the threshold, the collected value is set to 0 (i.e., the strategy is eliminated); if the value is greater than the threshold, a constrained weighted improved expected collection function is used for evaluation.

[0133] The specific calculation formula for the acquisition function is as follows:

[0134] (34)

[0135] in The average weight of the time window The expected improvement for the i-th objective is given by the formula:

[0136] (35)

[0137] In equation (35) , This is the current optimal observation value. and These are the standard normal distribution function and the density function, respectively.

[0138] By maximizing Generate candidates and in Internal simulation yielded the target value and constraint satisfaction results. ;

[0139] The agent is updated based on the evaluation results, and the optimal feasible strategy for the current time window is output. This optimal feasible strategy serves as the starting input for solving the next time window. Specifically, new samples are incorporated into the training set, and the algorithm is updated. and The posterior distribution is obtained. And satisfy all constraints, As The initial values ​​for a window's warm-start will be used in the next round of scrolling optimization.

[0140] This step constructs a rolling time-domain multi-objective constrained optimization model driven by regulation priority and solves it using constrained Bayesian optimization. This achieves coordinated optimization of operating costs and regeneration supply under hard constraints such as safety, water quality, and grid connection, and solves the problem of difficulty in generating operating strategies online under the conditions of rainstorm uncertainty, target conflict, and dynamic tightening of constraints.

[0141] S4 executes the operation strategy to complete the coordinated operation of the pre-treatment tank, main regulating reservoir, reclaimed water plant and photovoltaic grid connection.

[0142] In this embodiment, the operation strategy is obtained by solving the rolling time-domain multi-objective constrained decision problem in step S3. The specific contents of the operation strategy include pump station scheduling instructions, reservoir capacity allocation scheme, reclaimed water plant treatment load allocation, photovoltaic power generation consumption and grid connection plan, and pre-discharge and emergency discharge arrangements.

[0143] The execution process is as follows: at the beginning of each rolling time window The system control center receives the operation strategy vector. And then distribute it to the corresponding execution unit. Specifically:

[0144] Pre-reservoir scheduling: Based on the inflow diversion instructions in the operation strategy, control the gate opening and pump station start / stop to achieve interception and storage of initial rainwater pollution;

[0145] Scheduling of the main regulating reservoir: Adjust the outflow rate according to the reservoir capacity target allocated by the operation strategy, taking into account the demand for reclaimed water supply, the requirements for pre-discharge of rainstorms, and the control of safe reservoir capacity;

[0146] Reclaimed water plant operation: Based on the allocated water treatment volume and energy consumption constraints according to the operation strategy, adjust the operating load of each treatment unit to achieve the production of qualified reclaimed water;

[0147] Photovoltaic system grid connection: Based on the photovoltaic utilization power and grid connection capacity limit allocated by the operation strategy, the inverter power output is controlled to achieve priority consumption of photovoltaic power generation in the system load and grid connection of surplus power.

[0148] To ensure the executability of the strategy, the system monitors the status of each execution unit in real time and sends the data back to the control center. If the monitoring data deviates significantly from the strategy's expectations, the control center will re-optimize the strategy in the next rolling time window.

[0149] This step directly translates the rolling time-domain optimization strategy into control commands and executes them in each sub-unit of the system, realizing the coordinated scheduling and closed-loop control of the pre-storage tank, main storage tank, reclaimed water plant, and photovoltaic grid connection, thus solving the problem of the lack of dynamic coordinated operation mechanism in traditional rainwater storage systems.

[0150] Example 2, Figure 2 An optimized operation system for polluted rainwater storage is presented, including:

[0151] The comprehensive situation assessment module is used to calculate the water quantity situation index, water quality situation index and energy situation index based on system operation data, and to calculate the comprehensive situation index based on the three indices.

[0152] The control priority mapping module is used to convert the comprehensive situation index into a control priority index using a preset mapping relationship;

[0153] The multi-objective optimization decision module is used to construct and solve rolling time-domain multi-objective constrained decision problems based on the control priority index, and output the operation strategy.

[0154] The decision execution module is used to execute operational strategies and complete the coordinated operation of the pre-treatment tank, main regulating reservoir, reclaimed water plant, and photovoltaic grid connection.

[0155] Example 3,

[0156] An optimized operating device for a polluted rainwater storage system, such as Figure 3 As shown, it includes a memory and a processor: the memory is used to store a program; the processor is used to execute the program to implement any of the embodiments in Example 1.

[0157] Example 4,

[0158] A readable storage medium having a computer program stored thereon, such as Figure 4 As shown, when the computer program is executed by the processor, it implements any of the embodiments in Example 1.

[0159] Since the pollution-prone rainwater storage system optimization operation device described in this embodiment is the same device used to implement the method in Embodiment 1 of this invention, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the method described in Embodiment 1 of this application. Therefore, how the electronic device implements the method in this application embodiment will not be described in detail here. Any device used by those skilled in the art to implement the method in this application embodiment falls within the scope of protection of this application.

[0160] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0161] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0162] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0163] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0164] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0165] In conclusion, 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.

