Waterworks XGBoost intra-day regulation capability evaluation method considering process constraint

Abstract

The invention discloses a waterworks XGBoost intraday regulation capability evaluation method considering process constraints, which comprises the following steps of: firstly, setting upper and lower limits and start-stop logic of equipment operation according to a process flow, and forming a dynamic constraint model conforming to an actual operation rule of a waterworks; secondly, calculating the maximum up-regulation and down-regulation capability at each moment by extracting multi-dimensional features such as flow, power and equipment state and combining with equipment start-stop boundary conditions; then, an adjustment capability predictor based on XGBoost is constructed, a nonlinear mapping relation between the equipment state and the adjustment potential is established, and rapid prediction of the adjustment capability under any working condition is realized; and finally, generating an up-regulation and down-regulation capacity curve, carrying out capacity conversion on discrete loads such as sludge dewatering and the like, and realizing continuous expression of the all-day multi-period regulation capacity of the waterworks. The method can provide high-precision and dynamic decision support for power grid dispatching optimization and auxiliary service market response, and has remarkable engineering application value.

Description

Technical Field

[0001] The present invention relates to the field of power systems, and in particular to an XGBoost intraday regulation capability evaluation method for a water plant considering process constraints. Background Art

[0002] Assessing a water plant's intraday regulation capacity helps dynamically understand its adjustable power consumption range during grid load fluctuations, enabling timely participation in load shaving and valley shifting. This intraday regulation capacity assessment can improve power flexibility, optimize dispatch strategies, reduce peak-shaving pressure on the power system, encourage water plant load participation in real-time markets and ancillary service transactions, and enhance the safety and economic efficiency of grid operations.

[0003] The existing water plant regulation capacity assessment methods have the following shortcomings: 1) Process constraints are not fully considered. Existing models mostly use single equipment as the assessment object, ignoring the logic and process dependencies between water pumps, blowers, dosing equipment, dehydration equipment, etc. in the operation of the water plant, resulting in the regulation capacity assessment being divorced from actual operation constraints. 2) It is difficult to characterize the theoretical adjustable boundary. Traditional methods lack the theoretical boundary derivation of the equipment collaborative adjustment range and cannot quantify the potential response capacity under extreme regulation situations. 3) The prediction model has poor robustness. Existing machine learning methods have limited adaptability to the operating environment and historical load changes, making it difficult to accurately predict the intraday regulation capacity under different working conditions. 4) There is a lack of a unified time series aggregation mechanism, and a reasonable regulation capacity aggregation method in the time series dimension has not been established, making it difficult to support the continuous scheduling of load-side resources and the overall response capacity evaluation.

[0004] Therefore, the following improvements are needed when evaluating the regulation capacity of a water plant: clarify the timing logic and operation dependencies between devices through the process constraints of each power equipment; evaluate the maximum upward and downward regulation capacity boundaries while meeting the process continuity conditions; integrate historical data, operating conditions and environmental variables to improve the accuracy of regulation capacity prediction; and realize the integration and dynamic update of load regulation capabilities across time periods and equipment to provide stable flexible load support for power grid dispatching. Summary of the Invention

[0005] The problem to be solved by the present invention is to provide an XGBoost daily regulation capacity evaluation method for a water plant taking into account process constraints, fully consider the process dependencies and operation constraints between various power equipment, construct a theoretical evaluation model close to the actual operation boundary, and realize dynamic prediction and multi-time period integration of the regulation capacity by introducing the XGBoost algorithm and time series aggregation method, so as to improve the evaluation and prediction accuracy of the water plant's daily regulation capacity.

[0006] The present invention adopts the following technical solution: a method for evaluating the intraday regulation capacity of a water plant XGBoost considering process constraints, comprising the following steps:

[0007] Step 1: Construct a process constraint model for each electrical equipment in the water plant: Based on the process flow of each equipment in the water plant, establish process constraint relationships between water pumps, dosing equipment, blowers, sludge dewatering equipment, and lighting equipment. Set the upper and lower operating limits and switching logic of each electrical equipment to form a dynamic constraint model that meets actual operation requirements.

[0008] Step 2: Construct a theoretical daily regulation capacity assessment model for the water plant considering process constraints: Based on the process constraints, extract the flow, power, and state characteristics of each power device, construct a feature vector, and calculate the theoretical regulation potential based on the feature input combined with the operating range and start-stop constraints of each power device;

[0009] Step 3: Construct an XGBoost-based predictor for the water plant's daily regulation capacity: The extracted feature vectors of each device and the corresponding theoretical regulation capacity matrix are used as input and output. The XGBoost regression algorithm is used to train the water plant's daily regulation capacity predictor. The hyperparameters are optimized with the goal of minimizing the prediction error. The nonlinear relationship between the characteristics of each power device and its regulation potential is learned. The upward and downward regulation potential of the equipment is predicted based on the real-time operating conditions, and the daily regulation capacity is adaptively estimated under the operating condition.

