Waterworks day-ahead RWKV regulation capability evaluation method considering multiple factors
By collecting multi-factor data and constructing the RWKV model, the problems of insufficient information and modeling accuracy in the day-ahead regulation capacity assessment of water plants were solved, high-precision regulation capacity prediction and curve generation were achieved, and the water plants' ability to participate in power grid dispatch was enhanced.
Patent Information
- Application Number
- CN202510670802.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-23
- Publication Date
- 2025-09-16
AI Technical Summary
The existing methods for evaluating the regulation capacity of water plants on the day before have problems such as a single information source, limited accuracy in equipment power modeling, weak sequence modeling capabilities, and a lack of stability and interpretability in the generation of regulation capacity curves. These problems lead to large deviations in prediction results, insufficient continuity, and insufficient engineering feasibility.
Multi-factor operation data is collected, a refined equipment power calculation model is constructed, and the RWKV model is introduced for long-term series modeling to generate the day-ahead regulation capacity curve of the water plant, which is predicted using multi-factor environmental feature vectors and equipment operation feature sequences.
The accuracy and dynamics of the day-ahead regulation capacity assessment of water plants have been improved, and an upward/downward regulation capacity curve with engineering operability has been generated, providing a scientific basis for power grid dispatching.
Smart Images
Figure CN120654994A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of power systems, and in particular to a method for evaluating the day-ahead RWKV regulation capability of a water plant taking multiple factors into consideration. Background Art
[0002] Assessing a water plant's day-ahead regulation capability helps predict its maximum potential for load regulation within the next day, providing a basis for developing optimized dispatch plans for the power grid. This day-ahead regulation capability assessment enables planned utilization of load-side resources, reduces peak power purchase costs, improves the economic efficiency and operational reliability of the power system, and supports water plants' participation in day-ahead power markets and ancillary service transactions.
[0003] However, existing methods for assessing the day-ahead regulation capacity of water plants have the following shortcomings: 1) Single information source and insufficient input dimensions: Most methods are based solely on historical load data for evaluation, ignoring influencing factors such as temperature, humidity, water quality, water quantity, and operating shifts. This results in incomplete model input information and large deviations in prediction results. 2) Limited accuracy in equipment power modeling: Existing power models are mostly based on empirical formulas, making it difficult to accurately characterize the energy consumption characteristics of equipment under different operating conditions and control strategies, affecting the accurate judgment of the regulation capacity boundary. 3) Weak sequence modeling capabilities and insufficient prediction performance: Conventional machine learning or RNN models have limited ability to process long sequence dependencies, making it difficult to fully learn the temporal patterns and periodic variation characteristics of the day-ahead regulation capacity, affecting the continuity and rationality of the all-day capacity curve. 4) Lack of stability and interpretability in the generation of regulation capacity curves: Currently, there is a lack of a curve generation mechanism for day-ahead scheduling optimization, and the generated regulation capacity sequence lacks continuity, timeliness, and engineering feasibility.
[0004] Therefore, it is necessary to collect multi-factor operation data of the water plant, build a refined equipment power calculation model, realize accurate prediction and curve generation of the regulation potential of the water plant in each period, and improve the accuracy and dynamics of the water plant's regulation capacity assessment in the past few days. Summary of the Invention
[0005] The problem to be solved by the present invention is to provide a method for evaluating the RWKV regulation capacity of a water plant in the day ahead, which takes into account multiple factors. By collecting multi-factor operation data, constructing a refined power calculation model, and introducing the modeling advantage of RWKV for long time series, the regulation potential of the water plant in each time period can be accurately predicted and curve generated, thereby improving the accuracy and dynamism of the regulation capacity evaluation of the water plant in the day ahead.
[0006] The present invention adopts the following technical solution: a method for evaluating the RWKV regulation capacity of a water plant taking into account multiple factors, comprising the following steps:
[0007] Step 1: Multi-factor information collection: Collect key variable information of the water plant, including: raw water volume, water supply demand, rainfall, and ambient temperature, to form a multi-factor environmental feature vector;
[0008] Step 2: Construct a multi-factor water plant equipment power calculation model: Based on the collected multi-factor environmental feature vectors, the operating flow rate of water pumps, dosing equipment, blowers, and sludge pumps is estimated through the operating flow correlation model. The operating power calculation model, combined with physical mechanisms, converts the flow rate into the real-time power of each device, calculates the load rate, and quantifies the transmission process of environmental changes on power load.
