A permanent magnet micro-pump station energy efficiency optimization management method and system

Abstract

This invention provides a method and system for optimizing the energy efficiency of permanent magnet micro pump stations, relating to the field of intelligent energy management technology in the industrial internet. By constructing a cloud-edge-device collaborative optimization architecture, the energy efficiency management of permanent magnet micro pump stations is divided into three closely coupled time scales: long-term strategy, medium-term rolling, and instantaneous control. In the cloud, long-term scheduling planning is performed on a daily basis to generate long-term energy efficiency scheduling strategies, optimize start-stop combinations and baseline operating ranges, avoid high electricity price periods, and balance equipment fatigue. On the edge, predictive control is performed on a minute-by-minute basis, continuously solving for the optimal speed control setpoint sequence in the near future to ensure real-time operation in the high-efficiency zone. On the device side, millisecond-level adaptive control is executed to accurately track the setpoint. This not only improves the overall operating efficiency of the pump station cluster but also further reduces overall operating costs through intelligent response to peak and off-peak electricity prices, achieving dual optimization of energy efficiency and economy.

Description

Technical Field

[0001] This invention relates to the field of industrial internet intelligent energy management technology, specifically to a method and system for optimizing the energy efficiency of permanent magnet micro pump stations. Background Technology

[0002] Permanent magnet micro pump stations, with their advantages of high efficiency, high power density and compact structure, have been widely used in high-end industrial fields such as precision medical equipment, water supply, chip cooling systems, and micro-chemical processes. Currently, energy efficiency improvement technologies for permanent magnet micro pump stations are mainly developing along two paths: one is to optimize the hardware at the local level, such as by improving the electromagnetic design of the permanent magnet motor to reduce iron and copper losses, or by optimizing the impeller and flow channel structure of the pump body to reduce hydraulic losses; the other is to adopt variable frequency speed regulation technology based on fixed parameters at the control level to adjust the motor speed according to the set requirements. Existing technologies for optimizing the energy efficiency of permanent magnet micro pump stations present the following technical problems during use: Question 1: While hardware optimization for permanent magnet micro pump stations can partially improve energy efficiency, a permanent magnet micro pump station is a complex system consisting of a pump body, electronics, pipelines, valves, and control systems. Improving the efficiency of individual components can easily lead to poor system matching, preventing the high efficiency of individual components from being translated into high efficiency of the overall system operation. This results in some operating points deviating from the optimal efficiency range for a long time, leading to low efficiency. Consequently, there is a lack of global energy efficiency optimization for the system, making it impossible to perform global optimization under the premise of variable water supply demand. The coverage and systemic nature are insufficient, resulting in low efficiency. The second problem is that current control strategies rely on preset fixed models and parameters, such as the "Permanent Magnet Condensate Pump Speed ​​Control Method, System, Electronic Equipment and Storage Medium" disclosed in publication number "CN114251257A". Although it proposes a preset deviation curve for adjustment, it is still a pre-programmed static compensation. It cannot sense changes in the dynamic characteristics of the system and external disturbances. The static and rigid control strategy deteriorates under dynamic conditions, resulting in slow response, adjustment oscillation and additional energy loss. It cannot adapt to the long-term slow decline of equipment performance and cannot perform self-learning and adaptive dynamic adjustment based on the real-time system status. Summary of the Invention

[0003] To achieve the above objectives, the present invention provides the following technical solution: a method and system for optimizing the energy efficiency of a permanent magnet micro pump station, the method comprising: Acquire real-time and historical operating data of permanent magnet micro pump stations, and extract characteristic data reflecting the operating status of permanent magnet micro pump stations from real-time operating data; A long-term energy efficiency scheduling strategy for permanent magnet micro pump stations is generated based on historical operation data and characteristic data. Acquire the latest real-time operational and characteristic data, combine them with long-term energy efficiency scheduling strategies for rolling optimization, and generate a control setpoint sequence. The control parameters of the motor in the permanent magnet micro pump station are adjusted in real time according to the control setpoint sequence, and drive commands are generated. The execution effect of monitoring drive commands generates evaluation data, which is used to update the generation process of feature data and long-term energy efficiency scheduling strategies.

[0004] Furthermore, the extraction of feature data reflecting the operating status of the permanent magnet micro pump station from the real-time operating data includes: Acquire real-time operating data of the permanent magnet micro pump station, including flow rate data, rotational speed data, and current data; The real-time running data is processed to calculate static and dynamic feature values; The current operating point is determined based on flow rate data and rotation speed data. The preset optimal efficiency zone of the permanent magnet micro pump station is obtained. The Euclidean distance between the current operating point and the center point of the optimal efficiency zone is calculated as a static characteristic value. Within a preset time window, the rate of change of rotational speed is calculated as a dynamic characteristic value based on the rotational speed data; The static feature values ​​and dynamic feature values ​​are combined to form feature data.

[0005] Furthermore, the long-term energy efficiency scheduling strategy for permanent magnet micro pump stations generated based on historical operating data and feature data includes: A set of operating samples is constructed by acquiring feature data and historical operating data. Each operating sample in the set includes a set of operating parameters and corresponding real-time operating efficiency. The operating parameters include the target flow rate and the speed command. The efficiency prediction model is trained based on the running sample set. The input of the efficiency prediction model is the operating condition parameters, and the output is the predicted efficiency value. Using the predicted demand curve and efficiency prediction model for the future scheduling cycle as input, and minimizing the total system operating cost as the objective, an optimization solution is performed to generate a long-term energy efficiency scheduling strategy. The long-term energy efficiency scheduling strategy includes instruction information corresponding to multiple time points within the scheduling cycle, and the instruction information includes the target speed range.

[0006] Furthermore, the steps for obtaining the real-time operating efficiency include: Acquire historical flow data, historical pressure data, and historical power data from historical operating data; The output hydraulic power is calculated based on historical flow and pressure data, and the input electrical power is calculated based on historical electrical power data. The ratio of output hydraulic power to input electrical power is used as the actual operating efficiency.

[0007] Furthermore, the step of combining long-term energy efficiency scheduling strategies for rolling optimization to generate a control setpoint sequence includes: Obtain instruction information from long-term energy efficiency scheduling strategies and acquire the latest real-time operation data and characteristic data; Using instruction information as constraints, the latest real-time operating data and feature data as initial states, and a state-space prediction model for rolling optimization, a sequence of control setpoints is generated.

[0008] Furthermore, the control setpoint sequence is a series of speed setpoints over a future period of time. The input of the state-space prediction model is the speed setpoint, and the output is the flow prediction value. The flow prediction value and the demand flow value corresponding to the predicted demand curve in the long-term energy efficiency scheduling strategy are used together as the target for rolling optimization solution.

