Energy-saving scheduling method and system for secondary water supply pump house coordinated with time-of-use electricity price

Abstract

The present application relates to the technical field of intelligent water supply, especially to a MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method and system, comprising the following steps: S1, real-time acquisition of pump house equipment operation data, including equipment health state, load demand and external environment data, and preprocessing of the collected data; S2, based on the real-time operation data of the equipment, the health of the equipment is evaluated by using a health evaluation model, an equipment health index is generated, and it is used as an input of the optimization scheduling strategy. In the present application, by real-time acquisition of equipment data and combination of health evaluation, load and price prediction and robust MPC controller optimization scheduling, optimal start and stop of the equipment and load distribution are realized, thereby improving the problem that the traditional scheduling method mostly adopts artificial experience or fixed mode, and the health of the equipment and load fluctuation are not considered, resulting in equipment overload and high energy consumption.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent water supply, especially to a MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method and system. BACKGROUND

[0002] With the advancement of modern urbanization, the construction and management of water supply systems have become increasingly complex. In urban water supply systems, secondary water supply pump houses are an important link to ensure the stability of water resource supply, especially in high-rise buildings and large industrial parks. The operational efficiency of secondary water supply pump houses directly affects energy consumption and system stability.

[0003] Currently, most traditional scheduling methods use artificial experience or fixed modes, which fail to consider equipment health and load fluctuations, leading to equipment overload and high energy consumption. SUMMARY

[0004] To address the above shortcomings, the present application provides a MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method and system, aiming to improve the problem of traditional scheduling methods mostly using artificial experience or fixed modes, which fail to consider equipment health and load fluctuations, leading to equipment overload and high energy consumption.

[0005] In a first aspect, the present application provides a MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method, comprising the following steps:

[0006] S1, real-time collection of pump house equipment operation data, including equipment health status, load demand and external environment data, and preprocessing of the collected data;

[0007] S2, based on the real-time operation data of the equipment, using a health assessment model to assess the health status of the equipment, generating an equipment health index, and using it as an input for optimization scheduling strategy;

[0008] S3, using a deep learning model to predict future load demand, and using a time series model to predict future electricity price fluctuations, obtaining load demand prediction value and electricity price prediction value;

[0009] S4, combining equipment health status, load demand prediction and electricity price prediction results, using a robust MPC controller for scheduling optimization, obtaining the optimal equipment start-stop strategy and load distribution scheme;

[0010] S5, adjusting the start-stop state of the pump house equipment according to the optimized scheduling strategy, and adjusting the load distribution according to the real-time feedback data to ensure that the equipment operation matches the load demand;

[0011] S6, real-time monitoring device running state, load change and price fluctuation, through the feedback mechanism to the scheduling strategy dynamic correction.

[0012] By adopting the above technical scheme, real-time device data is collected and combined with health assessment, load and price prediction and robust MPC controller optimization scheduling, and then optimal start and stop of the device and load distribution are realized, thereby improving the problem that the traditional scheduling method mostly adopts artificial experience or fixed mode, and the device overload and high energy consumption are caused due to the failure to consider the device health and load fluctuation.

[0013] Preferably, the data preprocessing comprises:

[0014] Real-time monitoring of key operating parameters of the pump house equipment, such as vibration, temperature and pressure, and collecting these data through sensors;

[0015] The collected raw data is denoised, and a Kalman filter denoising algorithm is used to remove noise in the data, so as to ensure the accuracy and reliability of the data for subsequent analysis and use.

[0016] Preferably, the health state of the equipment comprises:

[0017] Fusing the real-time collected equipment vibration, temperature and pressure operating data with the historical operating data of the equipment;

[0018] Using a Kalman filter data fusion algorithm to denoise and integrate the fused data;

[0019] Combining with the equipment degradation model, generating the equipment health index through the health assessment algorithm, the equipment health index reflecting the current health state of the equipment, and taking it as the input data for subsequent scheduling optimization.

[0020] Preferably, the load demand prediction comprises:

[0021] Collecting historical load data, weather data and external factors affecting load of the pump house, such as holidays and seasonal changes;

[0022] Using a long short-term memory network deep learning model to train the collected historical data, learning the time sequence relationship between the load demand and the external factors;

[0023] Using the trained LSTM model to predict the load demand of the future period, obtaining the load demand prediction value, and taking it as the basis for scheduling optimization.

[0024] Preferably, the price prediction comprises:

[0025] Collecting historical price data and external factor data related to price fluctuation;

[0026] An autoregressive integrated moving average model is used to model historical electricity price data and capture the regularity of electricity price fluctuations.

[0027] The trained autoregressive integrated moving average model is used to predict future electricity price fluctuations, resulting in electricity price prediction results that serve as the basis for subsequent scheduling decisions.

[0028] Preferably, the robust MPC controller includes:

[0029] Based on the equipment health index, load demand prediction and electricity price prediction results, an optimization objective function of the robust MPC controller is constructed;

[0030] Constraints for the MPC controller are set, including equipment health state constraints, load demand constraints and electricity price fluctuation constraints, to ensure the robustness of the scheduling strategy;

[0031] Through an optimization algorithm, the optimization objective function is solved to obtain the optimal equipment start-stop strategy and load distribution scheme, ensuring the stability and effectiveness of the scheduling strategy in an uncertain environment.