[0009] Step 3: Build a water plant's day-ahead regulation capacity assessment model based on RWKV: Using the time series of each device's operating characteristics as input, a gating signal is generated through the Receptance mechanism. Combined with the accumulated Key-Value state, valid historical information is dynamically extracted to predict the water plant's day-ahead regulation capacity, and the regulation capacity prediction results at each moment are output.
[0010] Step 4. Generate the day-ahead regulation capacity curve of the water plant based on RWKV: Generate the day-ahead regulation capacity curve based on the regulation capacity prediction results at each moment output by the water plant's day-ahead regulation capacity assessment model. By aggregating the prediction results of each time slice from 0 to 24 hours, generate the upward regulation capacity curve and the downward regulation capacity curve respectively, describing the distribution of the water plant's flexibility resources in different time periods.
[0011] Preferably, in step 1, multi-factor information of the water plant is collected, and the formula is as follows:
[0012] E(d,t)=[Q raw (d,t),Q demand (d,t),R rain (d,t),T ambient (d,t)]
[0013] Where: E(d,t) is the environmental characteristic input at time t on the dth day, Q raw (d, t) is the raw water volume, Q demand (d,t) is the water supply demand, R rain (d,t) is the rainfall, T ambient (d,t) is the ambient temperature.
[0014] Preferably, in step 2, a water plant operation flow correlation model based on multiple factors is constructed, and the operation flow of water pumps, dosing equipment, blowers, and sludge pumps is calculated according to the water source quantity, demand, rainfall, and temperature changes. Specifically, it includes: a water pump flow correlation model based on multiple factors, a dosing equipment dosage correlation model based on multiple factors, a blower air volume correlation model based on multiple factors, and a sludge pump flow correlation model based on multiple factors. Finally, a water equipment operation quantity feature set X(d,t) is constructed.
[0015] Preferably, in step 2, a water plant operation power calculation model based on multiple factors is constructed to calculate the power and load rate of each device, specifically including: calculating the power and load rate of the water pump, calculating the power and load rate of the dosing equipment, calculating the power and load rate of the blower, and calculating the power and load rate of the sludge pump, and finally obtaining the power and load rate characteristic vector F(d, t) of each device in the water plant.
[0016] Preferably, in step 3, the RWKV-based water plant day-ahead regulation capability assessment model maps the device feature vector F(d, t) at time t to generate a Key vector k(t), a Value vector v(t), and a Receptance gating vector r(t):
[0017]
[0018] Where: k(t) is the Key vector at time t, which extracts features for memory accumulation; v(t) is the Value vector at time t, which contains important feature information; r(t) is the Receptance gate vector at time t, which controls the output ratio, W k is the Key mapping weight matrix, W v is the Value mapping weight matrix, W r is the Receptance mapping weight matrix, σ(·) is the Sigmoid activation function.
[0019] Preferably, the Key and Value are cumulatively updated over time steps, and the new Key and Value at the current moment are weighted and fused with the Key and Value accumulated at the previous moment to control the degree of historical information retention:
[0020] K acc (t) = λK acc (t-1)+k(t)V acc (t) = λV acc (t-1)+v(t)
[0021] Where: K acc (t) is the accumulated Key vector at time t, V acc (t) is the accumulated Value vector at time t, Kacc (t-1) is the key vector accumulated at the previous moment, and λ is the memory time decay factor.
[0022] Preferably, the water plant's regulation capacity is predicted, and the accumulated Value memory V is used to calculate the Receptance gate vector r(t). acc (t) is weighted element by element to obtain the predicted value Y(d,t) of the comprehensive regulation capacity of the water plant at the t-th moment on the d-th day, which is used to represent the regulation capacity vector predicted at the t-th moment on the d-th day.