[0010] Step 4. Construct a time series aggregation model for the water plant's intraday regulation capacity: Based on the single-moment regulation capacity results output by the water plant's intraday regulation capacity predictor, aggregate them in chronological order to form the water plant's intraday regulation capacity curve, and consider the conversion of discrete resource capacity to obtain an evaluation result describing the changes in the water plant's adjustable capacity at different time periods throughout the day.

[0011] Preferably, in the process constraint model of each power equipment in the water plant described in step 1:

[0012] Construct a water pump flow balance constraint and set the water output of the pump to meet the requirements of raw water transportation and process treatment at the same time. The formula is as follows:

[0013] Q pump (t) = Q raw (t)+Q process (t)

[0014] Where: Q pump (t) is the total water flow of the pump, Q raw (t) is the raw water delivery flow rate, Q process (t) is the internal process water flow rate (m 3 / h);

[0015] Construct linkage constraints for dosing equipment, set the dosing dosage of the dosing equipment to be proportional to the water supply flow of the pump, and the proportional coefficient is the dosing concentration requirement. The formula is as follows:

[0016] Qdose (t) = D ratio (t)×Q pump (t)

[0017] Where: Q dose (t) is the dosage of the dosing equipment, D ratio (t) is the ratio of dosage to water volume, Q pump (t) is the total water flow rate supplied by the pump;

[0018] Construct a blower air-water linkage constraint and set a fixed proportional relationship between the blower air volume and the process water flow rate. The formula is as follows:

[0019] Q air (t) = k air-water ×Q process (t)

[0020] Where: Q air (t) is the air output of the blower, k air-water is the gas-water ratio coefficient, Q process (t) is the internal process water flow rate;

[0021] Construct the start-stop synchronization constraint of the sludge dewatering equipment and set the start-stop of the sludge dewatering equipment to be synchronized or delayed with the operating status of the sludge pump and blower. The formula is as follows:

[0022] u dewater (t)≥u sludge,pump (t) and u dewater (t)≥u air (t)

[0023] Where: u dewater (t) is the switch status of the sludge dewatering equipment, u sludge,pump (t) is the sludge pump switch state, u air (t) is the blower switch status;

[0024] Construct auxiliary linkage constraints for lighting equipment, setting the lighting equipment to automatically start according to the operating status of production equipment, including water pumps, dosing equipment, blowers, and dehydration equipment. The formula is as follows:

[0025] u light (t)≥max(u pump (t),u dose (t),u air (t),u dewater (t))

[0026] Where: u light (t) is the lighting equipment switch state, u pump (t) is the water pump switch state, u dose(t) is the switch status of the dosing equipment, u dewater (t) is the switch status of the sludge dewatering equipment, u air (t) is the blower switch status.

[0027] Preferably, in step 2, the flow, power, and state features of each power device are extracted based on the process constraints of the water plant to construct a feature vector. The features include: flow, power, switch status of water pumps, dosing equipment, blowers, sludge pumps, and sludge dewatering equipment, as well as auxiliary lighting status, ambient temperature, and humidity. The formula is as follows:

[0028]

[0029] Where: Q pump (t) is the pump flow rate, P pump (t) is the pump power, u pump (t) is the start and stop state of the water pump, Q dose (t) is the dosing flow rate of the dosing equipment, P dose (t) is the power of the dosing equipment, u dose (t) is the start and stop status of the dosing equipment, Q air (t) is the air volume of the blower, P air (t) is the blower power, u air (t) is the blower start and stop state, Q sludge (t) is the sludge pump flow rate, P sludge (t) is the sludge pump power, u sludge (t) is the start and stop state of the sludge pump, P dewater (t) is the power of sludge dewatering equipment, u dewater (t) is the start and stop state of the sludge dewatering equipment, P light (t) is the lighting system power, u light (t) is the start and stop state of the lighting system, T ambient (t) is the ambient temperature, H ambient (t) is the ambient humidity.

[0030] Preferably, in step 2, based on the feature input and combined with the operating range and start-stop constraints of each power equipment, the theoretical regulation potential under the current operating state is calculated, including: water pump daily upward regulation capacity assessment, water pump daily downward regulation capacity assessment, dosing equipment upward regulation capacity assessment, dosing equipment downward regulation capacity assessment, blower upward regulation capacity assessment, blower downward regulation capacity assessment, sludge pump upward regulation capacity assessment, sludge pump downward regulation capacity assessment, sludge dewatering equipment switch switching regulation capacity assessment, lighting equipment switch switching regulation capacity assessment; the theoretical regulation capacity of the water plant is comprehensively obtained to form a moment-level theoretical regulation capacity matrix as the predictor training target.

[0031] Preferably, in step 3, a water plant regulation capacity predictor is constructed based on the XGBoost regression model, with the power equipment operation feature vector as input and the theoretical regulation capacity calculated in step 2 as the supervision label. The optimal parameter Θ is obtained by training through the objective function of minimizing the prediction error and model complexity.