Claims

1. A method for optimizing the operation of a polluted rainwater storage system, characterized in that, The method comprises the following steps: Based on system operation data, water quantity situation index, water quality situation index and energy situation index are calculated respectively to obtain a comprehensive situation index; the water quantity situation index, the water quality situation index and the energy situation index respectively represent system storage safety pressure, pollution load intensity and operation energy consumption cost; The water quantity situation index is determined based on inflow prediction and reservoir capacity water level in a rolling time window, and a reservoir capacity overrun risk is obtained by weighting the reservoir capacity overrun risk and the current inflow; The water quality situation index is determined based on water quality and flow prediction in a rolling time window, and a main pollutant exceeding standard measure is obtained; a relative deviation of exceeding standard and a unit time exceeding standard load are calculated according to a pollutant limit value; and the exceeding standard measure is weighted and summed according to a pre-reservoir removal rate to obtain the water quality situation index; The energy situation index is determined based on photovoltaic power generation power, pump station operation power and time-of-use electricity price, and a system net power purchase power and a unit time electricity charge load are obtained by weighting the net power purchase power and the unit time electricity charge load; The comprehensive situation index is converted into a regulation priority index by using a preset mapping relationship; A rolling time domain multi-objective constraint optimization model is constructed and solved based on the regulation priority index by dynamically adjusting the safety margin of the constraint condition and configuring a weighted combination form of the multi-dimensional objective function, and an operation strategy is output; The operation strategy is executed to coordinate and control the operation of the pre-reservoir, the main storage pool, the reclaimed water plant and the photovoltaic grid.

2. The method of claim 1, wherein, The comprehensive situation index is converted into a regulation priority index by using a preset mapping relationship, which comprises: The regulation priority index comprises a regulation urgency and a benefit weight; The water quantity situation index, the water quality situation index and the energy situation index are standardized and processed, and water quantity urgency, water quality urgency and energy urgency are calculated by using a monotonic mapping function; and the three are weighted and synthesized according to a preset coefficient to obtain the regulation urgency; The three types of situation indexes after standardization are normalized to obtain a benefit weight vector.

3. The method of claim 2, wherein, The rolling time domain multi-objective constraint optimization model is constructed and solved, which comprises: In a rolling prediction time domain, the safety margin of the constraint condition is adjusted according to the regulation urgency; Based on the adjusted constraint condition, a joint objective function weighted by the benefit weight is constructed and solved; The joint objective function is composed of a minimum water quality exceeding standard measure, a minimum net power purchase cost, a maximum standard reclaimed water supply and a minimum emergency discharge volume; The constraint condition comprises a water balance constraint, a reservoir capacity and water level constraint, a device operation capacity constraint, an effluent water quality constraint, a photovoltaic grid capacity constraint and an outflow control constraint.

4. The method of claim 3, wherein the method further comprises: The rolling time domain multi-objective constraint optimization model is solved by using a constraint Bayesian optimization algorithm.

5. The method of claim 4, wherein the optimization is performed by: The constraint Bayesian optimization algorithm is used for solving, which comprises: A target function proxy estimator and a constraint feasibility proxy estimator are constructed to respectively output a target prediction statistic and a constraint feasibility probability; A collection function is constructed based on the feasibility probability and the regulation priority index, and a candidate strategy vector is generated and evaluated in the current rolling time window to obtain an evaluation result of the target value and the constraint satisfaction state; The target function proxy estimator and the constraint feasibility proxy estimator are updated according to the evaluation result; and The rolling time domain multi-objective constraint optimization model is constructed and solved based on the regulation priority index by dynamically adjusting the safety margin of the constraint condition and configuring a weighted combination form of the multi-dimensional objective function, and an operation strategy is output. The optimal feasible strategy of the current time window is based on the updated estimator, and the optimal feasible strategy is taken as the starting input for solving the next time window.

6. The method of claim 5, wherein the optimization is performed by: Before the constructing the acquisition function with constraints based on the feasible probability and in combination with the regulation priority index, the method further comprises: An environmental prediction fluctuation index is calculated, which is obtained by weighting the rainfall prediction variance and the photovoltaic power prediction variance within a rolling time window; Based on the environmental prediction fluctuation index, a constraint satisfaction probability threshold is dynamically set; The generating and evaluating a candidate strategy vector within the current rolling time window specifically comprises: only when the predicted feasible probability of the candidate strategy vector on the constraint feasibility agent estimator is higher than the constraint satisfaction probability threshold, the objective function acquisition value of the strategy vector is calculated.

7. The method of claim 6, wherein the method further comprises: The dynamically setting the constraint satisfaction probability threshold based on the environmental prediction fluctuation index specifically comprises: The environmental prediction fluctuation index is mapped to an adjustment coefficient through a preset positive correlation mapping function; The basic probability threshold is corrected by using the adjustment coefficient, and the correction formula is: wherein, is a dynamic constraint satisfaction probability threshold, is a base threshold, is an environmental prediction volatility index, is a sensitivity coefficient.

8. The method of Claim 1, wherein, The system operation data is obtained through online monitoring sensors and a SCADA system.

Images