[0023] Preferably, in step 4, a day-ahead regulation capacity curve of the water plant is generated based on the regulation capacity results at each moment predicted by the RWKV model:
[0024]
[0025] Where: The daily regulation capacity curve of the water plant generated on the dth day includes the capacity curve and the downward regulation capacity curve. d is the predicted target day date, t is the time step index, and T is the total number of time steps in a day.
[0026] The technical solution of the present invention further provides: an electronic device, comprising:
[0027] one or more processors;
[0028] a storage device having one or more programs stored thereon;
[0029] When the one or more programs are executed by the one or more processors, the one or more processors implement any of the above-mentioned methods for evaluating the RWKV regulation capacity of a water plant taking multiple factors into account.
[0030] The technical solution of the present invention also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of any of the above-mentioned methods for evaluating the RWKV regulation capacity of a water plant taking into account multiple factors are implemented.
[0031] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:
[0032] 1. The evaluation method of the present invention collects multi-factor information of the water plant, establishes a data set that integrates multi-dimensional variables such as meteorology, water quality, water quantity, and load, constructs a refined water plant equipment power calculation model based on multiple factors, uses operating parameters and control logic to infer equipment energy consumption, enhances the physical basis for calculating the regulation capacity boundary, and effectively improves the integrity and representativeness of the evaluation input information.
[0033] 2. The evaluation method of the present invention combines the advantages of RWKV in long-sequence modeling. Through the RWKV-based water plant day-ahead regulation capacity evaluation model, high-precision prediction of regulation capacity under complex working conditions is achieved. Through the RWKV-based water plant day-ahead regulation capacity curve production method, the regulation capacity of the entire daily time series is modeled, and an upward / downward regulation capacity curve with engineering operability is generated, which provides a time boundary for the day-ahead scheduling plan and supports its scientific participation in the day-ahead scheduling and ancillary service market. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 This is a flowchart of the method for evaluating the RWKV regulation capacity of a water plant taking multiple factors into account according to the present invention. DETAILED DESCRIPTION
[0035] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the application are further elaborated in detail below with reference to the accompanying drawings. The described embodiments are only a part of the embodiments involved in the present invention. All non-innovative embodiments of other researchers in this field on this embodiment fall within the scope of protection of the present invention. At the same time, the step numbers in the embodiments of the present invention are only set for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.
[0036] In one embodiment of the present invention, a method for evaluating the RWKV regulation capacity of a water plant taking into account multiple factors is provided, such as Figure 1 As shown, it mainly includes: multi-factor information collection technology of water plants, multi-factor based water plant equipment power calculation model, RWKV based water plant day-ahead regulation capacity evaluation model, and RWKV based water plant day-ahead regulation capacity curve production method.
[0037] Step 1: Propose multi-factor information collection of water plants.
[0038] This embodiment addresses the issue of water plant operating power load being affected by multi-source information, and proposes a multi-factor environmental data acquisition technology. By collecting key variable information such as raw water volume, water supply demand, rainfall, and ambient temperature, a standardized environmental feature vector is formed.
[0039] This method can ensure that natural meteorological changes, water source dynamics and changes in water supply demand are uniformly incorporated into the load assessment system, support subsequent equipment operation volume estimation and regulation capacity prediction, and achieve accurate perception and high-frequency collection of external driving factors of power load.
[0040] Step 2: Propose a water plant equipment power calculation model based on multiple factors.
[0041] This embodiment proposes a model for calculating equipment flow and power load based on the collected multi-factor environmental characteristics.
[0042] Firstly, a multi-factor-based water plant operating flow correlation model is proposed. According to the water source quantity, demand, rainfall and temperature changes, the operating flow of water pumps, dosing equipment, blowers and sludge pumps is estimated.
[0043] Then, a water plant operating power calculation model based on multiple factors is proposed. Combined with the physical mechanism model, the flow rate is converted into the real-time power of each device and the load rate is calculated.
[0044] The equipment flow estimation and power load calculation model of this embodiment quantifies the transmission process of environmental changes on power load, laying the foundation for accurately evaluating the adjustable capacity of the water plant under different environmental conditions.
[0045] Step 3: Propose a day-ahead regulation capacity assessment model based on RWKV water plant.