[0009] Furthermore, the step of adjusting the control parameters of the motor in the permanent magnet micro pump station in real time according to the control setpoint sequence and generating drive commands includes: The current speed setting value is obtained from the control setpoint sequence as the setpoint value, and the actual speed value of the permanent magnet micro pump station at the current moment is obtained. Calculate the deviation between the setpoint value and the actual rotational speed value, calculate the rate of change of the deviation relative to time, and obtain the rate of change of the deviation. Based on the deviation and the rate of change of deviation, the fuzzy control rule table is queried to obtain the control parameter adjustment amount. The actual control parameters are obtained by adding a set of basic control parameters with the control parameter adjustment amount. The driving command at the current moment is calculated based on the actual control parameters, deviation, and rate of change of deviation.

[0010] Furthermore, the control parameter adjustment amount includes the proportional parameter adjustment amount, the integral parameter adjustment amount, and the derivative parameter adjustment amount, and the actual control parameters include the actual proportional parameter, the actual integral parameter, and the actual derivative parameter; The fuzzy control rule table defines the mapping relationship between deviation, deviation change rate and control parameter adjustment amount; The basic control parameters are initialized based on the static feature values ​​in the feature data.

[0011] Furthermore, the monitoring of the execution effect of the driving instructions generates evaluation data, including: Within a preset evaluation period, the control setpoint sequence, the corresponding drive commands, and the real-time operating data obtained after the drive commands are executed are recorded to generate recorded sample data. Key evaluation indicators are calculated based on recorded sample data, including setpoint tracking error and equipment operating efficiency. Key assessment indicators are linked with corresponding characteristic data and long-term energy efficiency scheduling strategies to form assessment data. The evaluation data is added to the historical database, and the parameters of the efficiency prediction model and the state-space prediction model are corrected based on the evaluation data.

[0012] A permanent magnet micro pump station energy efficiency optimization management system, the system including a data acquisition and processing module, an edge optimization control module and a cloud strategy management module; The data acquisition and processing module is used to acquire real-time and historical operating data of the permanent magnet micro pump station, and extract characteristic data reflecting the operating status of the permanent magnet micro pump station from the real-time operating data. The edge optimization control module is used to acquire the latest real-time operating data and feature data, combine them with long-term energy efficiency scheduling strategies, perform rolling optimization, generate a control setpoint sequence, adjust the control parameters of the motor in the permanent magnet micro pump station in real time according to the control setpoint sequence, and generate drive commands. The cloud-based strategy management module is used to generate long-term energy efficiency scheduling strategies for permanent magnet micro pump stations based on historical operating data and feature data, monitor the execution effect of drive commands to generate evaluation data, and update feature data and the generation process of long-term energy efficiency scheduling strategies based on evaluation data. The cloud-based strategy management module sends the generated long-term energy efficiency scheduling strategy to the edge optimization control module, and the data acquisition and processing module uploads the extracted feature data to the edge optimization control module and the cloud-based strategy management module; the edge optimization control module uploads the drive instructions to the cloud-based strategy management module.

[0013] This invention provides a method and system for optimizing the energy efficiency of permanent magnet micro pump stations. It offers the following advantages: 1. This invention constructs a cloud-edge-device collaborative optimization architecture, dividing the energy efficiency management of permanent magnet micro pump stations into three tightly coupled time scales: long-term strategy, medium-term rolling, and instantaneous control. In the cloud, long-term scheduling planning is performed on a daily basis based on historical operating data and efficiency prediction models, generating long-term energy efficiency scheduling strategies, optimizing start-stop combinations and baseline operating intervals, avoiding high electricity price periods, and balancing equipment fatigue. On the edge, predictive control is performed on a minute-by-minute basis using a state-space prediction model. Within the framework of the long-term energy efficiency scheduling strategy, the optimal speed control setpoint sequence for the near future is continuously solved to ensure real-time operation within the high-efficiency zone. On the device side, millisecond-level adaptive control is executed to accurately track the setpoint. This multi-scale collaborative approach breaks through the limitations of traditional local optimization, coordinating operational objectives across different time dimensions from a global perspective, resolving the contradiction between steady-state efficiency and dynamic response. This not only improves the overall operating efficiency of the pump station cluster but also further reduces overall operating costs through intelligent response to peak and off-peak electricity prices, achieving dual optimization of energy efficiency and economy.

[0014] 2. This invention constructs a complete learning closed loop that spans data perception, feature extraction, model updating, and control decision-making. By extracting feature data reflecting the health and dynamic characteristics of permanent magnet micro pump stations in real time, it provides a basis for online self-tuning of control parameters, facilitating flexible responses to various dynamic operating conditions, reducing flow overshoot under load changes, and significantly improving transient stability and process quality. Simultaneously, the execution effect of drive commands is transformed into evaluation data, continuously fed back to the cloud. This evaluation data is used to periodically correct and reconstruct the parameters of the efficiency prediction model and the state-space prediction model, enabling the prediction model to track the actual degradation of equipment performance and changes in operating conditions. This transforms the entire optimization management from a static program that is fixed upon deployment into an intelligent organism capable of self-perception, self-evaluation, and self-adjustment. It not only optimizes in the current moment but also continuously learns, allowing optimization strategies and models to co-evolve throughout the system's entire lifecycle. This effectively combats the natural degradation of equipment performance, maintaining a highly efficient and stable operating state from the initial stage to the long term, achieving a fundamental leap from one-time optimization to continuous optimization. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the steps of an energy efficiency optimization management method for a permanent magnet micro pump station according to the present invention. Figure 2 This is a data transmission flowchart of an energy efficiency optimization management method for permanent magnet micro pump stations according to the present invention; Figure 3 This is an architecture diagram of an energy efficiency optimization management system for permanent magnet micro pump stations according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] like Figures 1 to 2 As shown, a method for optimizing energy efficiency management of permanent magnet micro pump stations includes: Step S100: Obtain real-time and historical operating data of the permanent magnet micro pump station, and extract feature data reflecting the operating status of the permanent magnet micro pump station from the real-time operating data; First, real-time operating data of the permanent magnet micro pump station is acquired. This real-time operating data includes flow rate data, speed data, and current data. This data is primarily obtained through high-precision sensors installed at key locations within the permanent magnet micro pump station. These sensors include flow sensors (such as electromagnetic flowmeters or turbine flowmeters) for real-time acquisition of instantaneous flow rate data at the pump outlet; speed sensors (such as Hall effect sensors or photoelectric encoders) installed on the motor shaft for real-time acquisition of actual motor speed data; and current sensors (such as Hall effect current sensors) connected in series in the main circuit of the motor drive circuit for real-time acquisition of phase current or bus current data during motor operation. In addition to flow rate, speed, and current data, to comprehensively reflect the pump station's operating status and health, real-time operating data should also include inlet and outlet pressure data acquired through pressure transmitters; temperature data of key parts of the motor and pump body acquired through thermocouples or resistance temperature detectors; three-phase voltage data for calculating complete electrical power; and vibration data acquired through vibration acceleration sensors for monitoring mechanical conditions. All these data together constitute the real-time operating dataset. Secondly, the real-time running data is processed to calculate static and dynamic feature values; The current operating point is determined based on flow rate and rotational speed data. The preset optimal efficiency zone of the permanent magnet micro pump station is obtained. The Euclidean distance between the current operating point and the center point of the optimal efficiency zone is calculated as a static characteristic value. The static characteristic value is a scalar value calculated based on the current steady-state operating conditions, reflecting the degree of deviation from the preset optimal efficiency zone. It is used to quantify the static energy efficiency level of the current operating state. The larger the Euclidean distance, the lower the energy efficiency, providing a steady-state benchmark for long-term scheduling and parameter initialization. Within a preset time window, the rate of change of rotational speed is calculated based on the rotational speed data as a dynamic characteristic value. The dynamic characteristic value is a scalar value that reflects the speed of change of operating status and is used to capture transient response characteristics and dynamic behaviors, such as the degree of acceleration, deceleration or load fluctuation. Finally, the static and dynamic feature values ​​are combined to form feature data. The calculated static and dynamic feature values ​​are normalized to eliminate the influence of dimensions and scaled to the interval of [0,1] or [-1,1]. Normalization can be based on the maximum and minimum values ​​in historical statistics or the theoretical range. The two normalized scalar values ​​are concatenated in a fixed order to form a two-dimensional feature vector, which is directly used as feature data reflecting the steady-state deviation and dynamic intensity of the permanent magnet micro pump station.