[0032] Preferably, the scheduling strategy includes:

[0033] According to the optimized scheduling strategy, the start-stop state of the pump house equipment is adjusted, with low-energy-consumption equipment being preferentially activated or high-energy-consumption tasks being delayed;

[0034] Based on the equipment health assessment results, the load distribution is adjusted in real time, with equipment in good health being given priority in load distribution to ensure long-term stable operation of the equipment and the system.

[0035] Preferably, the dynamic correction includes:

[0036] The running state of the equipment, load changes and electricity price fluctuations are monitored in real time;

[0037] When changes in equipment health state, large fluctuations in electricity price or fluctuations in load demand are detected, the system automatically adjusts the scheduling strategy;

[0038] Based on new real-time data, the equipment start-stop state and load distribution are recalculated to ensure that the scheduling strategy can adapt to real-time changes.

[0039] Preferably, the load distribution adjustment includes:

[0040] The equipment health index obtained from the equipment health assessment module is used to assess the health status of the equipment;

[0041] Based on the equipment health status, the load is preferentially distributed to equipment in good health to reduce the load on equipment in poor health, thereby avoiding overload or failure;

[0042] When adjusting the load distribution, ensure that the overall load demand is met, while prolonging the service life of the equipment.

[0043] In a second aspect, the present application provides the following technical solutions, MPC and time-of-use electricity price coordinated secondary water pump house energy-saving scheduling system, comprising the following modules:

[0044] The data acquisition module is used for real-time acquisition of the running data of the pump house equipment, including the device health state, load demand and external environment data, and pre-processing of the collected data;

[0045] The device health assessment module is used for assessing the health state of the device based on the real-time running data of the device, using a health assessment model to generate a device health index, and using the device health index as an input of the optimized scheduling strategy;

[0046] The load and electricity price prediction module is used for predicting future load demand using a deep learning model, and predicting future electricity price fluctuations using a time series model, to obtain a load demand prediction value and an electricity price prediction value;

[0047] The robust MPC controller module is used for combining the device health state, load demand prediction and electricity price prediction results, and performing scheduling optimization through a robust MPC controller to obtain an optimal device start-stop strategy and load distribution scheme;

[0048] The scheduling execution module is used for adjusting the start-stop state of the pump house equipment according to the optimized scheduling strategy, and adjusting the load distribution according to the real-time feedback data to ensure that the device operation matches the load demand;

[0049] The feedback adjustment module is used for real-time monitoring of the device running state, load change and electricity price fluctuation, and dynamically correcting the scheduling strategy through a feedback mechanism.

[0050] The present application has the following beneficial effects:

[0051] 1、In the present application, by real-time acquisition of device data and combination of health assessment, load and electricity price prediction and robust MPC controller optimized scheduling, optimal start-stop and load distribution of the device are realized, thereby improving the problem that the traditional scheduling method mostly adopts artificial experience or fixed mode, and the device overload and high energy consumption problem caused by failure to consider device health and load fluctuation.

[0052] 2、In the present application, by real-time acquisition of the running data of the pump house equipment, the accuracy and reliability of the data are ensured, thereby improving the problem that the traditional device monitoring method mostly relies on artificial patrol or periodic inspection, and the problem of device failure that cannot be discovered and processed in time caused by failure to realize real-time monitoring.

[0053] 3、The device health index is generated through the device health evaluation model in the application, and then a basis is provided for the scheduling strategy, thereby improving the traditional device management method which mostly relies on fixed-period maintenance and manual judgment, and the risk of device overload and failure is caused due to the failure to dynamically evaluate the device health state.

[0054] 4、The device start-stop and load distribution are optimized through the robust MPC controller in the application, and then the efficient operation of the device is ensured, thereby improving the traditional scheduling strategy which mostly adopts a fixed mode, and the problems of energy waste and low system operation efficiency are caused due to the failure to adapt to the changes of device health, load demand and electricity price fluctuation. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 A method flowchart of the MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method is provided in the application.

[0056] Figure 2 A data acquisition and preprocessing flowchart of the MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method is provided in the application.

[0057] Figure 3 A device health evaluation flowchart of the MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method is provided in the application.

[0058] Figure 4 A load and electricity price prediction principle diagram of the MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method is provided in the application.

[0059] Figure 5 A robust MPC controller optimization scheduling principle diagram of the MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method is provided in the application.

[0060] Figure 6 A module architecture diagram of the MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling system is provided in the application. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings, obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0062] Embodiment one:

[0063] In the first embodiment of the application, the application provides a MPC and time-of-use electricity price coordinated secondary water supply pump house energy-saving scheduling method, as shown inFigures 1-5 As shown, comprising the following steps:

[0064] S1, real-time acquisition of pump house equipment operation data, including equipment health status, load demand and external environment data, and preprocessing the collected data;

[0065] Further, the data preprocessing includes:

[0066] Real-time monitoring of key operating parameters of pump house equipment, such as vibration, temperature, pressure, and collecting these data through sensors;

[0067] The collected raw data is denoised, and Kalman filter denoising algorithm is used to remove noise in the data, ensuring the accuracy and reliability of the data for subsequent analysis.