[0046] This embodiment addresses the problem that water plant operating data has strong time-series dependence. A day-ahead regulation capability assessment model based on the RWKV architecture is proposed. This model uses the time series of equipment operating characteristics as input, adopts the Receptance mechanism to generate gating signals, and dynamically extracts valid historical information based on the accumulated Key-Value state to output the predicted regulation capability at each moment.
[0047] By introducing the time-decay memory mechanism, the model takes into account both historical state retention and adaptability to new information, significantly improving the accuracy and temporal consistency of regulatory ability prediction.
[0048] Step 4: Propose a method for producing the day-ahead regulation capacity curve of a water plant based on RWKV.
[0049] This implementation is based on the regulation capacity prediction results output by the RWKV model at each moment, and proposes a method for producing the day-ahead regulation capacity curve. By aggregating the prediction results of each time slice from 0 to 24 hours, an upward regulation capacity curve and a downward regulation capacity curve are generated respectively, which fully describe the distribution of flexibility resources of the water plant in different time periods.
[0050] The method for producing the day-ahead regulation capability curve of the waterworks in this embodiment can provide highly timely and high-resolution decision support for power dispatch optimization, load management, and demand response strategy formulation, thereby improving the response capability of the waterworks in participating in the flexibility regulation of the power system.
[0051] Specifically, for step 1, a multi-factor information collection technology for water plants is proposed to collect key variable information such as raw water volume, water supply demand, rainfall, and ambient temperature:
[0052] E(d,t)=[Qraw (d,t),Q demand (d,t),R rain (d,t),T ambient (d,t)]
[0053] Where: E(d,t) is the environmental characteristic input at time t on the dth day, Q raw (d,t) is the raw water volume (m 3 / h), Q demand (d,t) is the water supply demand (m 3 / h)、R rain (d, t) is rainfall (mm / h), T ambient (d, t) is the ambient temperature (℃).
[0054] Specifically, for step 2, a water plant equipment power calculation model based on multiple factors is proposed.
[0055] First, a multi-factor water plant operating flow correlation model is proposed. According to the water source, demand, rainfall and temperature changes, the operating flow of water pumps, dosing equipment, blowers and sludge pumps is estimated. Specifically,
[0056] Pump flow correlation model based on multiple factors:
[0057] Q pump (d,t)=max(Q raw (d,t),Q demand (d,t))+η rain R rain (d,t)-η evap T ambient (d,t)
[0058] Where: Q pump (d,t) is the water supply flow of the water system at time t on day d (m 3 / h), Q raw (d,t) is the raw water flow (m 3 / h), refers to the amount of raw water extracted from the water source, Q demand (d,t) is the water supply demand (m 3 / h), refers to the water demand supplied to the city or users, R rain (d,t) is the amount of showers (mm / h), which affects the water supply and water quality, T ambient (d, t) is the ambient humidity (℃), high temperature causes the evaporation to increase, η rain is the sensitivity coefficient of shower to the increase of water source (mm -1 ), reflecting the contribution of each millimeter of rainfall to the flow, η evap is the temperature evaporation influence coefficient (℃ -1), reflecting the flow loss caused by each degree of temperature increase.
[0059] Dosage correlation model of dosing equipment based on multiple factors:
[0060] Q dose (d,t)=β dose Q pump (d,t)(1+η′ rain R rain (d,t))
[0061] Where: Q dose (d,t) is the dosing flow rate of the dosing equipment at time t on day d (kg / h), β dose The standard dosing ratio (kg / m 3 ), that is, the dosage of medicine required per cubic meter of water, Q pump (d, t) is the pump flow rate, η′ rain is the dosing adjustment coefficient for water quality deterioration caused by rainfall (mm -1 ), the greater the rainfall, the more medicine needs to be added, R rain (d,t) is the rainfall (mm / h).
[0062] Blower air volume correlation model based on multiple factors:
[0063] Q air (d,t)=β air Q pump (d,t)(1+η temp T ambient (d,t))
[0064] Where: Q air (d,t) is the air volume of the blower at the tth moment on the dth day (Nm 3 / h, standard conditions cubic meters per hour), β air is the standard gas-water ratio (Nm 3 / m 3 ), refers to the standard aeration volume required per unit water volume, Q pump (d,t) is the pump flow rate (m 3 / h), taking the output of the multi-factor water pump flow correlation model as input, η temp is the sensitivity coefficient of temperature to aeration demand changes (℃ -1 ), the temperature rises and the dissolved oxygen decreases, so it is necessary to increase aeration, T ambient (d, t) is the ambient temperature (℃).