[0018] The current operating point refers to the representation of the steady-state operating condition of the permanent magnet micro pump station at a certain instant on the performance curve. It is usually determined by one or more pairs of key operating parameter coordinates. In this scheme, the current operating point is determined based on flow rate data and rotational speed data, by synchronously collecting the instantaneous flow rate value at the current moment. and the corresponding actual motor speed value This set of values As a two-dimensional coordinate point, this two-dimensional coordinate point represents the flow output condition at the current speed, that is, the current working point. By using the current working point setting method, the relationship between the input speed and the output flow is directly associated, avoiding the need to directly measure the pressure. The optimal efficiency zone is an operating range predetermined by the manufacturer through pump type testing. Within this range, the overall operating efficiency of the pump station (the combination of hydraulic and motor efficiency) exceeds a set threshold, such as above 90% of the maximum efficiency. This is typically represented as a closed graphical region, such as an ellipse or polygon, on a two-dimensional performance curve plotted on flow rate and rotational speed. The optimal efficiency zone can be obtained by consulting the equipment performance manual, analyzing performance curve data points provided by the manufacturer, or through preliminary experimental testing and fitting efficiency contour lines. The center point of the optimal efficiency zone is the center coordinate of this closed graphical region, or it can be directly taken as the average flow rate of all known high-efficiency sample points within the region. and average speed Coordinates of the center point ; Current work point Center point of the optimal efficiency zone Euclidean distance The calculation formula is: ; This formula calculates the straight-line distance between two points on a two-dimensional plane. This represents the square of the difference between the current flow and the optimal flow. This represents the square of the difference between the current speed and the optimal speed. Adding these two squared terms and then taking the square root yields a scalar distance value that integrates the deviations of both flow rate and speed. The larger the scalar distance value, the farther the current operating point is from the center of the high-efficiency zone, and the lower the static energy efficiency potential. Using Euclidean distance as a static feature value, it is possible to unbiasedly integrate two parameters of different dimensions and orders of magnitude, flow rate and speed, into a single, comparable scalar value through the sum of squares and the square root operation, directly quantifying the degree of deviation. The preset time window is a recent period of fixed length, such as the past 5 or 10 seconds; the rate of change of rotational speed. The value is obtained by calculating the slope of a linear regression or simple difference of the rotational speed data sequence within the time window; m rotational speed samples are collected at equal intervals within the time window. Calculate the rate of change , where t is the timestamp; This is the first rotational speed sample; This is the m-th rotational speed sample; This is the timestamp corresponding to the first rotational speed sample. The timestamp corresponding to the m-th speed sample is denoted as . The speed change rate directly reflects the degree of urgency of the motor's execution of speed commands and the system's response speed to external disturbances. A large absolute value of the speed change rate means that the system is undergoing a violent acceleration or deceleration process. At this time, dynamic losses increase, control difficulty increases, and the requirements for the controller's response characteristics are different from those in steady state. Therefore, the speed change rate effectively characterizes the transient dynamic characteristics of the system.

[0019] Step S200: Generate a long-term energy efficiency scheduling strategy for permanent magnet micro pump stations based on historical operating data and feature data; Step S201: Obtain feature data and historical operating data to construct an operating sample set. Each operating sample in the operating sample set includes a set of operating parameters and corresponding real-time operating efficiency. The operating parameters include the target flow rate and the speed command. Constructing the operating sample set includes the following processes: First, extract historical operating data for a past period (such as the past three months) from the historical database. The historical operating data includes historical flow data, historical pressure data, and historical power data. Obtain the flow target value and speed command for that period as operating parameters. Clean and match the acquired data to ensure timestamp alignment and remove obviously abnormal or deviated data points to obtain valid data points. Then, the real-time operating efficiency is calculated based on the effective data points; historical traffic data at the same timestamp is extracted. Historical export pressure data and historical import pressure data Normally, if the inlet pressure is atmospheric pressure, then It can be considered as 0; Calculate the net head H. , where ρ is the fluid density and g is the gravitational acceleration; Calculate the output hydraulic power , The output hydraulic power characterizes the effective power of the pump station in doing work on the fluid. Obtain historical voltage data at the same timestamp from historical operational data. and historical current data Calculate the input electrical power , In this context, PF represents the power factor, which can be calculated or estimated from the phase difference between voltage and current. Input power characterizes the total electrical power consumed by the pumping station from the power grid and is a direct indicator of its energy cost. For pumping stations driven by three-phase permanent magnet synchronous motors, it is typically necessary to collect three-phase voltage and current data using the formula... ;in Line voltage, Where PF is the line current and PF is the power factor. The real-time operating efficiency is calculated as the ratio of output hydraulic power to input electrical power multiplied by 100%. The real-time operating efficiency reveals the degree of loss in the energy conversion process and quantitatively reflects the true energy conversion efficiency of the pumping station under certain historical operating parameters. It is the target variable for the efficiency prediction model to learn. Finally, each set of operating parameters is combined with the calculated actual operating efficiency to form a complete operating sample. The sum of all operating samples constitutes the operating sample set. The operating sample set is a structured historical data set used to train the efficiency prediction model. Each operating sample records the operating parameters of the pump station under specific operating conditions and its corresponding actual operating efficiency reflecting energy efficiency.