[0068] Specifically, data collection, through sensors to monitor the key operating parameters of pump house equipment, such as vibration, temperature, pressure, etc. These data provide basic information for subsequent analysis of the system. These sensors can be various types of temperature sensors, pressure sensors and vibration sensors, which record the working state and operating environment of the equipment in real time. The collected data is transmitted to the system processing unit through the data acquisition module.

[0069] Data preprocessing, the raw data collected has certain noise, which may be caused by the accuracy limit of the sensor, external environmental interference and other factors. In order to improve the data quality, the raw data collected must be denoised. Specifically, Kalman filter algorithm is used to filter the raw data to remove noise and ensure data accuracy. Kalman filter is a recursive estimation algorithm suitable for extracting system state from noisy data.

[0070] The basic process of Kalman filter includes the following formulas, state update formula:

[0071] ; wherein is the prediction of state variable, is the state transition matrix, is the estimated value at the last time, is the control input matrix, is the control input.

[0072] Kalman gain: ; wherein is the Kalman gain, is the prediction error covariance, is the variance of measurement noise.

[0073] State estimation update: ; wherein For filtered state estimation, For measurement, For prediction.

[0074] Error covariance update: ; Where Updated error covariance.

[0075] Through the above steps, the Kalman filter algorithm can effectively remove the noise in the sensor data, ensuring the accuracy of the data. The filtered data can truly reflect the running state of the equipment, providing accurate basic data for subsequent health assessment, load prediction and electricity price prediction.

[0076] Input and output process, input: original equipment running data collected by sensors, including vibration, temperature, pressure and other parameters. Output: accurate data processed by Kalman filter, used for subsequent analysis and optimization scheduling decision.

[0077] By collecting real-time running data of pump house equipment and performing Kalman filter denoising processing, the noise in the data can be effectively eliminated, and the influence of external interference on the data can be reduced. This processing method provides high-quality input data for the system, thereby supporting accurate decision-making in equipment health assessment, load demand prediction and other aspects. The high accuracy of data preprocessing ensures the reliability of subsequent optimization scheduling scheme, so as to achieve the purpose of energy saving and efficient operation.

[0078] S2, based on the real-time running data of the equipment, using the health assessment model to evaluate the health status of the equipment, generating the equipment health index, and taking it as the input of the optimization scheduling strategy;

[0079] Further, the health status of the equipment includes:

[0080] Fusing the real-time collected equipment vibration, temperature, pressure running data with the historical running data of the equipment;

[0081] Using Kalman filter data fusion algorithm to denoise and integrate the fused data;

[0082] Combined with the equipment degradation model, the equipment health index is generated through the health assessment algorithm, which reflects the current health status of the equipment, and is taken as the input data of subsequent scheduling optimization.

[0083] Specifically, data fusion, real-time collected equipment running data (such as vibration, temperature, pressure, etc.) are fused with the historical running data of the equipment. Historical data may include equipment failure records, maintenance records, equipment service life, repair history and other information. These data help the system understand the running background and potential risks of the equipment more comprehensively. By fusing real-time data with historical data, the system can more accurately assess the health status of the equipment.

[0084] Kalman filter denoising, for the fused data, the Kalman filter algorithm will be denoised and integrated. Because the sensor may be disturbed by various external factors when collecting data, the Kalman filter algorithm is used to remove noise and accurately extract the key running information of the device. Through the Kalman filter algorithm in S1, the noise in the sensor data will be effectively removed, and the remaining is the reliable data that truly reflects the current state of the device. This process ensures that the data used by the health assessment model has high accuracy and credibility.

[0085] Health assessment and health index generation, based on the data after Kalman filter denoising, combined with the device degradation model, the health index of the device is generated through the health assessment algorithm. The device degradation model predicts the degradation trend and future state of the device according to the historical running data, load, working conditions and other factors of the device. The device health index is calculated according to the degradation model and the data fusion result, and the specific generation formula is: ; wherein is the device health index, is the filtered and denoised device running data, is the degradation parameter of the device, is the external environmental influence factor of the device. The device health index is usually between [0, 1], the closer the value is to 1, the better the health condition of the device; the closer the value is to 0, the worse the health condition of the device, which may need to be maintained or replaced. Through this formula, the health status of the device is quantified as an index, reflecting the running stability, degradation level and failure risk of the device. The index will be used as input for subsequent scheduling optimization, guiding the adjustment and optimization of scheduling strategy.

[0086] Input and output process, input: real-time collected device vibration, temperature, pressure and other running data; historical running data of the device; degradation model parameters. Output: generated device health index, as input for subsequent scheduling optimization.

[0087] By fusing real-time data with historical data and processing through Kalman filter denoising, the health assessment process can assess the health status of the device based on more accurate input. The health index generated by combining the device degradation model provides a key input for scheduling optimization. The health index can effectively reflect the current condition of the device, helping the system to real-time grasp the health status of the device, thereby avoiding device failure, optimizing energy scheduling, and prolonging the service life of the device.