[0032] Preferably, in step 3, the hyperparameters of the XGBoost model include: tree depth and learning rate; through training, the nonlinear relationship between the characteristics of each power equipment and the regulation potential is fully learned, and a model is obtained that predicts the equipment's potential for adjustment up and down according to real-time operating conditions, and an adaptive estimation of the intraday regulation capacity under the perception of the water plant's operating conditions is performed.

[0033] Preferably, in step 4, the time series aggregation model of the water plant's intraday regulation capacity performs time series aggregation on the regulation capacity results output by the XGBoost predictor at all times throughout the day to generate the water plant's intraday regulation capacity curve, and separately counts the upward and downward regulation potentials of water pumps, blowers, and dehydration equipment, and incorporates switching resources such as sludge dehydration equipment into the total capacity after equivalent conversion, outputs the upward regulation capacity curve and the downward regulation capacity curve, and reflects the water plant's support potential for the power grid in different time periods.

[0034] The technical solution of the present invention further provides: an electronic device, comprising:

[0035] one or more processors;

[0036] a storage device having one or more programs stored thereon;

[0037] 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 intraday regulation capacity of a water plant XGBoost considering process constraints.

[0038] 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 intraday regulation capacity of a water plant XGBoost considering process constraints are implemented.

[0039] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0040] 1. The XGBoost daily regulation capacity assessment method for water plants in the present invention fully considers the process dependencies and operational constraints between various power equipment. By constructing a theoretical assessment model close to the actual operating boundaries, it comprehensively improves the assessment and prediction accuracy of the daily regulation capacity of water plants.

[0041] 2. The present invention introduces the XGBoost algorithm and time series aggregation method to achieve dynamic prediction and multi-period integration of regulation capabilities, providing sustainable and controllable flexible load resource support for power grid dispatching. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flowchart of the XGBoost daily regulation capacity evaluation method for a water plant considering process constraints in the present invention. DETAILED DESCRIPTION

[0043] 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.

[0044] In one embodiment of the present invention, an XGBoost intraday regulation capacity evaluation method for a water plant considering process constraints mainly includes: a process constraint model of each power equipment in the water plant, a theoretical intraday regulation capacity evaluation model of the water plant considering process constraints, an XGBoost-based intraday regulation capacity predictor of the water plant, and a time series aggregation model of the intraday regulation capacity of the water plant.

[0045] like Figure 1 As shown, the XGBoost daily regulation capacity evaluation method of a water plant considering process constraints in this embodiment includes the following steps:

[0046] Step 1: Construct the process constraint model of each power equipment in the water plant.

[0047] First, based on the process flow of each piece of waterworks equipment, establish process constraints between water pumps, dosing equipment, blowers, sludge dewatering equipment, and lighting. It's clear that water pump flow determines the dosing ratio, blower air volume is linked to process water flow, sludge dewatering requires synchronization with sludge pump and aeration operations, and lighting equipment starts up simultaneously with the main equipment.

[0048] At the same time, the operation sequence logic and coupling relationship between the equipment are clarified, and the upper and lower operating limits and switch start and stop logic of electrical equipment such as water pumps and blowers are set to form a dynamic constraint model that meets the actual operation requirements, laying the foundation for subsequent regulation capability evaluation.

[0049] Step 2: Propose a theoretical daily regulation capacity evaluation model for a water plant considering process constraints.

[0050] Based on process constraints, we extract multidimensional features such as flow rate, power, and status for each device to construct a feature vector. Based on these feature inputs and combined with the device's operating range and start / stop constraints, we calculate the theoretical maximum upward and downward adjustment capabilities. Using specific formulas to quantify the adjustable range of each device under current operating conditions, we develop a moment-by-moment theoretical adjustment capacity matrix. This matrix fully reflects the impact of process constraints on intraday adjustment capacity and serves as the training target for subsequent machine learning-based predictors.

[0051] Step 3: Build a water plant’s daily regulation capacity predictor based on XGBoost.

[0052] The extracted feature vectors and the corresponding theoretical regulation capacity are used as input and output to train a water plant's intraday regulation capacity predictor using the XGBoost regression algorithm. The model aims to minimize prediction error by optimizing hyperparameters such as tree depth and learning rate to fully learn the complex nonlinear relationship between equipment characteristics and regulation capacity. Through training, a model is developed that can rapidly predict equipment's potential for upward and downward regulation based on real-time operating conditions, enabling adaptive estimation of intraday regulation capacity based on operating condition awareness.

[0053] Step 4: Propose a time series aggregation model for the daily regulation capacity of the water plant.

[0054] Based on the single-moment regulation capacity output of the XGBoost predictor, the results are aggregated chronologically to form a complete intraday regulation capacity curve for the water plant. Separate upward and downward regulation capacity curves are generated, taking into account the capacity conversion of discrete resources such as sludge dewatering equipment. This time-series aggregation model accurately describes the changes in the water plant's regulation capacity at different time periods throughout the day, providing dynamic, fine-grained decision support for power dispatch optimization and demand response strategy formulation.