[0020] Among them, the operating parameters are a set of key setpoints describing the required or instructed operating status of the pumping station. The flow target value comes from the upper management system or historical records and is the required delivery flow rate of the pumping station at a certain moment. The speed command is the speed setpoint calculated by the control system and issued to the motor driver in order to achieve the flow target value. The operating parameters are usually obtained directly from the historical command log of the control system and serve as input features for the efficiency prediction model. The efficiency prediction model learns the mapping relationship between the actual operating efficiency that the pumping station can achieve under a specific combination of commands, that is, how the pumping station is required to work, thereby providing predictive capabilities for future scheduling optimization.

[0021] Step S202: Train an efficiency prediction model based on the running sample set. The input of the efficiency prediction model is the operating condition parameters, and the output is the predicted efficiency value. The efficiency prediction model is a mathematical mapping function or calculation model that takes the operating condition parameters as input and outputs the predicted efficiency value that the pump station is expected to achieve under the operating condition parameters. In this embodiment, a trained shallow neural network containing multiple nonlinear transformations is used. The efficiency prediction model includes an input layer with two neurons, which receive the flow target value and speed command from the running sample set, respectively; multiple hidden layers, each containing several neurons, each neuron performing a weighted summation of the output of the previous layer and applying a nonlinear activation function to learn the complex nonlinear relationship between the operating condition parameters and efficiency; and an output layer with one neuron, which linearly combines the output of the last hidden layer to finally output a predicted efficiency value in the range of [0,1] or [0,100%]. The efficiency prediction model encapsulates the efficiency performance spectrum of the pump station under different operating condition parameters. The following process is included when training the efficiency prediction model: First, the constructed running sample set is randomly divided into a training set and a validation set in a ratio (e.g., 8:2); the operating parameters in the training set are normalized so that their mean is 0 and their standard deviation is 1. Next, the neural network is initialized by assigning a random small value close to zero to all connection weights and biases from the input layer to the first hidden layer, between each hidden layer, and from the last hidden layer to the output layer. Next, the iterative training loop begins: For each running sample in the training set, the operating parameters are input into the efficiency prediction model. This input is received by the input layer and passed to the first hidden layer. In each neuron of the first hidden layer, the input is weighted and summed, and a bias is added. Then, a non-linear transformation is performed using the ReLU activation function to generate the output of the neuron in the first hidden layer. This output is used as the input of the next hidden layer and propagates sequentially. When the propagation reaches the neuron in the output layer, a final weighted sum is performed and added to the bias. Here, the ReLU activation function is not used; instead, the Sigmoid activation function is used to limit the final output value to between 0 and 1. This value is the predicted efficiency value of the sample. Then, the squared error between the predicted efficiency value and the real-time operating efficiency is calculated. The squared error is used as the basis for guiding the adjustment of neural network parameters. The real-time operating efficiency is obtained from the operating samples corresponding to the operating parameters input in the iterative loop. Using the gradient descent principle, the value of each weight and bias in the neural network is fine-tuned along the direction of decreasing squared error. The gradient of the squared error with respect to each parameter is calculated. The Adam optimizer is used to update all weights according to the gradient to reduce the squared error and make the predicted efficiency value approach the real-time operating efficiency. Among them, the Adam optimizer is a widely used optimization algorithm in deep learning. It can be used directly in the default neural network and can adaptively adjust the learning step size of each parameter, thereby updating all weights and biases more stably and efficiently. Finally, the above process is repeated across all running samples in the training set, which is called a training cycle. After each cycle, an independent validation set is input into the current neural network, and the average prediction error on the validation set is calculated. This process is repeated for multiple training cycles, and the change in the average prediction error on the validation set is monitored. When the average prediction error on the validation set no longer decreases or even begins to increase after several consecutive training cycles, it indicates that the model may be overfitting, and training should be stopped. At this point, the weights and biases of each layer within the neural network have been optimized to a stable set of values, which can better capture the general mapping relationship between operating parameters and operating efficiency. The efficiency prediction model is then considered trained. When the efficiency prediction model is put into use, new normalized operating parameters are input, and after the same forward propagation process, accurate predicted efficiency values ​​can be obtained at the output layer.

[0022] Step S203: Using the predicted demand curve and efficiency prediction model for the future scheduling cycle as input and minimizing the total system operating cost as the objective, perform optimization to generate a long-term energy efficiency scheduling strategy. The long-term energy efficiency scheduling strategy includes instruction information corresponding to multiple time points within the scheduling cycle, and the instruction information includes the target speed range.

[0023] Optimization is a mathematical programming process aimed at minimizing the total operating cost of the system. The objective includes direct energy costs (based on time-of-use pricing and predicted power consumption) as well as potentially hidden equipment start-up and shutdown losses or fatigue costs. The solution process is as follows: First, the future scheduling cycle is discretized into T time steps, and the decision variable is defined as the planned rotational speed of the pump station in each time period t. Define binary integer decision variables as the pump station start-up and shutdown status for each time period t. (1 indicates running, 0 indicates stopped); Then, construct the objective function to minimize the system cost. ;

[0024] in, The input electrical power is calculated from the efficiency prediction model. The function; This refers to the electricity price during time period t. This is the cost per startup; An auxiliary binary variable indicating whether startup occurs during time period t. ; The constraint is that the requirement satisfies the following conditions: ;in, It is based on the planned rotation speed Flow rate estimated from pump characteristic curves; This is the actual required flow rate; the upper and lower limits of the rotational speed are constrained as follows: Equipment protection constraints include measures to avoid frequent start-stop cycles, such as minimum continuous operation and downtime constraints; speed change rate constraints are... ; Finally, the efficiency prediction model is used as the key internal mapping between rotational speed and efficiency / flow rate in the objective function and constraints. The above objective function and constraints are input into a commercial or open-source MILP solver, such as CPLEX or Gurobi, for solving. This solving process is performed directly by computer software to obtain the rotational speed command with the lowest total cost for each time period in the future. After summarizing the rotational speed commands, a set of rotational speed command sequences that minimize the total cost and cover the entire scheduling cycle is obtained, which is the long-term energy efficiency scheduling strategy.