[0088] S3, use a deep learning model to predict future load demand and use a time series model to predict future electricity price fluctuations to obtain load demand prediction value and electricity price prediction value;

[0089] Further, the load demand prediction includes:

[0090] Collecting historical load data of the pump house, meteorological data, and external factors such as holidays and seasonal changes affecting load;

[0091] Using a long short-term memory network deep learning model to train the collected historical data and learn the time series relationship between load demand and external factors;

[0092] Using the trained LSTM model to predict the load demand of the future period, obtaining the load demand prediction value, and serving as the basis for dispatch optimization.

[0093] Specifically, the first step of historical data collection and load demand prediction is to collect historical data related to the operation of the pump house. The data includes: historical load data, which records the actual load demand changes of the pump house. Meteorological data, such as temperature, humidity, and air pressure, which may affect the load demand of the pump house. Holidays and seasonal changes, these external factors also affect the load demand, so we need to consider holidays, seasonal changes and other factors. By collecting these historical data, we can fully understand the fluctuation rules of load demand and external influencing factors.

[0094] LSTM model training, using the collected historical load data, meteorological data, and external factors such as holidays and seasonal changes, through long short-term memory network (LSTM) for training. LSTM is a special recurrent neural network suitable for processing and predicting time series data, which can effectively capture the time series relationship in long time period. The basic structure of LSTM model includes input layer, multiple LSTM units, fully connected layer and output layer. LSTM units can maintain long-term memory of important information in time dimension through its unique memory unit and gating mechanism, so as to more accurately predict load demand.

[0095] The basic formula of LSTM network, forget gate: ; Where is the output of the forget gate, is the sigmoid activation function, is the weight matrix, is the hidden layer state at the last time, is the input at the current time, is the bias term.

[0096] Input gate: ; Where is the output of the input gate, is the weight matrix of the input gate, is the bias term.

[0097] Candidate memory cell: ; wherein is a candidate memory cell, is a hyperbolic tangent activation function.

[0098] updating the memory cell: ; wherein is the memory cell at the current time step, is the forget gate output, is the input gate output, is the candidate memory cell.

[0099] the output gate: ; wherein is the output gate output.

[0100] the final output: ; wherein is the hidden layer output at the current time step, is the memory cell, is the output gate.

[0101] load demand prediction, after the training is completed, the trained LSTM model is used to predict the future load demand. The prediction will be based on historical load data and external factors such as weather data, and output the load demand value in the future period. Load demand prediction formula: ; wherein is the load demand prediction value of the future period, is the hidden layer state calculated by the LSTM model, is the current input external factor data.

[0102] input and output process, input: historical load data: actual load demand of pump house; weather data: including temperature, humidity, air pressure, etc.; holidays and seasonal changes: external factor data. Output: predicted load demand value, used for subsequent dispatching optimization.

[0103] Through training of historical load data and external factor data, the LSTM model can capture the time sequence rule of load demand and accurately predict the future load demand. The predicted load demand value provides a basis for dispatching optimization, so that reasonable allocation of equipment load can be realized in the dispatching process, energy consumption is optimized, and efficient operation of the water supply system is guaranteed.

[0104] S4, combined with the device health state, load demand prediction and electricity price prediction results, the robust MPC controller is used for dispatching optimization to obtain the optimal device start-stop strategy and load allocation scheme;

[0105] Further, the electricity price prediction includes:

[0106] Collect historical electricity price data and external factor data related to electricity price fluctuations;

[0107] Model historical electricity price data using an autoregressive integrated moving average (ARIMA) model to capture the patterns of electricity price fluctuations.

[0108] Use the trained ARIMA model to predict future electricity price fluctuations, obtaining electricity price prediction results as the basis for subsequent dispatch decisions.

[0109] Specifically, electricity price prediction is a key factor in dispatch optimization. To ensure that the system can respond flexibly in the case of large electricity price fluctuations, historical electricity price data and external factor data related to electricity price fluctuations need to be collected first. These data provide the necessary basis for establishing an electricity price fluctuation model. The prediction of electricity price uses an autoregressive integrated moving average (ARIMA) model. The ARIMA model is a statistical method widely used in time series prediction, which can capture the historical fluctuation patterns of electricity prices and make reasonable predictions of future electricity prices.

[0110] The modeling process of the ARIMA model includes: where is the electricity price data at a certain time in the time series, is the model parameter, is white noise.

[0111] Difference, used to eliminate non-stationary sequences, by differencing the sequence, the electricity price sequence becomes a stationary sequence.

[0112] Moving average, where is the parameter of the moving average model, is white noise.

[0113] The final form of the ARIMA model is: .

[0114] Through this model, the patterns of electricity price fluctuations can be learned from historical electricity price data and external factor data, and predictions can be made. The trained ARIMA model can be used for future electricity price prediction.

[0115] Electricity price prediction results and dispatch optimization: Use the trained ARIMA model to predict future electricity prices, obtaining electricity price prediction results. Electricity price prediction results will play an important role in dispatch optimization as one of the basis for future load distribution.

[0116] Robust MPC Scheduling Optimization: Combining electricity price prediction results, device health status, and load demand prediction, the robust model predictive control (MPC) is used for scheduling optimization. The robust MPC controller can optimize the start-stop strategy and load distribution scheme of the device in the presence of uncertainty, to reduce energy consumption and maintain system stability. The goal of the robust MPC controller is to minimize the overall cost of the system, considering the impact of device health status, load demand, and electricity price fluctuations on the system. The basic model of the controller is as follows:

[0117] Robust MPC Controller Objective Function: ; where is the load distribution cost at time ; is the energy consumption cost at time , based on electricity price prediction and load demand; is the device failure cost at time , based on device health status and device start-stop strategy.