[0065] Sludge pump flow correlation model based on multiple factors:
[0066] Q sludge (d,t)=β sludge Qpump (d,t)
[0067] Where: Q sludge (d,t) is the sludge pump flow rate at time t on day d (m 3 / h), β sludge is the sludge production ratio corresponding to unit water supply, Q pump (d,t) is the pump flow rate (m 3 / h), with the output of the pump flow correlation model based on multiple factors as input.
[0068] Get the water supply equipment operation quantity feature set:
[0069] X(d,t)=[Q pump (d,t),Q dose (d,t),Q air (d,t),Q sludge (d,t)].
[0070] Then, a water plant operating power calculation model based on multiple factors is proposed. Combined with the physical mechanism model, the flow rate is converted into the real-time power of each device, and the power and load rate are calculated. Specifically, it includes:
[0071] Calculate the power and load rate of the pump:
[0072]
[0073] Where: P pump (d,t) is the pump power (kW) at the tth moment on the dth day, ρ w is the density of water (kg / m 3 ), g is the acceleration due to gravity (m / s 2 ), H pump (d, t) is the pump head (m), that is, the height difference of the water level raised by the pump, Q pump (d,t) is the pump flow rate at time t on day d (m 3 / h), η pump,mech is the mechanical efficiency of the pump (between 0 and 1), η pump (d,t) is the pump load rate at time t on day d, P pump (d, t) is the current electric power of the water pump (kW), P pump,rated is the rated electrical power of the pump (kW).
[0074] Calculate the power and load rate of the dosing equipment:
[0075] P dose (d,t)=β dose,power ×Q dose (d,t)
[0076]
[0077] Where: P dose (d,t) is the electrical power of the dosing equipment at time t on day d (kW), β dose,power The electric power coefficient corresponding to the unit dosing flow rate (kW / (kg / h)), Q dose (d,t) is the dosing flow rate at time t on day d (kg / h), η dose (d,t) is the dosing equipment load rate at time t on day d, P dose (d, t) is the current power of the dosing equipment (kW), P dose,rated It is the rated electrical power of the dosing equipment (kW).
[0078] Calculate the power and load rate of the blower:
[0079]
[0080] Where: P air (d,t) is the blower power (kW) at the tth moment on the dth day, k air is the air constant (kg / m 3 ), usually 1.2kg / m 3 , Q air (d,t) is the blower air volume at time t on day d (Nm 3 / h)、Δp air (d, t) is the blower air supply pressure difference (Pa), η air,mech is the mechanical efficiency of the blower (between 0 and 1), η air (d,t) is the blower load rate at time t on day d, P air (d, t) is the current electric power of the blower (kW), P air,rated is the blower rated electrical power (kW).
[0081] Calculate the power and load rate of the sludge pump:
[0082]
[0083] Where: P sludge (d, t) is the sludge pump power (kW), ρ sludge is the density of sludge mixture (kg / m 3 ), in this embodiment, 1050-1100 kg / m 3 ; g is the acceleration due to gravity (m / s 2 ), H sludge (d, t) is the sludge pump head (m), Q sludge (d,t) is the sludge pump flow rate (m 3 / h), η sludge,mecis the mechanical efficiency of the sludge pump (between 0 and 1), η sludge (d,t) is the sludge pump load rate at time t on day d, P sludge (d, t) is the current electric power of the sludge pump (kW), P sludge,rated is the rated electrical power of the sludge pump (kW).
[0084] Get the power and load rate characteristic vectors of each equipment in the water plant:
[0085] F(d,t)=[Q pump ,P pump ,η pump ,...]
[0086] Where F(d,t) is the equipment operation feature vector at time t on day d.