[0025] The future scheduling cycle is usually set to 24 hours, which is in line with the daily shift planning, time-of-use electricity pricing cycle (peak, flat, valley) and the usual span of daily demand forecasting in most industrial scenarios, and can effectively coordinate the production plan, energy consumption cost and equipment maintenance within a day. The forecast demand curve is a predicted trajectory of the fluid flow rate required by the pumping station over time within a future scheduling cycle. It includes the predicted flow rate demand values ​​at a series of equally spaced time points (e.g., every 15 minutes) within the future cycle. The forecast demand curve is generated by analyzing historical water consumption patterns, fluid consumption patterns, combining production planning and scheduling, and integrating external information such as weather forecasts. It provides a core and forward-looking load target for long-term scheduling, enabling optimization strategies to move beyond passive responses and proactively plan, arrange high-efficiency operating intervals in advance, and avoid periods of high electricity prices. Long-term energy efficiency scheduling strategy is a global and forward-looking operation guidance plan for permanent magnet micro pump stations based on the prediction of future demand and global cost optimization. It covers a future scheduling cycle (such as 24 hours) and includes optimization instruction information corresponding to each decision point (such as every 15 minutes) within the cycle. From the perspective of global optimization, it determines when the equipment should operate in which efficiency range to balance energy consumption, electricity costs and equipment lifespan, and sets boundaries and reference targets for the next level of real-time rolling optimization. The instruction information is a control guide issued to the execution layer for a specific future point in time in the long-term energy efficiency scheduling strategy. In addition to the target speed range, the instruction information also includes a planned operating status flag (such as "running", "standby", "suggested shutdown") and the associated expected operating efficiency threshold. The target speed range gives the recommended speed range for achieving economical and efficient operation within this time period, providing the core constraint for the rolling optimization of step S300. The planned operating status flag guides the basic start-up and shutdown of the equipment. The expected operating efficiency threshold can be used for subsequent performance evaluation and comparison. These pieces of information work together to transform the macroeconomic economic strategy into an executable physical operating boundary that takes efficiency into account.

[0026] Step S300: Obtain the latest real-time operating data and feature data, combine them with long-term energy efficiency scheduling strategies for rolling optimization, and generate a control setpoint sequence; Step S301: Obtain instruction information from the long-term energy efficiency scheduling strategy, and obtain the latest real-time operation data and feature data; Step S302: Using instruction information as constraints, the latest real-time running data and feature data as the initial state, and a rolling optimization solution using a state-space prediction model, a control setpoint sequence is generated.

[0027] The control setpoint sequence is a series of speed setpoints over a future period. The input of the state-space prediction model is the speed setpoint, and the output is the flow prediction value. The flow prediction value and the demand flow value corresponding to the predicted demand curve in the long-term energy efficiency scheduling strategy are used together as the target for rolling optimization.

[0028] The state-space prediction model is a mathematical model used to describe the dynamic characteristics of a permanent magnet micro pump station. It precisely expresses the relationship between the system's internal states (such as current rotational speed, historical flow inertia) and the rotational speed setpoint and predicted flow rate using a set of discrete-time linear or linearized difference equations. The state-space prediction model includes a state vector, a state transition matrix, a control input matrix, and an output matrix. The state vector contains the current rotational speed and the flow rate values ​​at several past moments. The state transition matrix describes how the state evolves from the previous moment to the current moment. The control input matrix describes the impact of changes in the rotational speed setpoint on the state. The output matrix is ​​used to calculate the predicted flow rate value from the current state. The rolling optimization solution using a state-space prediction model includes the following process: Step S3021: Initialize the state-space prediction model. At the beginning of each rolling optimization cycle (time k), based on the latest real-time running data and feature data, construct or estimate the initial value of the current state vector of the state-space model. , ; in, This represents the system state vector at time k; This represents the actual rotational speed measured at time k, which is directly obtained from the speed sensor. This represents the actual flow rate measured at time k, which is directly obtained from the flow sensor. The historical flow value at time k-1 (the previous sampling period) is obtained from the cache. The purpose of introducing historical flow is to contain the dynamic inertial information of the system in the state. Step S3022: Based on the internal architecture of the state-space model, construct an optimization problem in the future finite time domain, set constraints based on the model architecture, and define the optimization problem and set the constraint conditions; Model architecture constraints: The future dynamics of the system must follow the following equations: State transition equation: ; in, A represents the predicted system state vector at time k+1; B represents the state transition matrix, obtained from system identification, describing how the system state evolves spontaneously; C represents the control input matrix, obtained from system identification, describing the control input... How to influence state changes; This represents the control input at time k, specifically the speed setpoint. ; Output equation: ; in, This represents the system output at time k, specifically the predicted flow rate. C represents the output matrix, which is obtained by the system and describes how the output is extracted from the state vector. The optimization problem is defined as follows: Given a prediction time domain of... Step, control time domain is Step, usually The decision variables are a series of control inputs within the future control time domain: The objective function is typically designed to minimize the deviation between the predicted output and the reference trajectory, while suppressing drastic changes in the control input. The formula for minimizing the objective function J is as follows:

[0029] Where J represents the objective function (performance metric) to be minimized. Indicates based on model and decision variables The predicted flow output at time k+i; The reference value for flow at time k+i is derived from the predicted demand curve in the long-term energy efficiency scheduling strategy. λ represents the control weight coefficient (a positive scalar), used to balance tracking accuracy and control stability; It represents the increment of the control input, and its sum of squares term is used to penalize excessive control changes, making the control action smoother; The constraints include: System dynamic constraints: All predicted states and outputs must strictly satisfy the above state-space model equations; The input constraints are: , ,in, and The target speed range in the long-term scheduling strategy Sure; Input rate constraints: That is, there is a physical upper limit to the range of change of control quantity in adjacent cycles; Step S3023: Perform rolling optimization solution. At each sampling time k, the above problem (with a quadratic objective function and linear constraints) is constructed into a standard quadratic programming (QP) problem. The QP solver is called to solve it online to obtain the optimal control sequence.

[0030] Take only the first element in the optimal control sequence The speed setpoint that should be issued immediately at this moment. At the same time, the entire sequence The output is a sequence of control setpoints for use by subsequent logic. At time k+1, the state estimate is updated using the new measurements. Scroll the entire optimization window forward one step and repeat the process from steps S3021 to S3023 above to perform a new round of solution and implementation. This process ensures that the system dynamically generates a sequence of optimal control setpoints that allows the flow output to accurately and smoothly track the demand curve, while satisfying the long-term policy boundary, by repeatedly solving constrained optimization problems online.

[0031] Step S400: Adjust the control parameters of the motor in the permanent magnet micro pump station in real time according to the control setpoint sequence, and generate drive commands; Step S401: Obtain the current speed setpoint value at time k from the control setpoint sequence as the setpoint value. Obtain the actual rotational speed of the permanent magnet micro pump station at time k. ; Step S402: Calculate the setpoint value Compared with actual speed value deviation , ; Calculate the rate of change of the deviation relative to time to obtain the rate of change of the deviation. ; ;in, It is the deviation from the previous control cycle. It is the control period, such as 1 millisecond. The rate of change of the deviation relative to time; Among them, the deviation reflects the magnitude and direction of the current tracking error. A positive deviation indicates that the speed is too low, and a negative deviation indicates that the speed is too high. The deviation change rate reflects the trend and speed of error change. A positive deviation change rate indicates that the deviation is increasing, and a negative deviation change rate indicates that the deviation is decreasing.