[0118] Constraints: ; where is the power output of device at time , is the total load demand at time .

[0119] ; where and are the minimum and maximum power output of device , respectively.

[0120] By optimizing the above objective function, the robust MPC controller can calculate the optimal device start-stop strategy and load distribution scheme, ensuring that the pump house meets the load demand while considering electricity price fluctuations and device health status, to achieve the minimization of energy cost and the optimal state of device operation.

[0121] Input and output flow, input: electricity price prediction results: electricity price fluctuations predicted by ARIMA model; device health status: device health index obtained by health assessment model; load demand prediction: future load demand predicted by LSTM model. Output: optimal device start-stop strategy: device start-stop decision optimized by robust MPC controller; optimal load distribution scheme: load distribution scheme calculated by robust MPC controller.

[0122] By combining electricity price prediction, load demand prediction and equipment health status, and using a robust MPC controller for scheduling optimization, the present application can obtain the optimal equipment start-stop strategy and load distribution scheme under multiple constraints. This scheme not only maximizes energy saving effect, but also ensures stable operation of equipment, avoids fault risk, and reduces equipment maintenance and replacement cost.

[0123] S5, adjusting the start-stop state of the pump house equipment according to the optimized scheduling strategy, and adjusting the load distribution according to the real-time feedback data to ensure that the equipment operation matches the load demand;

[0124] Further, the robust MPC controller comprises:

[0125] According to the equipment health index, load demand prediction and electricity price prediction results, an optimization objective function of the robust MPC controller is constructed;

[0126] The MPC controller is set with constraint conditions, including equipment health state constraint, load demand constraint, and electricity price fluctuation constraint, to ensure the robustness of the scheduling strategy;

[0127] Through an optimization algorithm, the optimization objective function is solved to obtain the optimal equipment start-stop strategy and load distribution scheme, ensuring the stability and effectiveness of the scheduling strategy in an uncertain environment.

[0128] Specifically, the optimization objective function of the robust MPC controller is constructed by the equipment health index, load demand prediction and electricity price prediction results. The core of the objective function is to minimize the overall operation cost of the system, including equipment operation cost, energy consumption cost, fault risk cost, etc.

[0129] The expression of the optimization objective function is: ; wherein is the load distribution cost at time , which reflects the matching degree of load demand and equipment operation state at that time; is the energy consumption cost at time , which is distributed according to the electricity price prediction and load demand; is the equipment failure cost at time , which is based on the equipment health index; is the equipment maintenance cost, which is based on the health assessment and fault prediction of the equipment. The objective function ensures the optimization of scheduling strategy in terms of energy consumption and cost under the multiple constraints of equipment health status, load demand and electricity price fluctuation.

[0130] The constraint conditions of the MPC controller, in order to ensure the stability and effectiveness of the robust MPC controller in the actual scheduling process, multiple constraint conditions need to be set for the controller. These constraints include:

[0131] Device health state constraints consider the health state of devices, ensuring that devices operate in a suitable working condition, avoiding overloading operation leading to device damage or failure. ; wherein is the health state threshold of device At time , the health index, is the health state threshold of device .

[0132] Load demand constraints ensure that the total load output of devices can meet the system's load demand. ; wherein is the load output of all devices at time , the load demand at this time.

[0133] Electricity price fluctuation constraints ensure that energy costs are minimized in the case of electricity price fluctuations. Considering the uncertainty of electricity price fluctuations, the MPC controller needs to make predictions and adjustments within each period to ensure the stability of the scheduling strategy. ; wherein is the energy consumption at time , the electricity price fluctuation limit at time . Through these constraints, the robust MPC controller can ensure that the scheduling strategy remains stable and effective in the case of electricity price fluctuations and device health state uncertainty.

[0134] Solution process of robust MPC controller: Once the optimization objective function and constraints are determined, the next step is to solve the objective function through optimization algorithms. The robust MPC controller uses appropriate optimization methods such as linear programming (LP), quadratic programming (QP), etc., for online solving, obtaining the optimal device start-stop strategy and load distribution scheme.

[0135] The optimization problem of robust MPC can be represented as: ; wherein is the scheduling decision vector (i.e., device start-stop strategy and load distribution), is the th constraint condition.

[0136] Real-time feedback data adjustment: During the execution of the scheduling strategy, the system will monitor the running state of the device in real time, and adjust the load distribution and device start-stop state according to the real-time feedback data. When the device health state, load demand or electricity price changes, the system will dynamically update the scheduling strategy to ensure that the device operation and load demand always match.

[0137] ​​Input-output flow, input: device health index: generated by device health assessment model; load demand prediction: future load demand data generated by LSTM model; electricity price prediction: future electricity price prediction data generated by ARIMA model. Output: optimized device start-stop strategy: device start-stop decision obtained according to optimization objective function and constraint condition; optimized load distribution scheme: optimized load distribution calculated according to predicted load demand and electricity price fluctuation.