[0087] Specifically, for step 3, a model for evaluating the day-ahead regulation capacity of a water plant based on the RWKV model is proposed. This model takes the time series of equipment operating characteristics as input, uses a receptance mechanism to generate gating signals, and dynamically extracts valid historical information based on the accumulated key-value state to output the predicted regulation capacity at each moment. By introducing a time-decaying memory mechanism, the model balances historical state retention with adaptability to new information, significantly improving the accuracy and temporal consistency of regulation capacity predictions.
[0088] The key, value, and receptance generation method based on the RWKV water plant day-ahead regulation capacity assessment model is shown in the following formula. This step maps the device feature vector F(d, t) at time t to generate the key vector k(t), value vector v(t), and receptance gate vector r(t):
[0089]
[0090] Where: k(t) is the Key vector at time t, which extracts features for memory accumulation, v(t) is the Value vector at time t, which contains important feature information, r(t) is the Receptance gate vector at time t, which controls the output ratio, and W k is the Key mapping weight matrix, W v Value mapping weight matrix, W r is the Receptance mapping weight matrix, σ(·) is the Sigmoid activation function.
[0091] In this embodiment, W k 、W v 、W r Both are trainable parameters, and the range of σ(·) is (0,1).
[0092] Furthermore, based on the RWKV water plant's day-ahead regulation capacity assessment model, cumulative updates are performed through Key-Value. This step cumulatively updates the Key and Value over the time step. The new Key and Value at the current moment are weightedly fused with the Key and Value accumulated at the previous moment. λ is the time decay factor that controls the degree of historical information retention:
[0093] K acc (t) = λK acc (t-1)+k(t)V acc (t) = λV acc (t-1)+v(t)
[0094] Where: K acc (t) is the accumulated Key vector at time t, V acc (t) is the accumulated Value vector at time t, K acc (t-1) is the Key vector accumulated at the previous moment, V acc (t-1) is the accumulated Value vector of the previous moment, k(t) is the Key vector of the current moment, v(t) is the Value vector of the current moment, and λ is the memory time decay factor (0<λ<1), which controls the amount of historical memory retention.
[0095] Furthermore, the water plant’s regulation capacity prediction based on RWKV is carried out. This step uses the Receptance gate r(t) to store the accumulated Value memory V acc (t) is weighted element by element to obtain the final predicted value of the comprehensive regulation capacity of the water plant at the tth moment on the dth day Y(d,t):
[0096] Y(d,t)=r(t)V acc (t)
[0097] Where: Y(d,t) is the predicted regulatory capacity vector at time t on day d, which is the element-by-element multiplication (Hadamard product), that is, the corresponding elements are multiplied one by one.
[0098] Specifically, for step 4, a method for producing the day-ahead regulation capacity curve of a water plant based on RWKV is proposed. The formula is as follows: Based on the regulation capacity results at each moment predicted by the RWKV model, the day-ahead regulation capacity curve of the water plant is generated:
[0099]
[0100] Where: is the daily regulation capacity curve of the water plant generated on day d (upward or downward adjustment),
[0101] The unit is usually kW or m3 / h, is the comprehensive adjustment capability vector predicted by the RWKV model at time t on day d, including components such as upward / downward adjustment capability, d is the predicted target day date, t is the time step index, and T is the total number of time steps in a day, which is set according to the time granularity.
[0102] In an embodiment of the present invention, an electronic device is also provided, including: one or more processors; a storage device on which one or more programs are stored; when the one or more programs are executed by the one or more processors, the one or more processors implement the method for evaluating the RWKV regulation capacity of a water plant described in any of the above embodiments.
[0103] In an embodiment of the present invention, a computer-readable storage medium is further provided, on which a computer program is stored. When the program is executed by a processor, the steps of any of the methods for evaluating the RWKV regulation capacity of a water plant in the above embodiments are implemented.