[0032] Step S403: Based on the deviation and the rate of change of deviation, query the fuzzy control rule table to obtain the control parameter adjustment amount, add a set of basic control parameters with the control parameter adjustment amount to obtain the actual control parameters; The fuzzy control rule table defines the mapping relationship between deviation, deviation rate of change, and control parameter adjustment in tabular form. The fuzzy control rule table includes the deviation (E) and deviation rate of change (E...). C The fuzzy linguistic values ​​(such as "negative large NB", "negative medium NM", "zero ZO", "positive medium PM", "positive large PB") are used as inputs, and the fuzzy linguistic values ​​of the proportional parameter adjustment ΔKp, integral parameter adjustment ΔKi, and derivative parameter adjustment ΔKd are used as outputs for all possible combinations of rules. The fuzzy control rule table is set based on the dynamic characteristics of the permanent magnet motor drive system and the domain knowledge of the control engineer. The core principle of the setting is: based on the data of the deviation value and the deviation change rate, the deviation interval and the deviation change rate interval are divided into large deviation, medium deviation, and small deviation. When the deviation is large, the proportional parameter is increased first to quickly eliminate the deviation; when the deviation is medium and the deviation change rate is large, the derivative parameter is enhanced to suppress overshoot; when the deviation is small and the change rate is small, the integral parameter is enhanced to eliminate steady-state error. Through the fuzzy control rule table, the precise numerical deviation and change rate are transformed into a reasonable, nonlinear parameter adjustment strategy through fuzzification, rule reasoning, and defuzzification, so that the PID controller can adapt to different dynamic operating conditions. When formulating the fuzzy control rule table, the universe of discourse for the deviation and the rate of change of deviation is determined (based on the historical maximum deviation), and divided into 5-7 overlapping fuzzy subsets (such as NB, NM, ZO, PM, PB), and the membership function (such as a trigonometric function) for each subset is defined. Similarly, the output universe of discourse and fuzzy subsets for the three adjustment variables are determined. Then, based on the above control principles, the complete rule base is manually formulated in the form of "IF-THEN", for example: "IF E is PB AND E..." C"isZO, THEN ΔKp is PB, ΔKi is NB, ΔKd is ZO"; the rules are verified and fine-tuned through simulation or experimentation. The control parameter adjustment amounts, when the query is generated, include: First, the calculated precise deviation E (k) and the rate of change of deviation e c(k) Based on the preset membership function, they are converted into membership degrees for the input fuzzy subsets (such as NB, ZO, etc.). Secondly, activate all relevant rules in the fuzzy control rule table. For each activated rule, prune or scale the output fuzzy set of its conclusion part according to the matching degree of its preconditions. Finally, the output fuzzy sets of all activated rules are superimposed and aggregated to obtain three total output fuzzy sets for ΔKp, ΔKi, and ΔKd. Each total output fuzzy set is then defuzzified to calculate the precise values ​​of ΔKp, ΔKi, and ΔKd, thus obtaining the control parameter adjustments. These control parameter adjustments include the proportional parameter adjustment ΔKp, the integral parameter adjustment ΔKi, and the derivative parameter adjustment ΔKd. The control parameter adjustments are the changes in the proportional, integral, and derivative parameters that need to be added to or reduced above the basic control parameters within the current control cycle. This is used to achieve online adaptive tuning of the PID parameters, enabling the controller's dynamic response characteristics (such as speed and stability) to match the system's current tracking state (characterized by deviation and the rate of change of deviation) in real time.

[0033] The basic control parameters are a set of initial static parameters for the PID controller, including the basic proportional parameters. Basic integral parameters and fundamental differential parameters The basic control parameters are initialized based on the static feature values ​​in the feature data extracted in step S100. When generating the basic control parameters, firstly, different static feature values ​​corresponding to different typical steady-state operating points are established in advance through experiments or simulations. A set of PID parameters that yields good stability and response speed is obtained; then, these data pairs are fitted or tabulated, and during online operation, the results are calculated in real time. The value is dynamically set by looking up a table or interpolating; for example, when When the system is large (operating in an inefficient region or under extreme conditions), the system dynamics may change, affecting the basic proportional parameters. Set it slightly larger to ensure responsiveness; Actual control parameters include actual proportional parameters Actual integration parameters and actual differential parameters The actual control parameters are the PID parameter values ​​ultimately used to calculate the drive command at the current moment. They integrate the basic control parameters based on steady-state conditions and the control parameter adjustments based on transient tracking states to form a set of optimal control parameters, thereby achieving accurate and stable setpoint tracking and directly performing vector addition. ; ; This yields the proportional, integral, and derivative coefficients applicable to the current precise control cycle.

[0034] Step S404: Calculate the drive command at the current moment based on the actual control parameters, deviation, and rate of change of deviation; the drive command generation includes the following process: First, use the actual control parameters and the current deviation. and rate of change of deviation The incremental PID algorithm is used to calculate the change in the required control output at time k in the current control cycle compared to the previous cycle, which is the control increment. and output value ; ; in, This represents the increment of the control quantity at time k. , , These represent the actual proportional parameter, the actual integral parameter, and the actual differential parameter, respectively. This represents the rotational speed deviation at time k. The integral term acts on this term, aiming to continuously accumulate and eventually eliminate the steady-state error. , These represent the speed deviation values ​​for the previous (k-1 time) and the previous two (k-2 time) control cycles, respectively, representing historical information about the deviation. The first difference, representing the deviation, is approximately the rate of change of the deviation and reflects the trend of error change. The proportional term acts on this term to suppress the trend of error change. The second-order difference representing the deviation is approximately equal to the rate of change of the deviation (acceleration); the differential term acts on this term to predict the future trend of the error and apply a suppressive force in advance, thereby improving the dynamic response of the system and reducing overshoot; Then, the calculated control quantity u(k), representing the required torque or current scalar, is limited to ensure it is within the safe range of the driver. Finally, the control quantity after limiting. According to the torque constant of the permanent magnet synchronous motor Converted to q-axis current command: ;in, This is the final generated drive command, which is sent to the current loop of the motor driver.

[0035] Among them, the drive command is the control signal that is ultimately sent to the permanent magnet micro pump station motor driver. In typical vector control, its core is the q-axis current command value. The target speed obtained from the upper-level optimization is transformed into a direct and executable electrical command for the motor torque through an actual PID controller that incorporates adaptive parameters, thereby achieving closed-loop precise control of the motor speed.

[0036] Step S500: Monitor the execution effect of drive commands to generate evaluation data. The evaluation data is used to update the generation process of feature data and long-term energy efficiency scheduling strategy.

[0037] Step S501: Within the preset evaluation period, record the control setpoint sequence, the corresponding drive commands, and the real-time running data obtained after the drive commands are executed, and generate recorded sample data; The preset evaluation period is a fixed time period that is much longer than the control period and the rolling optimization period. It is the end of a complete long-term scheduling cycle (such as 24 hours) or a fixed equipment operation period (such as weekly). It is used to systematically evaluate and learn from the overall control effect and strategy execution over a relatively long period of time.

[0038] Recorded sample data is a complete snapshot of data, including input commands, control actions, and system responses, recorded at a fixed frequency (e.g., every minute) or at key event points within an evaluation cycle; it includes timestamps, control setpoint sequences, actual issued drive commands, and real-time operating data fed back by sensors after the drive commands are executed.