[0138] Through the solving process of the robust MPC controller, combined with the device health state, load demand prediction and electricity price prediction, the system can effectively schedule in an uncertain environment, reduce energy consumption, reduce failure risk, and ensure that the device always operates in the best state. This scheme dynamically adjusts the device start-stop state and load distribution to ensure the matching of device operation and load demand, achieving efficient operation and energy saving of the system.

[0139] S6, real-time monitoring of device operating state, load change and electricity price fluctuation, dynamic correction of scheduling strategy through feedback mechanism.

[0140] Further, the scheduling strategy includes:

[0141] According to the optimized scheduling strategy, adjust the start-stop state of the pump house equipment, preferentially start low-energy-consumption equipment or delay high-energy-consumption tasks;

[0142] According to the device health assessment result, adjust the load distribution in real time, and preferentially consider the devices with good health status in load distribution to ensure the long-term stable operation of the device health and system.

[0143] Dynamic correction includes:

[0144] Real-time monitoring of device operating state, load change and electricity price fluctuation;

[0145] When the device health state changes, the electricity price fluctuates greatly or the load demand fluctuates, the system will automatically adjust the scheduling strategy;

[0146] Based on the new real-time data, recalculate the device start-stop state and load distribution to ensure that the scheduling strategy can adapt to real-time changes.

[0147] Load distribution adjustment includes:

[0148] According to the device health index obtained by the device health assessment module, evaluate the health status of the device;

[0149] According to the device health state, preferentially distribute the load to the devices with good health status, and reduce the load of the devices with poor health status to avoid overload or failure;

[0150] When adjusting the load distribution, ensure that the overall load demand is met while extending the service life of the equipment.

[0151] Specifically, the scheduling strategy adjustment: according to the optimized scheduling strategy, the system dynamically adjusts the start-stop state of the pump house equipment, preferentially starts low-energy-consumption equipment, and postpones high-energy-consumption tasks, thereby minimizing energy consumption and optimizing equipment use efficiency. Through the intelligent scheduling system, the most suitable equipment can be flexibly selected for operation according to the current state of the equipment and the fluctuation of electricity price.

[0152] The mathematical expression for adjustment is: ; wherein is the power output of the equipment at time , is the health status of the equipment , is the load demand of the equipment , is the electricity price at time . In this expression, the optimal equipment start-stop scheme is determined by calculating the comprehensive influence of the equipment health status, load demand, and electricity price fluctuation.

[0153] Equipment health assessment and load distribution: the system adjusts the load distribution in real time according to the equipment health assessment results. Specifically, the load is preferentially distributed to equipment with better health status, which can effectively avoid overloading of equipment with poor health status, thereby reducing the risk of failure and extending the service life of the equipment.

[0154] Adjustment strategy for load distribution: ; wherein is the load distribution of the equipment , is the sum of the health status of all equipment at time , is the total load demand of the system at the current time. This load distribution method ensures that equipment with better health status bears more load, while equipment with poor health status bears less load, thereby avoiding failure.

[0155] Dynamic adjustment mechanism: to ensure that the scheduling strategy can respond to real-time changes, the system dynamically adjusts the scheduling strategy by monitoring the running state of the equipment, load demand, and electricity price fluctuation in real time. When the equipment health status changes, the electricity price fluctuates greatly, or the load demand fluctuates, the system will automatically adjust the scheduling strategy to ensure that the start-stop state and load distribution of the equipment always match the actual situation.

[0156] The mathematical expression for the adjustment process is: ; wherein is the adjusted load of the equipment , allocating for current load, scheduling data from real-time feedback, including device status, load demand, and electricity price changes. Through the processing of real-time feedback data, the system can recalculate the start-stop state of devices and load allocation according to new data.

[0157] Device start-stop and load allocation correction: According to real-time feedback data, the system recalculates the start-stop state and load allocation of each device to ensure that the scheduling strategy can adapt to real-time changes and avoid device load mismatch or overload operation.

[0158] Through the following process, the system adjusts the schedule according to real-time data: if the electricity price fluctuates greatly, the system will automatically adjust the start-stop state of high-energy consumption devices and delay the operation time of these devices; if the device health status changes, the system will reduce the load allocation to devices with poor health status to avoid device overload; if the load demand changes, the system will adjust the load allocation of all devices accordingly to ensure that the devices can run continuously and stably.

[0159] Input and output process, input: device operating state: real-time monitoring of device health status, vibration, temperature and other parameters through sensors; load change: real-time collection of load demand data; electricity price fluctuation: real-time electricity price data, obtained through power market or electricity price prediction model. Output: adjusted device start-stop state: whether the device is enabled, the adjustment result of the running load; dynamic load allocation: the specific load allocation adjustment value of each device.

[0160] By real-time monitoring of device operating state, load change and electricity price fluctuation, and dynamically adjusting the scheduling strategy based on feedback mechanism, it can ensure that the start-stop state of the device and the load demand are always matched, so as to maximize the energy utilization efficiency and prolong the service life of the device. This technical solution not only enhances the adaptability of the system to uncertain environment, but also effectively improves the stability and reliability of the pump house equipment.

