Online intelligent cloud control method and cloud control platform for energy-saving and consumption-reducing management of motor

Abstract

The invention discloses an online smart cloud control method and cloud control platform for motor energy conservation and consumption reduction management. The method comprises the following steps: acquiring real-time operation parameters of all monitoring points; inputting the real-time operation parameters into the fault prediction model to obtain fault monitoring points, and screening out the fault monitoring points to obtain residual monitoring points; acquiring historical operation parameters of all the monitoring points, and constructing a power supply scene set according to the historical operation parameters; acquiring operation current of all load equipment connected with the remaining monitoring points, and determining a target characteristic curve in a database according to the operation current; if the target characteristic curve is a load characteristic curve, determining a load demand of each remaining monitoring point in each power supply scene; transmitting each load demand to an edge node; determining a current power supply scene according to the real-time operation parameters; and taking the optimal load scheduling result in the current power supply scene as a target load scheduling result, and scheduling the loads of the remaining monitoring points according to the target load scheduling result.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of cloud computing technology, and in particular to an online intelligent cloud control method and cloud control platform for motor energy saving and consumption reduction management. Background Art

[0002] With the development of industrial automation and intelligence, the power supply system, as the core support of industrial activities, undertakes the important task of providing stable power for various industrial equipment.

[0003] In the prior art, the electric power supply system mainly includes permanent magnet synchronous motors and three-phase asynchronous motors. In order to facilitate maintenance personnel to maintain the electric power supply system, the motors are usually set to fixed speeds and voltages. That is, the speed and voltage of the motors are set according to preset typical load conditions. However, when the actual load does not match the typical load, the fixed speed and voltage may cause the motor to run at an unnecessary high power, thereby wasting energy. For example, if the motor still runs at a high speed when the load is low, it will cause unnecessary energy consumption, resulting in low energy utilization efficiency of the electric power supply system.

[0004] Therefore, in response to the above problems, a method is urgently needed to maximize energy utilization efficiency. Summary of the invention

[0005] The embodiments of the present application provide an online intelligent cloud control method and cloud control platform for motor energy saving and consumption reduction management, which are used to maximize energy utilization efficiency.

[0006] To achieve the above objectives, the embodiments of the present application adopt the following technical solutions:

[0007] In a first aspect, an online intelligent cloud control method for motor energy saving and consumption reduction management is provided, which is applied to a cloud control platform, wherein the cloud control platform includes a database, wherein the database stores no-load characteristic curves and load characteristic curves of monitoring points, wherein the monitoring points include permanent magnet synchronous motors and three-phase asynchronous motors, and the method includes:

[0008] Obtaining real-time operating parameters of all monitoring points in the electric power supply system;

[0009] Inputting the real-time operating parameters into a pre-built fault prediction model to obtain a fault monitoring point, a fault type and a fault time, and filtering out the fault monitoring point from all the monitoring points to obtain remaining monitoring points, wherein each of the remaining monitoring points is connected to at least one load device;

[0010] Acquire historical operating parameters of all the monitoring points in the electric power supply system, and construct a power supply scenario set according to the historical operating parameters of all the monitoring points, wherein the power supply scenario set includes at least one power supply scenario;

[0011] For each of the remaining monitoring points in each of the power supply scenarios, the operating currents of all the load devices connected to the remaining monitoring points within a preset time period are obtained, and according to the operating currents, a target characteristic curve matching the remaining monitoring points is determined in the database;

[0012] In a case where the target characteristic curve is the load characteristic curve, determining the load requirement of each of the remaining monitoring points in each of the power supply scenarios according to the load characteristic curve in each of the power supply scenarios;

[0013] The load demand of each of the remaining monitoring points in each of the power supply scenarios is transmitted to the edge node, and the edge node is used to perform load optimization calculation according to the load demand, obtain the best load scheduling result that maximizes the energy utilization efficiency of the power supply system in each of the power supply scenarios, and upload the best load scheduling result in each of the power supply scenarios to the cloud control platform;

[0014] Determining a current power supply scenario in the power supply scenario set according to the real-time operating parameters;

[0015] The optimal load scheduling result under the current power supply scenario is used as the target load scheduling result, and the loads of the remaining monitoring points are scheduled according to the target load scheduling result.

[0016] In a possible implementation manner of the first aspect, constructing a power supply scenario set according to the historical operating parameters of all the monitoring points includes:

[0017] Performing feature extraction on the historical operating parameters of all the monitoring points to obtain at least one target feature for describing a power supply scenario;

[0018] Clustering all the target features to obtain at least one clustering result, wherein each clustering result is used to characterize one of the power supply scenarios, and a set of all the clustering results is the power supply scenario set;

[0019] Each of the power supply scenarios includes power supply characteristics, power supply time, load level and total available power of the power supply system.