[0055] Specifically, for step 1, a process constraint model for each power equipment in the water plant is constructed, including:

[0056] Water pump flow balance constraint: The total water supply flow of the water pump is composed of the sum of the raw water extraction flow and the internal water flow of the process. Through the flow balance relationship, it is ensured that the water output of the pump can simultaneously meet the needs of raw water transportation and process treatment. It is the basic guarantee condition for the hydraulic flow continuity of the entire water plant.

[0057] The total water supply flow of the water pump is shown in the following formula:

[0058] Q pump (t) = Q raw (t)+Q process (t)

[0059] Where: Q pump (t) is the total water flow of the pump (m 3 / h), Qraw (t) is the raw water delivery flow (m 3 / h), Q process (t) is the internal process water flow rate (m 3 / h).

[0060] Dosing Equipment Linkage Constraint: This specifies that the dosing dosage of the dosing equipment should be proportional to the water flow rate of the pump, with the proportionality coefficient being the required dosing concentration. This constraint ensures that the dosing amount is adjusted synchronously with changes in pump flow rate, thereby stabilizing treated water quality and preventing under- or overdosing due to flow rate fluctuations.

[0061] The linkage constraints of the dosing equipment are as follows:

[0062] Q dose (t) = D ratio (t)×Q pump (t)

[0063] Where: Q dose (t) is the dosage of the dosing equipment (kg / h), D ratio (t) is the ratio of dosage to water volume (kg / m 3 ), Q pump (t) is the total water flow of the pump (m 3 / h).

[0064] Blower air-water linkage constraint: describes the fixed proportional relationship between the blower air volume and the process water flow rate. Through the air-water linkage constraint, the aeration intensity can be reasonably adjusted with the changes in process water volume, maintaining the optimal working conditions for biochemical treatment or dissolved oxygen control, and preventing insufficient aeration or energy waste.

[0065] The blower air-water linkage constraint is as follows:

[0066] Q air (t) = k air-water ×Q process (t)

[0067] Where: Q air (t) is the air volume of the blower (Nm3 / h), k air-water is the gas-water ratio coefficient, Q process (t) is the internal process water flow rate (m 3 / h).

[0068] Synchronous constraints on the start and stop of sludge dewatering equipment: The start and stop of the sludge dewatering equipment must be synchronized or delayed with the operating status of the sludge pump and blower to ensure that the dewatering equipment can be put into operation only when there is sludge supply and aeration concentration conditions, to avoid equipment idling, poor dewatering effect or sludge retention problems.

[0069] The start-stop synchronization constraint of the sludge dewatering equipment is as follows:

[0070] u dewater (t)≥u sludge,pump (t) and u dewater (t)≥u air (t)

[0071] Where: u dewater (t) is the switch status of the sludge dewatering equipment (0 / 1), u sludge,pump (t) is the sludge pump switch status (0 / 1), u air (t) is the blower switch status (0 / 1).

[0072] The auxiliary linkage constraints of lighting equipment require that lighting equipment be automatically linked and turned on according to the operating status of the main production equipment (water pumps, dosing equipment, blowers, dehydration equipment). Through lighting linkage, personnel are guaranteed to have sufficient lighting during equipment operation and maintenance, thereby improving operational safety and emergency response capabilities.

[0073] The auxiliary linkage constraint of lighting equipment is as follows:

[0074] u light (t)≥max(u pump (t),u dose (t),u air (t),u dewater (t))

[0075] Where: u light (t) is the lighting equipment switch status (0 / 1), u pump (t) is the water pump switch status (0 / 1), u dose (t) is the switch status of the dosing equipment (0 / 1), u dewater (t) is the switch status of the sludge dewatering equipment (0 / 1), u air (t) is the blower switch status (0 / 1).

[0076] Specifically, for step 2, a theoretical daily regulation capacity evaluation model of a water plant considering process constraints is proposed, and the method is as follows:

[0077] Firstly, considering the process constraints of the water plant, a feature vector that comprehensively reflects the operating status of each power equipment is constructed.

[0078] In this embodiment, the characteristics include the flow rate, power, and switch status of the water pump, dosing equipment, blower, sludge pump, and sludge dewatering equipment, as well as the auxiliary lighting status, ambient temperature, and humidity, etc., which are expressed as follows:

[0079]

[0080] Where: Q pump (t) is the pump flow rate (m 3 / h), P pump (t) is the pump power (kW), u pump (t) is the start / stop state of the water pump (0 / 1), Q dose (t) is the dosing flow rate of the dosing equipment (kg / h), P dose (t) is the power of the dosing equipment (kW), u dose (t) is the start / stop status of the dosing equipment (0 / 1), Q air (t) is the air volume of the blower (Nm 3 / h), P air (t) is the blower power (kW), u air (t) is the blower start / stop status (0 / 1), Q sludge (t) is the sludge pump flow rate (m 3 / h), P sludge (t) is the sludge pump power (kW), u sludge (t) is the start / stop state of the sludge pump (0 / 1), P dewater (t) is the power of sludge dewatering equipment (kW), u dewater (t) is the start / stop status of the sludge dewatering equipment (0 / 1), P light (t) is the lighting system power (kW), u light (t) is the lighting system start / stop status (0 / 1), T ambient (t) is the ambient temperature (℃), H ambient (t) is the ambient humidity (%).