[0039] Step S502: Calculate key evaluation indicators based on recorded sample data. Key evaluation indicators include setpoint tracking error and equipment operating efficiency. Associate the key evaluation indicators with corresponding feature data and long-term energy efficiency scheduling strategies to form evaluation data. The setpoint tracking error is represented by the root mean error, and the speed setpoint values ​​at all time points are extracted from the recorded sample data. and the corresponding actual speed value Calculate the root mean error (RMSE);

[0040] Where N represents the total number of samples collected during the evaluation period; i represents the i-th sampling time. This represents the speed setpoint (target value) issued by the control setpoint sequence at the i-th sampling time. It originates from the upper-level optimization results and is the ideal speed that the motor is expected to achieve. This represents the actual motor speed value measured by the sensor at the i-th sampling time. Represents the instantaneous tracking error at a single sampling moment; The square operation represents the instantaneous error, which is used to eliminate the effect of positive and negative errors canceling each other out, ensuring that all deviations contribute positive values; To assess equipment operating efficiency, the actual flow rate, real-time pressure, and real-time electrical power at each time point are extracted from the recorded sample data. The hydraulic power and real-time operating efficiency at each moment are calculated separately, and then the arithmetic average is taken over the entire evaluation period to obtain the average operating efficiency.

[0041] Step S503: Add the evaluation data to the historical database, and perform parameter correction on the efficiency prediction model and state-space prediction model based on the evaluation data.

[0042] When calibrating the parameters of the efficiency prediction model, the efficiency prediction model update process is triggered when new evaluation data is added to the historical database. The recorded sample data in the new evaluation period is added to or partially replaces the original running sample set to form an updated training dataset. Using this updated training dataset, the internal weight parameters of the original efficiency prediction model are retrained or fine-tuned according to the complete training process described in step S202, so that the efficiency prediction model can learn the latest changes that may occur in the equipment performance (such as the shift in efficiency characteristics due to wear). When calibrating the parameters of the state-space prediction model, the calibration of the state-space prediction model focuses on the calibration of its internal matrices. From the latest recorded sample data, a segment of time-series data containing rich dynamic changes (such as changes in speed setpoint and load fluctuations) of input (speed command / drive command) and output (actual flow) is extracted. This time-series data is used as a new system identification dataset to re-estimate and update the matrix parameters of the state-space model. In this way, the state-space prediction model can more accurately reflect the current dynamic characteristics of the system and improve the prediction accuracy of rolling optimization. The historical database is a structured, time-series data warehouse that includes raw historical operating data, constructed operating sample sets, generated evaluation data, and various model parameters used by the system, such as efficiency prediction model weights and state-space model matrices. It provides reliable data support for all energy efficiency optimization processes and enables the system to track performance degradation and learn continuously by comparing historical and current data.

[0043] like Figure 3As shown, a permanent magnet micro pump station energy efficiency optimization management system is disclosed. The system includes a data acquisition and processing module, an edge optimization control module, and a cloud strategy management module. The data acquisition and processing module is used to acquire real-time and historical operating data of the permanent magnet micro pump station, and extract characteristic data reflecting the operating status of the permanent magnet micro pump station from the real-time operating data. The edge optimization control module is used to acquire the latest real-time operating data and feature data, combine them with long-term energy efficiency scheduling strategies, perform rolling optimization, generate a control setpoint sequence, adjust the control parameters of the motor in the permanent magnet micro pump station in real time according to the control setpoint sequence, and generate drive commands. The cloud-based strategy management module is used to generate long-term energy efficiency scheduling strategies for permanent magnet micro pump stations based on historical operating data and feature data, monitor the execution effect of drive commands to generate evaluation data, and update feature data and the generation process of long-term energy efficiency scheduling strategies based on evaluation data. The cloud-based strategy management module distributes the generated long-term energy efficiency scheduling strategy to the edge optimization control module, while the data acquisition and processing module uploads the extracted feature data to both the edge optimization control module and the cloud-based strategy management module. The edge optimization control module then uploads the drive commands to the cloud-based strategy management module.

[0044] In this embodiment, by constructing a cloud-edge-device collaborative optimization architecture, the energy efficiency management of permanent magnet micro pump stations is divided into three closely coupled time scales: long-term strategy, medium-term rolling, and instantaneous control. In the cloud, long-term scheduling planning is performed on a daily basis based on historical operating data and efficiency prediction models to generate long-term energy efficiency scheduling strategies, optimize start-stop combinations and baseline operating intervals, avoid high electricity price periods, and balance equipment fatigue. On the edge, predictive control is performed on a minute-by-minute basis using a state-space prediction model. Within the framework of the long-term energy efficiency scheduling strategy, the optimal speed control setpoint sequence for the near future is solved on a rolling basis to ensure real-time operation in the high-efficiency zone. On the device side, millisecond-level adaptive control is executed to accurately track the setpoint. By using a multi-scale collaborative method, the limitations of traditional local optimization are broken, and the operating objectives at different time dimensions are coordinated from a global perspective. This resolves the contradiction between steady-state efficiency and dynamic response, not only improving the overall operating efficiency of the pump station cluster but also further reducing overall operating costs through intelligent response to peak and valley electricity prices, achieving dual optimization of energy efficiency and economy.

[0045] This invention constructs a complete learning loop encompassing data perception, feature extraction, model updating, and control decision-making. By extracting feature data reflecting the health and dynamic characteristics of permanent magnet micro pump stations in real time, it provides a basis for online self-tuning of control parameters, facilitating flexible responses to various dynamic operating conditions, reducing flow overshoot under load surges, and significantly improving transient stability and process quality. Simultaneously, it transforms the execution effect of drive commands into evaluation data, continuously feeding it back to the cloud. This evaluation data is used to periodically correct and reconstruct the parameters of the efficiency prediction model and the state-space prediction model, enabling the prediction model to track the actual degradation of equipment performance and changes in operating conditions. This transforms the entire optimization management from a static program that is fixed upon deployment into an intelligent organism capable of self-perception, self-evaluation, and self-adjustment. It not only achieves optimization in the current moment but also, through continuous learning, allows the optimization strategy and model to co-evolve throughout the system's entire lifecycle, effectively combating the natural degradation of equipment performance and maintaining a highly efficient and stable operating state from the initial stage to the long term. This represents a fundamental leap from one-time optimization to continuous optimization.

[0046] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code, which, when executed by the one or more processors, can perform the energy efficiency optimization management method and system for a permanent magnet micro-pump station as described above.

[0047] The methods and systems according to the embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as a ROM or hard disk, may store the energy efficiency optimization management method and system for permanent magnet micro pump stations provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components of the electronic device shown in this application may be omitted according to actual needs.