[0161] Example two:

[0162] In modern urban secondary water supply pump rooms, the pump room equipment is responsible for providing stable water supply for high-rise buildings and industrial parks. Due to the large energy consumption of the pump room, especially under the background of fluctuating load demand, uneven equipment health status and fluctuating electricity price, how to optimize the start-stop strategy and load distribution of the pump room equipment to achieve energy saving goal has become a technical problem to be solved. In order to solve this challenge, a scheduling system based on MPC and time-of-use electricity price coordination is introduced, which combines real-time monitoring, equipment health assessment, load demand prediction and electricity price fluctuation for dynamic scheduling and optimization. However, the existing problems are as follows: firstly, how to adjust the equipment start-stop and load distribution according to the changing load demand and electricity price fluctuation in real time to avoid energy waste and high peak period consumption. Secondly, how to adjust the load distribution according to the health status of the equipment to ensure that the equipment with poor health status is not overloaded, avoid equipment failure and prolong the service life. Thirdly, how to design a robust scheduling strategy to cope with uncertain factors such as equipment health change, load fluctuation and electricity price change, to ensure stable and efficient operation of the system. Fourthly, how to quickly respond to equipment operating state, electricity price fluctuation and load change, and adjust the scheduling strategy in real time to ensure that the start-stop state and load distribution of the equipment can adapt to changes in time, avoiding system mismatch and resource waste. In order to solve the above problems, the MPC and time-of-use electricity price coordinated secondary water supply pump room energy saving scheduling system provided by the present application is adopted, and the structure is as shown in Figure 6 The specific implementation process of the system is as follows:

[0163] The data acquisition module is used for real-time acquisition of the operating data of the pump room equipment, including the health status of the equipment, the load demand and the external environment data, and pre-processing of the collected data;

[0164] The equipment health assessment module is used for evaluating the health status of the equipment based on the real-time operating data of the equipment, using a health assessment model to generate an equipment health index, and using the equipment health index as an input for optimizing the scheduling strategy;

[0165] The load and electricity price prediction module is used for predicting future load demand using a deep learning model, and predicting future electricity price fluctuation using a time series model to obtain load demand prediction value and electricity price prediction value;

[0166] The robust MPC controller module is used for combining the equipment health status, load demand prediction and electricity price prediction results, and performing scheduling optimization through a robust MPC controller to obtain the optimal equipment start-stop strategy and load distribution scheme;

[0167] The scheduling execution module is used for adjusting the start-stop state of the pump room equipment according to the optimized scheduling strategy, and adjusting the load distribution according to the real-time feedback data to ensure that the equipment operation matches the load demand;

[0168] A feedback adjustment module is configured to monitor the equipment running state, load changes and electricity price fluctuations in real time, and dynamically correct the scheduling strategy through a feedback mechanism.

[0169] Specifically, the data acquisition module acquires the running data of the pump house equipment in real time, including the equipment health state, load demand and external environment data. After obtaining the data through sensors, the data is preprocessed such as denoising and standardization to ensure data quality and provide reliable basis for subsequent analysis and decision-making.

[0170] The equipment health evaluation module generates an equipment health index based on a health evaluation model using real-time running data. The equipment health index reflects the current running state of the equipment, and is used to optimize the scheduling strategy to avoid equipment overload or failure and ensure stable operation of the equipment.

[0171] The load and electricity price prediction module uses a deep learning model to predict future load demand, combined with external factors such as weather and holidays. At the same time, a time series model is used to predict electricity price fluctuations to provide load and electricity price prediction data for scheduling optimization and optimize equipment start-stop strategies.

[0172] The robust MPC controller module combines the equipment health state, load demand and electricity price prediction results, and the robust MPC controller performs scheduling optimization. This module optimizes the equipment start-stop strategy and load distribution according to uncertainty to ensure the stability and efficiency of scheduling in uncertain environments.

[0173] The scheduling execution module adjusts the equipment start-stop state and dynamically adjusts the load distribution according to the optimized scheduling strategy. This module ensures that the equipment always operates in the optimal way, avoids equipment overload and inefficient operation, and maximizes energy use efficiency and equipment health.

[0174] The feedback adjustment module monitors the equipment state, load changes and electricity price fluctuations in real time, and dynamically corrects the scheduling strategy. According to the feedback data, the equipment start-stop and load distribution are adjusted to ensure that the scheduling strategy can adapt to real-time changes and improve the flexibility and stability of the system.

[0175] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent replacements to some technical features, as long as they are within the spirit and principles of the present application. Any modification, equivalent replacement, improvement, etc. made within the scope of the present application should be included in the protection scope of the present application.