[0081] This formula not only extracts the physical operating volume of the equipment, but also incorporates environmental influencing factors. It can provide accurate and rich input information for the subsequent intraday regulation capacity prediction model, significantly improving the overall modeling accuracy.

[0082] Then, a calculation method for the theoretical regulation potential of each major equipment in the water plant under the current operating status is proposed.

[0083] Construct a water pump daily adjustment capacity evaluation model as shown below:

[0084]

[0085] Where: The regulating capacity of the water pump to increase the flow upward, Q pump,max is the maximum allowable flow rate of the pump, Q pump (t) is the water pump flow rate.

[0086] This formula defines the maximum upward adjustment capacity of the pump at time t, that is, the remaining adjustment space between the current pump flow and its maximum allowable flow. If the current flow is close to the upper limit, the upward adjustment potential is zero; otherwise, the upward adjustment potential is the remaining flow difference, reflecting the pump's adjustment ability to increase the flow upward.

[0087] Construct a water pump intraday downward adjustment capacity evaluation model as shown below:

[0088]

[0089] Where: The adjustment capability of the water pump to increase the flow downward, Q pump,min is the minimum allowable flow rate of the pump, Q pump (t) is the water pump flow rate.

[0090] This formula describes the potential reduction of the pump's current flow rate relative to its minimum allowable flow rate, reflecting the maximum extent to which the pump can be lowered at time t. If the current flow rate is close to the lower limit, there is no potential for downward adjustment. If there is a surplus, the potential for downward adjustment is the difference between the two, which is used for flexible load reduction.

[0091] Construct a dosing equipment adjustment capacity evaluation model as shown below:

[0092]

[0093] Where: Adjust the capacity and Q of the dosing equipment dose (t) is the dosing flow rate, Q dose,max It is the maximum allowable dosing flow rate of the dosing equipment.

[0094] This formula quantifies the maximum upward adjustment capacity of the dosing equipment at time t, that is, the dosing flow rate can be increased to the remaining adjustment space of the equipment upper limit, ensuring that when the water pump flow increases or the water quality requirements are improved, the dosing equipment has sufficient capacity to synchronously increase the dosage to maintain stable water quality.

[0095] Construct a dosing equipment down-regulation capability evaluation model as shown below:

[0096]

[0097] Where: Lower the capacity and Q of the dosing equipment dose (t) is the dosing flow rate, Q dose,min It is the minimum allowable dosing flow rate of the dosing equipment.

[0098] This formula represents the range of dosage that can be reduced by the dosing equipment under the current state. If the dosage is much higher than the minimum requirement, the dosing equipment has a certain potential for downward adjustment, and the dosage of the agent can be reduced simultaneously when the water pump load is reduced or the water quality load is reduced.

[0099] Construct a blower adjustment capability evaluation model as shown below:

[0100]

[0101] Where: To increase the blower capacity, Q air (t) is the blower air volume, Q air,max The maximum allowable air volume of the blower.

[0102] This formula quantifies the blower's upward adjustment potential, that is, the room for the current air volume to increase relative to the maximum allowable air volume. It is used to characterize the blower's flexibility in increasing aeration intensity to meet the needs of increased water treatment load or increased dissolved oxygen levels.

[0103] Construct a blower down-regulation capability evaluation model as shown below:

[0104]

[0105] Where: To add a blower, reduce the capacity and Q air (t) is the blower air volume, Q air,min The minimum allowable air volume of the blower.

[0106] This formula describes the potential reduction of the blower at time t, reflecting the maximum amount the blower can reduce airflow when the load is reduced or energy conservation is required. If the airflow is close to the lower limit, the potential reduction is zero.

[0107] Construct a sludge pump adjustment capacity evaluation model as shown below:

[0108]

[0109] Where: Increase the capacity and Q of the sludge pump sludge,max is the maximum allowable flow rate of the sludge pump, Q sludge (t) is the sludge pump flow rate.

[0110] This formula calculates the potential of the sludge pump to increase the sludge pumping flow rate. It is applicable to situations where the sludge accumulation increases and the dewatering load rises, and the sludge pump needs to increase the operating volume to ensure the matching and continuity of the processing capacity.

[0111] Construct a sludge pump down-regulation capacity evaluation model as shown below:

[0112]

[0113] Where: Lower the capacity and Q of the sludge pump sludge,min is the minimum allowable flow rate of the sludge pump, Q sludge(t) is the sludge pump flow rate.

[0114] This formula indicates that the sludge pump has the potential to reduce the sludge pumping flow rate. When the sludge load is reduced or energy-saving scheduling is required, the sludge pump can moderately reduce the operating flow rate to improve the flexibility and energy efficiency of the overall system.

[0115] Construct a sludge dewatering equipment switch adjustment capacity evaluation model as shown in the following formula:

[0116] Δu dewater (t)=1-u dewater (t)

[0117] Where: Δu dewater (t) is the switching potential of the sludge dewatering equipment (0 / 1), aw u ed (t) is the start / stop status of the sludge dewatering equipment (0 / 1).