[0104] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for evaluating the RWKV regulation capacity of a water plant taking into account multiple factors, characterized in that: The steps include: Step 1: Multi-factor information collection: Collect key variable information of the water plant, including: raw water volume, water supply demand, rainfall, and ambient temperature, to form a multi-factor environmental feature vector; Step 2: Construct a multi-factor water plant equipment power calculation model: Based on the collected multi-factor environmental feature vectors, the operating flow rate of water pumps, dosing equipment, blowers, and sludge pumps is estimated through the operating flow correlation model. The operating power calculation model, combined with physical mechanisms, converts the flow rate into the real-time power of each device, calculates the load rate, and quantifies the transmission process of environmental changes on power load. Step 3: Build a water plant's day-ahead regulation capacity assessment model based on RWKV: Using the time series of each device's operating characteristics as input, a gating signal is generated through the Receptance mechanism. Combined with the accumulated Key-Value state, valid historical information is dynamically extracted to predict the water plant's day-ahead regulation capacity, and the regulation capacity prediction results at each moment are output. Step 4. Generate the day-ahead regulation capacity curve of the water plant based on RWKV: Generate the day-ahead regulation capacity curve based on the regulation capacity prediction results at each moment output by the water plant's day-ahead regulation capacity assessment model. By aggregating the prediction results of each time slice from 0 to 24 hours, generate the upward regulation capacity curve and the downward regulation capacity curve respectively, describing the distribution of the water plant's flexibility resources in different time periods.
2. The method for evaluating the RWKV regulation capacity of a water plant taking multiple factors into account according to claim 1, characterized in that: In step 1, multi-factor information of the water plant is collected, and the formula is as follows: E(d,t)=[Q raw (d,t),Q demand (d,t),R rain (d,t),T ambient (d,t)] Where: E(d,t) is the environmental characteristic input at time t on the dth day, Q raw (d, t) is the raw water volume, Q demand (d,t) is the water supply demand, R rain (d,t) is the rainfall, T ambient (d,t) is the ambient temperature.
3. The method for evaluating the RWKV regulation capacity of a water plant taking multiple factors into account according to claim 1, characterized in that: In step 2, a multi-factor-based water plant operation flow correlation model is constructed. Based on the water source, demand, rainfall, and temperature changes, the operation flow of water pumps, dosing equipment, blowers, and sludge pumps is estimated. Specifically, the following are the factors: Pump flow correlation model based on multiple factors: Q pump (d,t)=max(Q raw (d,t),Q demand (d,t))+η rain R rain (d,t)-η evap T ambient (d,t) Where: Q pump (d,t) is the water supply flow of the water system at time t on day d, Q raw (d, t) is the raw water flow, Q demand (d,t) is the water supply demand, R rain (d,t) is the amount of shower, T ambient (d,t) is the ambient humidity, η rain is the sensitivity coefficient of showers to the increase in water source volume, η evap is the temperature evaporation influence coefficient; Dosage correlation model of dosing equipment based on multiple factors: Q dose (d,t)=β dose Q pump (d,t)(1+η′ rain R rain (d,t)) Where: Q dose (d,t) is the dosing flow of the dosing equipment at time t on day d, β dose is the standard dosing ratio, Q pump (d,t) is the pump flow rate, η′ rain is the dosing adjustment coefficient for water quality deterioration caused by rainfall, R rain (d,t) is the rainfall; Blower air volume correlation model based on multiple factors: Q air (d,t)=β air Q pump (d,t)(1+η temp T ambient (d,t)) Where: Q air (d,t) is the air volume of the blower at the tth moment on the dth day, β air is the standard gas-water ratio, Q pump (d,t) is the pump flow rate, η temp is the sensitivity coefficient of temperature to aeration demand changes, T ambient (d,t) is the ambient temperature; Sludge pump flow correlation model based on multiple factors: Q sludge (d,t)=β sludge Q pump (d,t) Where: Q sludge (d,t) is the sludge pump flow rate at time t on day d, β sludge is the sludge production ratio corresponding to unit water supply, Q pump (d, t) is the pump flow rate; Construct the water supply equipment operation quantity feature set, which is expressed as: X(d,t)=[Q pump (d,t),Q dose (d,t),Q air (d,t),Q sludge (d,t)]。 