Claims

1. An energy-saving scheduling method for secondary water supply pump stations that coordinates MPC (Multi-Purpose Control) and time-of-use pricing, characterized in that: Includes the following steps: S1. Real-time collection of pump room equipment operation data, including equipment health status, load demand and external environment data, and preprocessing of the collected data; S2. Based on the real-time operating data of the equipment, the health status of the equipment is assessed using a health assessment model, an equipment health index is generated, and the index is used as the input for the optimization scheduling strategy. The assessment of the equipment's health status includes: fusing real-time collected equipment vibration, temperature, and pressure operating data with the equipment's historical operating data; using a Kalman filter data fusion algorithm to denoise and integrate the fused data; and combining the equipment degradation model with a health assessment algorithm to generate an equipment health index, which reflects the equipment's current health status and serves as input data for subsequent scheduling optimization. S3. Use a deep learning model to predict future load demand and a time series model to predict future electricity price fluctuations, and obtain the predicted load demand and electricity price values. The load demand forecasting includes: collecting historical load data, meteorological data, and external factors affecting the load such as holidays and seasonal changes; training the collected historical data using a long short-term memory network deep learning model to learn the temporal relationship between load demand and external factors; and using the trained LSTM model to predict the load demand for future periods to obtain the predicted load demand value, which serves as the basis for scheduling optimization. The electricity price forecasting includes: collecting historical electricity price data and external factor data related to electricity price fluctuations; using an autoregressive integral moving average model to model the historical electricity price data and capture the patterns of electricity price fluctuations; using the trained autoregressive integral moving average model to predict electricity price fluctuations in future periods, and obtaining the electricity price forecast results as the basis for subsequent dispatching decisions. S4. Combining the equipment health index, load demand forecast, and electricity price forecast results, the scheduling strategy is optimized through the robust MPC controller to obtain the optimal equipment start-up and shutdown strategy and load allocation scheme. The robust MPC controller constructs an optimization objective function based on equipment health index, load demand forecast, and electricity price forecast results; it sets constraints for the MPC controller, including equipment health status constraints, load demand constraints, and electricity price fluctuation constraints, to ensure the robustness of the scheduling strategy; and it solves the optimization objective function through an optimization algorithm to obtain the optimal equipment start-up and shutdown strategy and load allocation scheme, ensuring the stability and effectiveness of the scheduling strategy under uncertain environments. S5. Adjust the start-up and shutdown status of the pump room equipment and perform load distribution according to the optimal equipment start-up and shutdown strategy and load distribution scheme in the optimized scheduling strategy; S6. Monitor equipment operating status, load changes and electricity price fluctuations in real time, dynamically correct scheduling strategies through feedback mechanisms, and adjust load allocation to ensure that equipment operation matches load demand.

2. The energy-saving scheduling method for secondary water supply pump stations based on MPC and time-of-use pricing as described in claim 1, characterized in that, The data preprocessing includes: real-time monitoring of key operating parameters of the pump room equipment, such as vibration, temperature, and pressure, and collecting this data through sensors; denoising the collected raw data using a Kalman filter denoising algorithm to remove noise from the data, ensuring the accuracy and reliability of the data for subsequent analysis.

3. The energy-saving scheduling method for secondary water supply pump stations based on MPC and time-of-use pricing as described in claim 1, characterized in that, The scheduling strategy includes: adjusting the start-up and shutdown status of pump room equipment according to the optimized scheduling strategy, prioritizing the use of low-energy-consuming equipment or postponing high-energy-consuming tasks; and adjusting load allocation in real time based on equipment health assessment results, prioritizing equipment with better health status to ensure equipment health and long-term stable operation of the system.

4. The energy-saving scheduling method for secondary water supply pump stations based on MPC and time-of-use pricing as described in claim 1, characterized in that, The dynamic correction includes: real-time monitoring of equipment operating status, load changes, and electricity price fluctuations; when changes in equipment health status, large fluctuations in electricity prices, or fluctuations in load demand are detected, the system will automatically adjust the scheduling strategy; and recalculate equipment start-up and shutdown status and load allocation based on new real-time data to ensure that the scheduling strategy can adapt to real-time changes.

5. The energy-saving scheduling method for secondary water supply pump stations based on MPC and time-of-use pricing as described in claim 1, characterized in that, Load allocation adjustment includes: assessing the health status of equipment based on the equipment health index obtained from the equipment health assessment module; prioritizing load allocation to equipment with better health status and reducing the load on equipment with poor health status to avoid overload or failure; and ensuring that overall load demand is met while extending the service life of equipment during load allocation adjustment.

6. An energy-saving dispatching system for secondary water supply pump stations that coordinates MPC (Multi-Purpose Control) and time-of-use pricing, characterized in that: The energy-saving scheduling method for secondary water supply pump stations, which integrates MPC and time-of-use pricing as described in any one of claims 1-5, includes the following modules: The data acquisition module is used to collect real-time operating data of the pump room equipment, including equipment health status, load demand and external environmental data, and to preprocess the collected data. The equipment health assessment module is used to assess the health status of the equipment based on its real-time operating data using a health assessment model, generate an equipment health index, and use it as input for optimizing scheduling strategies. The load and electricity price forecasting module is used to predict future load demand using a deep learning model and to predict future electricity price fluctuations using a time series model, thus obtaining load demand forecasts and electricity price forecasts. The robust MPC controller module is used to combine equipment health index, load demand forecast and electricity price forecast results, and optimize scheduling through the robust MPC controller to obtain the optimal equipment start-up and shutdown strategy and load allocation scheme. The scheduling execution module is used to adjust the start-up and shutdown status of pump room equipment and perform load distribution according to the optimized scheduling strategy; The feedback adjustment module is used to monitor the equipment operating status, load changes and electricity price fluctuations in real time. It dynamically corrects the scheduling strategy through the feedback mechanism and adjusts the load allocation to ensure that the equipment operation matches the load demand.

Images