[0118] This formula defines the on / off switching potential of the sludge dewatering equipment. If the equipment is currently off (u=0), it has the potential to be turned on; if the equipment is currently on (u=1), it has the potential to be turned off.

[0119] In particular, in this embodiment, the dehydration equipment participates in the regulation in a discrete manner, which usually corresponds to a large step load change.

[0120] Construct a lighting equipment switching adjustment capability evaluation model as shown in the following formula:

[0121] Δu light (t)=1-u light (t)

[0122] Where: Δu light (t) is the lighting equipment switching potential (0 / 1), u light (t) is the lighting system start / stop status (0 / 1).

[0123] This formula quantifies the start-stop switching potential of lighting equipment. Although the lighting load accounts for a small proportion of the total system load, it still has regulatory significance in specific emergency energy saving or maintenance operation switching, and can flexibly respond to external control commands through on-off switching.

[0124] Specifically, for step 3, a water plant daily regulation capacity predictor based on XGBoost is constructed as follows:

[0125] Based on the XGBoost regression model, a water plant regulation capacity predictor is constructed. The constructed equipment operation feature vector X(t) is used as input, and the theoretical regulation capacity Y(t) calculated in step 2 is used as the supervision label. The optimal parameter Θ is obtained by training through the objective function of minimizing the prediction error and model complexity. The formula is as follows:

[0126]

[0127] Where: is the loss function (usually the mean square error (MSE) is used), Ω(Θ) is the regularization term (to limit the complexity of the model and avoid overfitting), X(t) is the operating characteristic vector of each device in the water plant at the current moment, Y(t) is the theoretical regulation capacity vector at the current moment, Y(t) is the regulation capacity result predicted by XGBoost at the current moment, and Θ is the hyperparameter set of the XGBoost model.

[0128] This formula indicates that the model controls complexity while ensuring accuracy, thereby achieving stable and efficient prediction of regulation capability. It can automatically capture the nonlinear mapping between the characteristics of each power equipment and its regulation potential, and adapt to multi-source disturbances and complex constraints in actual operation.

[0129] Specifically, for step 4, a time series aggregation model of the water plant's daily regulation capacity is proposed, which is expressed as follows:

[0130]

[0131] Where: The total capacity of the water plant increased at time t, The total capacity reduction of the water plant at time t, Increase the flow potential of the pump, The potential for increasing air volume of the blower, Δu dewater (t) is the switching potential of the sludge dewatering equipment, γ dewater The capacity conversion value corresponding to the sludge dewatering equipment action, Reduce the flow potential of the pump, Potential for lowering air volume on blowers.

[0132] In this embodiment, first, the regulation capacity results output by the XGBoost predictor at all times of the day are time-series aggregated to generate the daily regulation capacity curve of the water plant; then, the upward and downward regulation potentials of water pumps, blowers, dehydration equipment, etc. are counted separately, and switching resources such as sludge dehydration equipment are equivalently converted and included in the total capacity; finally, a variety of flexibility indicators including upward regulation capacity curve and downward regulation capacity curve are output to comprehensively reflect the support potential of the water plant to the power grid in different time periods, providing a time-series decision-making basis for scheduling optimization, demand response and operation evaluation.

[0133] 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 XGBoost intraday regulation capacity evaluation method of a water plant considering process constraints as described in any of the above embodiments.

[0134] 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 above-mentioned methods for evaluating the intraday regulation capacity of a water plant XGBoost considering process constraints are implemented.

[0135] 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 intraday regulation capacity of a water plant using XGBoost considering process constraints, characterized in that: The steps include: Step 1: Construct a process constraint model for each electrical equipment in the water plant: Based on the process flow of each equipment in the water plant, establish process constraint relationships between water pumps, dosing equipment, blowers, sludge dewatering equipment, and lighting equipment. Set the upper and lower operating limits and switching logic of each electrical equipment to form a dynamic constraint model that meets actual operation requirements. Step 2: Construct a theoretical daily regulation capacity assessment model for the water plant considering process constraints: Based on the process constraints, extract the flow, power, and state characteristics of each power device, construct a feature vector, and calculate the theoretical regulation potential based on the feature input combined with the operating range and start-stop constraints of each power device; Step 3: Construct an XGBoost-based predictor for the water plant's daily regulation capacity: The extracted feature vectors of each device and the corresponding theoretical regulation capacity matrix are used as input and output. The XGBoost regression algorithm is used to train the water plant's daily regulation capacity predictor. The hyperparameters are optimized with the goal of minimizing the prediction error. The nonlinear relationship between the characteristics of each power device and its regulation potential is learned. The upward and downward regulation potential of the equipment is predicted based on the real-time operating conditions, and the daily regulation capacity is adaptively estimated under the operating condition. Step 4. Construct a time series aggregation model for the water plant's intraday regulation capacity: Based on the single-moment regulation capacity results output by the water plant's intraday regulation capacity predictor, aggregate them in chronological order to form the water plant's intraday regulation capacity curve, and consider the conversion of discrete resource capacity to obtain an evaluation result describing the changes in the water plant's adjustable capacity at different time periods throughout the day.