[0048] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0049] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing energy efficiency management of permanent magnet micro pump stations, characterized in that, The method includes: Acquire real-time and historical operating data of permanent magnet micro pump stations, and extract characteristic data reflecting the operating status of permanent magnet micro pump stations from real-time operating data; A long-term energy efficiency scheduling strategy for permanent magnet micro pump stations is generated based on historical operation data and characteristic data. Acquire the latest real-time operational and characteristic data, combine them with long-term energy efficiency scheduling strategies for rolling optimization, and generate a control setpoint sequence. The control parameters of the motor in the permanent magnet micro pump station are adjusted in real time according to the control setpoint sequence, and drive commands are generated. The execution effect of monitoring drive commands generates evaluation data, which is used to update the generation process of feature data and long-term energy efficiency scheduling strategies.

2. The energy efficiency optimization management method for a permanent magnet micro pump station according to claim 1, characterized in that, The extraction of feature data reflecting the operating status of the permanent magnet micro pump station from real-time operating data includes: Acquire real-time operating data of the permanent magnet micro pump station, including flow rate data, rotational speed data, and current data; The real-time running data is processed to calculate static and dynamic feature values; The current operating point is determined based on flow rate data and rotation speed data. The preset optimal efficiency zone of the permanent magnet micro pump station is obtained. The Euclidean distance between the current operating point and the center point of the optimal efficiency zone is calculated as a static characteristic value. Within a preset time window, the rate of change of rotational speed is calculated as a dynamic characteristic value based on the rotational speed data; The static feature values ​​and dynamic feature values ​​are combined to form feature data.

3. The method for optimizing energy efficiency management of a permanent magnet micro pump station according to claim 2, characterized in that, The long-term energy efficiency scheduling strategy for permanent magnet micro pump stations, generated based on historical operating data and characteristic data, includes: A set of operating samples is constructed by acquiring feature data and historical operating data. Each operating sample in the set includes a set of operating parameters and corresponding real-time operating efficiency. The operating parameters include the target flow rate and the speed command. The efficiency prediction model is trained based on the running sample set. The input of the efficiency prediction model is the operating condition parameters, and the output is the predicted efficiency value. Using the predicted demand curve and efficiency prediction model for the future scheduling cycle as input, and minimizing the total system operating cost as the objective, an optimization solution is performed to generate a long-term energy efficiency scheduling strategy. The long-term energy efficiency scheduling strategy includes instruction information corresponding to multiple time points within the scheduling cycle, and the instruction information includes the target speed range.

4. The energy efficiency optimization management method for a permanent magnet micro pump station according to claim 3, characterized in that, The steps for obtaining the real-time operating efficiency include: Acquire historical flow data, historical pressure data, and historical power data from historical operating data; The output hydraulic power is calculated based on historical flow and pressure data, and the input electrical power is calculated based on historical electrical power data. The ratio of output hydraulic power to input electrical power is used as the actual operating efficiency.

5. The energy efficiency optimization management method for a permanent magnet micro pump station according to claim 1, characterized in that, The step of combining long-term energy efficiency scheduling strategies for rolling optimization to generate a control setpoint sequence includes: Obtain instruction information from long-term energy efficiency scheduling strategies and acquire the latest real-time operation data and characteristic data; Using instruction information as constraints, the latest real-time operating data and feature data as initial states, and a state-space prediction model for rolling optimization, a sequence of control setpoints is generated.

6. The energy efficiency optimization management method for a permanent magnet micro pump station according to claim 5, characterized in that, The control setpoint sequence is a series of speed setpoints over a future period of time. The input of the state-space prediction model is the speed setpoint, and the output is the flow prediction value. The flow prediction value and the demand flow value corresponding to the predicted demand curve in the long-term energy efficiency scheduling strategy are used together as the target for rolling optimization.

7. The energy efficiency optimization management method for a permanent magnet micro pump station according to claim 1, characterized in that, The process of adjusting the control parameters of the motor in the permanent magnet micro pump station in real time according to the control setpoint sequence and generating drive commands includes: The current speed setting value is obtained from the control setpoint sequence as the setpoint value, and the actual speed value of the permanent magnet micro pump station at the current moment is obtained. Calculate the deviation between the setpoint value and the actual rotational speed value, calculate the rate of change of the deviation relative to time, and obtain the rate of change of the deviation. Based on the deviation and the rate of change of deviation, the fuzzy control rule table is queried to obtain the control parameter adjustment amount. The actual control parameters are obtained by adding a set of basic control parameters with the control parameter adjustment amount. The driving command at the current moment is calculated based on the actual control parameters, deviation, and rate of change of deviation.

8. The energy efficiency optimization management method for a permanent magnet micro pump station according to claim 7, characterized in that, The control parameter adjustment includes proportional parameter adjustment, integral parameter adjustment and derivative parameter adjustment, and the actual control parameters include actual proportional parameter, actual integral parameter and actual derivative parameter; The fuzzy control rule table defines the mapping relationship between deviation, deviation change rate and control parameter adjustment amount; The basic control parameters are initialized based on the static feature values ​​in the feature data.

9. The method for optimizing energy efficiency management of a permanent magnet micro pump station according to claim 1, characterized in that, The monitoring and driving instructions generate evaluation data on the execution effect, including: Within a preset evaluation period, the control setpoint sequence, the corresponding drive commands, and the real-time operating data obtained after the drive commands are executed are recorded to generate recorded sample data. Key evaluation indicators are calculated based on recorded sample data, including setpoint tracking error and equipment operating efficiency. Key assessment indicators are linked with corresponding characteristic data and long-term energy efficiency scheduling strategies to form assessment data. The evaluation data is added to the historical database, and the parameters of the efficiency prediction model and the state-space prediction model are corrected based on the evaluation data.

10. A permanent magnet micro pump station energy efficiency optimization management system, characterized in that, The system includes a data acquisition and processing module, an edge optimization control module, and a cloud strategy management module; The data acquisition and processing module is used to acquire real-time and historical operating data of the permanent magnet micro pump station, and extract characteristic data reflecting the operating status of the permanent magnet micro pump station from the real-time operating data. The edge optimization control module is used to acquire the latest real-time operating data and feature data, combine them with long-term energy efficiency scheduling strategies, perform rolling optimization, generate a control setpoint sequence, adjust the control parameters of the motor in the permanent magnet micro pump station in real time according to the control setpoint sequence, and generate drive commands. The cloud-based strategy management module is used to generate long-term energy efficiency scheduling strategies for permanent magnet micro pump stations based on historical operating data and feature data, monitor the execution effect of drive commands to generate evaluation data, and update feature data and the generation process of long-term energy efficiency scheduling strategies based on evaluation data. The cloud-based strategy management module distributes the generated long-term energy efficiency scheduling strategy to the edge optimization control module, and the data acquisition and processing module uploads the extracted feature data to the edge optimization control module and the cloud-based strategy management module. The edge optimization control module uploads the driving instructions to the cloud policy management module.

Images