4. The method for evaluating the RWKV regulation capacity of a water plant taking multiple factors into account according to claim 3, characterized in that: In step 2, a water plant operating power calculation model based on multiple factors is constructed to calculate the power and load rate of each device, including: Calculate the power and load rate of the pump: Where: P pump (d,t) is the electric power of the water pump at the tth moment on the dth day, ρ w is the density of water, g is the acceleration due to gravity, H pump (d, t) is the pump head, η pump,mech is the mechanical efficiency of the pump, η pump (d,t) is the pump load rate at time t on day d, P pump,rated is the rated electrical power of the water pump; Calculate the power and load rate of the dosing equipment: P dose (d,t)=β dose,power ×Q dose (d,t) Where: P dose (d,t) is the electrical power of the dosing equipment at time t on day d, β dose,power is the electric power coefficient corresponding to the unit dosing flow rate, η dose (d,t) is the dosing equipment load rate at time t on day d, P dose,rated Rated electrical power of the dosing equipment; Calculate the power and load rate of the blower: Where: P air (d,t) is the electric power of the blower at the tth moment on the dth day, k air is the air constant, Δp air (d, t) is the blower air supply pressure difference, η air,mech is the mechanical efficiency of the blower, η air (d,t) is the blower load rate at time t on day d, P air,rated is the rated electrical power of the blower; Calculate the power and load rate of the sludge pump: Where: P sludge (d,t) is the sludge pump power, ρ sludge is the density of sludge mixed liquor, H sludge (d, t) is the sludge pump head, η sludge,mec is the mechanical efficiency of the sludge pump, η sludge (d,t) is the sludge pump load rate at time t on day d, P sludge,rated Rated electrical power of the sludge pump; Calculate the power and load rate characteristic vectors of each device in the water plant: F(d,t)=[Q pump ,P pump ,η pump ,...] Where: F(d,t) is the equipment operation feature vector at time t on day d.
5. The method for evaluating the RWKV regulation capacity of a waterworks taking multiple factors into account according to claim 4, characterized in that: In step 3, the RWKV-based water plant day-ahead regulation capacity assessment model maps the device feature vector F(d,t) at time t to generate the Key vector k(t), the Value vector v(t), and the Receptance gate vector r(t): Where: k(t) is the Key vector at time t, which extracts features for memory accumulation; v(t) is the Value vector at time t, which contains important feature information; r(t) is the Receptance gate vector at time t, which controls the output ratio, W k is the Key mapping weight matrix, W v is the Value mapping weight matrix, W r is the Receptance mapping weight matrix, σ(·) is the Sigmoid activation function.
6. The method for evaluating the RWKV regulation capacity of a waterworks taking multiple factors into account according to claim 5, characterized in that: In step 3, the RWKV-based water plant regulation capacity assessment model performs cumulative updates on the Key and Value in time steps, weightedly integrating the new Key and Value at the current moment with the Key and Value accumulated at the previous moment to control the degree of historical information retention: K acc (t)=λK acc (t-1)+k(t)V acc (t)=λV acc (t-1)+v(t) Where: K acc (t) is the accumulated Key vector at time t, V acc (t) is the accumulated Value vector at time t, K acc (t-1) is the key vector accumulated at the previous moment, λ is the memory time decay factor, 0<λ<1.
7. The method for evaluating the RWKV regulation capacity of a waterworks taking multiple factors into account according to claim 6, characterized in that: In step 3, the water plant’s regulation capacity is predicted, and the accumulated Value memory V is used to calculate the Receptance gate vector r(t). acc (t) is weighted element by element to obtain the predicted value of the comprehensive regulation capacity of the water plant at time t on day d: Y(d,t)=r(t)⊙V acc (t) Where Y(d,t) is the predicted regulatory capacity vector at time t on day d, and ⊙ is the element-by-element multiplication, which represents the Hadamard product.
8. The method for evaluating the RWKV regulation capacity of a water plant taking multiple factors into account according to claim 7, characterized in that: In step 4, the day-ahead regulation capacity curve of the water plant is generated based on the regulation capacity results at each moment predicted by the RWKV model: Where: The daily regulation capacity curve of the water plant generated on the dth day includes the capacity curve and the downward regulation capacity curve. d is the predicted target day date, t is the time step index, and T is the total number of time steps in a day.
9. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method for evaluating the RWKV regulation capability of a water plant taking multiple factors into consideration as described in any one of claims 1 to 8.
10. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the program is executed by a processor, the steps of the method for evaluating the RWKV regulation capacity of a water plant taking into account multiple factors as claimed in any one of claims 1 to 8 are implemented.