2. The XGBoost daily regulation capacity evaluation method for a water plant considering process constraints according to claim 1 is characterized in that: In the process constraint model of each power equipment in the water plant described in step 1: Construct a water pump flow balance constraint and set the water output of the pump to meet the requirements of raw water transportation and process treatment at the same time. The formula is as follows: Q pump (t)=Q raw (t)+Q process (t) Where: Q pump (t) is the total water flow of the pump, Q raw (t) is the raw water delivery flow rate, Q process (t) is the internal process water flow rate (m 3 / h); Construct linkage constraints for dosing equipment, set the dosing dosage of the dosing equipment to be proportional to the water supply flow of the pump, and the proportional coefficient is the dosing concentration requirement. The formula is as follows: Q dose (t)=D ratio (t)×Q pump (t) Where: Q dose (t) is the dosage of the dosing equipment, D ratio (t) is the ratio of dosage to water volume, Q pump (t) is the total water flow rate supplied by the pump; Construct a blower air-water linkage constraint and set a fixed proportional relationship between the blower air volume and the process water flow rate. The formula is as follows: Q air (t)=k air-water ×Q process (t) Where: Q air (t) is the air output of the blower, k air-water is the gas-water ratio coefficient, Q process (t) is the internal process water flow rate; Construct the start-stop synchronization constraint of the sludge dewatering equipment and set the start-stop of the sludge dewatering equipment to be synchronized or delayed with the operating status of the sludge pump and blower. The formula is as follows: u dewater u(t) ≥ u sludge,pump u(t) and u dewater u(t) ≥ u air u(t) Where: u dewater (t) is the switch status of the sludge dewatering equipment, u sludge,pump (t) is the sludge pump switch state, u air (t) is the blower switch status; Construct auxiliary linkage constraints for lighting equipment, setting the lighting equipment to automatically start according to the operating status of production equipment, including water pumps, dosing equipment, blowers, and dehydration equipment. The formula is as follows: u light (t)≥max(u pump (t),u dose (t),u air (t),u dewater (t)) Where: u light (t) is the lighting equipment switch state, u pump (t) is the water pump switch state, u dose (t) is the switch status of the dosing equipment, u dewater (t) is the switch status of the sludge dewatering equipment, u air (t) is the blower switch status.

3. The XGBoost daily regulation capacity evaluation method for a water plant considering process constraints according to claim 1 is characterized in that: In step 2, based on the constraints of the water plant process, the flow, power, and state characteristics of each power equipment are extracted to construct a feature vector. The features include: the flow, power, and switch status of the water pump, dosing equipment, blower, sludge pump, and sludge dewatering equipment, as well as the auxiliary lighting status, ambient temperature, and humidity. The formula is as follows: Where: Q pump (t) is the pump flow rate, P pump (t) is the pump power, u pump (t) is the start and stop state of the water pump, Q dose (t) is the dosing flow rate of the dosing equipment, P dose (t) is the power of the dosing equipment, u dose (t) is the start and stop status of the dosing equipment, Q air (t) is the air volume of the blower, P air (t) is the blower power, u air (t) is the blower start and stop state, Q sludge (t) is the sludge pump flow rate, P sludge (t) is the sludge pump power, u sludge (t) is the start and stop state of the sludge pump, P dewater (t) is the power of sludge dewatering equipment, u dewater (t) is the start and stop state of the sludge dewatering equipment, P light (t) is the lighting system power, u light (t) is the lighting system start and stop state, T ambient (t) is the ambient temperature, H ambient (t) is the ambient humidity.

4. The XGBoost daily regulation capacity evaluation method for a water plant considering process constraints according to claim 3 is characterized in that: In step 2, based on the feature input and the operating range and start-stop constraints of each power equipment, the theoretical regulation potential under the current operating state is calculated, including: water pump daily upward regulation capacity assessment, water pump daily downward regulation capacity assessment, dosing equipment upward regulation capacity assessment, dosing equipment downward regulation capacity assessment, blower upward regulation capacity assessment, blower downward regulation capacity assessment, sludge pump upward regulation capacity assessment, sludge pump downward regulation capacity assessment, sludge dewatering equipment switch switching regulation capacity assessment, lighting equipment switch switching regulation capacity assessment; the theoretical regulation capacity of the water plant is comprehensively obtained to form a moment-level theoretical regulation capacity matrix as the predictor training target.

5. The XGBoost daily regulation capacity evaluation method for a water plant considering process constraints according to claim 4 is characterized in that: In step 2, the water pump's daily adjustment capacity is evaluated, and the maximum adjustment capacity of the water pump at time t is obtained by the remaining adjustment space between the current water pump flow rate and the maximum allowed flow rate: Where: Q is the regulating capacity of the water pump to increase the flow upward, pump,max is the maximum allowable flow rate of the pump, Q pump (t) is the water pump flow rate; The daily reduction capacity of the water pump is evaluated by calculating the reduction space of the current flow rate of the water pump relative to the minimum allowable flow rate to obtain the maximum reduction range of the water pump at time t: Where: Q is the regulating ability of the water pump to increase the flow downward, pump,min is the minimum allowable flow rate of the pump, Q pump (t) is the water pump flow rate; The evaluation of the dosing equipment's upward adjustment capacity is performed by increasing the dosing flow rate to the remaining adjustment space of the equipment's upper limit, and quantifying the maximum upward adjustment capacity of the dosing equipment at time t: Where: To adjust the capacity of the dosing equipment, Q dose (t) is the dosing flow rate, Q dose,max The maximum allowable dosing flow rate of the dosing equipment; The dosing equipment's downward adjustment capacity assessment calculates the range of dosing amounts that can be reduced under the current state of the dosing equipment: Where: To reduce the capacity of the dosing equipment, Q dose (t) is the dosing flow rate, Q dose,min The minimum allowable dosing flow rate for the dosing equipment; The blower adjustment capacity assessment quantifies the blower's adjustment potential by comparing the current air volume with the maximum allowable air volume: Where: To increase the blower's capacity, Q air (t) is the air volume of the blower, Q air,max is the maximum allowable air volume of the blower; The blower lowering capacity evaluation is to calculate the lowering space of the blower at time t: Where: To increase the blower capacity, Q air (t) is the air volume of the blower, Q air,min is the minimum allowable air volume of the blower; The sludge pump upward adjustment capacity assessment calculates the potential of the sludge pump to increase the sludge flow rate: Where: To increase the capacity of the sludge pump, Q sludge,max is the maximum allowable flow of the sludge pump, Q sludge (t) is the sludge pump flow rate; The sludge pump down-regulation capacity assessment calculates the potential of the sludge pump to reduce the sludge flow rate: Where: To reduce the capacity of the sludge pump, Q sludge,min is the minimum allowable flow rate of the sludge pump, Q sludge (t) is the sludge pump flow rate; The on / off switching adjustment capability of the sludge dewatering equipment is evaluated and the on / off switching potential of the sludge dewatering equipment is calculated: Δu dewater (t)=1-u dewater (t) Where: Δu dewater (t) is the switching potential of the sludge dewatering equipment, u dewater (t) is the start and stop status of the sludge dewatering equipment; The lighting equipment switching adjustment capability assessment quantifies the lighting equipment's start-stop switching potential: Δu light (t)=1-u light (t) Where: Δu light (t) is the switching potential of the lighting equipment, u light (t) is the start and stop status of the lighting system.

6. The XGBoost daily regulation capacity evaluation method for a water plant considering process constraints according to claim 5 is characterized in that: In step 3, a water plant regulation capacity predictor is constructed based on the XGBoost regression model. The power equipment operation feature vector is used as input, and the theoretical regulation capacity calculated in step 2 is used as the supervision label. The optimal parameter Θ is obtained by training through the objective function of minimizing the prediction error and model complexity. The formula is as follows: Where: is the loss function, Ω(Θ) is the regularization term, X(t) is the operating characteristic vector of each device in the water plant at the current moment, Y(t) is the theoretical regulation capacity vector at the current moment, Y(t) is the regulation capacity result predicted by XGBoost at the current moment, and Θ is the hyperparameter set of the XGBoost model.

7. The XGBoost daily regulation capacity evaluation method for a water plant considering process constraints according to claim 6 is characterized in that: In step 3, the hyperparameters of the XGBoost model include: tree depth and learning rate. Through training, the nonlinear relationship between the characteristics of each power equipment and the regulation potential is fully learned, and a model is obtained to predict the equipment's potential for regulation up and down based on real-time operating conditions, and to perform adaptive estimation of the intraday regulation capacity under the operating condition perception of the water plant.

8. The XGBoost daily regulation capacity evaluation method for a water plant considering process constraints according to claim 6 is characterized in that: In step 4, the water plant's intraday regulation capacity time series aggregation model performs time series aggregation on the regulation capacity results output by the XGBoost predictor at all times throughout the day to generate the water plant's intraday regulation capacity curve. The upward and downward regulation potentials of water pumps, blowers, dehydration equipment, etc. are counted separately, and switching resources such as sludge dehydration equipment are converted into the total capacity after equivalent conversion. The upward and downward regulation capacity curves are output to reflect the water plant's support potential for the power grid in different time periods. The formula is as follows: Where: The total capacity of the water plant increased at time t, The total capacity reduction of the water plant at time t, Increase the flow potential of the pump, The potential for increasing air volume of the blower, Δu dewater (t) is the switching potential of the sludge dewatering equipment, γ dewater The capacity conversion value corresponding to the sludge dewatering equipment action, Reduce the flow potential of the pump, Potential for lowering air volume on blowers.

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 XGBoost intraday regulation capacity evaluation method of a water plant considering process constraints 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, which, when executed by a processor, implements the steps of the XGBoost intraday regulation capacity evaluation method for a water plant considering process constraints as described in any one of claims 1 to 8.

Images