[0020] In another possible implementation manner of the first aspect, determining, according to the load characteristic curve in each power supply scenario, the load requirement of each of the remaining monitoring points in each power supply scenario includes:

[0021] Determine the operating efficiency and output power of each of the remaining monitoring points in each of the power supply scenarios according to the load characteristic curve in each of the power supply scenarios;

[0022] For each of the remaining monitoring points, the input power of the remaining monitoring points in each of the power supply scenarios is calculated based on the operating efficiency and the output power, and the input power is used to characterize the load demand of the remaining monitoring points in each of the power supply scenarios, wherein the total value of the load demand of all the remaining monitoring points in each of the power supply scenarios is less than the total available power.

[0023] In another possible implementation manner of the first aspect, the edge node performs load optimization calculation according to the load demand to obtain an optimal load scheduling result that minimizes the energy utilization efficiency of the electric power supply system in each power supply scenario, including:

[0024] In response to receiving the transmission instruction of the load demand, the edge node obtains the voltage, current, rotation speed of each of the remaining monitoring points and the flow and wind pressure of each load device connected to the remaining monitoring points through the cloud control platform, and the transmission instruction is used to transmit the load demand of each of the remaining monitoring points in each of the power supply scenarios to the edge node;

[0025] According to the voltage, current, rotation speed of each of the remaining monitoring points and the flow rate and wind pressure of each load device connected to the remaining monitoring points, a load optimization model is constructed, wherein the optimization goal of the load optimization model is to maximize energy utilization efficiency, and the constraint condition is that the total value of the load demand of all the remaining monitoring points in each power supply scenario is less than the total available power;

[0026] Using a preset optimization algorithm to solve the load optimization model and obtain at least one load scheduling result;

[0027] According to the load scheduling result, calculating the actual working efficiency of each of the remaining monitoring points in each of the power supply scenarios;

[0028] The load scheduling result with the highest actual working efficiency is taken as the optimal load scheduling result.

[0029] In another possible implementation manner of the first aspect, after taking the load scheduling result with the highest actual working efficiency as the optimal load scheduling result, the method further includes:

[0030] According to the operating efficiency and the actual working efficiency of each of the remaining monitoring points in each of the power supply scenarios, the power saving rate in each of the power supply scenarios is calculated, and the power saving rate is uploaded to the cloud control platform.

[0031] In another possible implementation manner of the first aspect, determining a current power supply scenario in the power supply scenario set according to the real-time operating parameter includes:

[0032] Extracting features of the real-time operating parameters to obtain real-time feature vectors;

[0033] Performing time series decomposition on the real-time feature vector to obtain residual features;

[0034] Extracting random fluctuation features and mutation features from the residual features;

[0035] A current power supply scenario is determined in the power supply scenario set according to the random fluctuation feature and the mutation feature.

[0036] In another possible implementation manner of the first aspect, determining a current power supply scenario in the power supply scenario set according to the random fluctuation feature and the mutation feature includes:

[0037] Acquire historical random fluctuation characteristics and historical mutation characteristics of each of the power supply scenarios in the power supply scenario set;

[0038] Calculating a first similarity between the random fluctuation feature and the historical random fluctuation feature;

[0039] Calculating a second similarity between the mutation feature and the historical mutation feature;

[0040] Calculating a comprehensive similarity based on the first similarity and the second similarity;

[0041] The power supply scenario corresponding to the largest comprehensive similarity is used as the current power supply scenario.

[0042] In another possible implementation manner of the first aspect, after inputting the real-time operating parameters into a pre-built fault prediction model to obtain a fault monitoring point, a fault type, and a fault time, the method further includes:

[0043] Constructing a digital twin model corresponding to the electric power supply system, wherein the digital twin model includes the permanent magnet synchronous motor and the three-phase asynchronous motor consistent with the electric power supply system;

[0044] Simulating the fault monitoring point, the fault type and the fault time in the digital twin model, and determining the impact of the fault monitoring point on other monitoring points in the digital twin model;

[0045] According to the influence of the fault monitoring point and the fault time on other monitoring points in the digital twin model, the influence scope is predicted and displayed in the digital twin model.

[0046] In another possible implementation manner of the first aspect, determining the influence of the fault monitoring point on other monitoring points in the digital twin model includes:

[0047] Injecting the fault type and the fault time into the fault monitoring point in the digital twin model, and running the digital twin model to obtain operation data of each of the monitoring points when the digital twin model is running;

[0048] Determining whether the operating data of each of the monitoring points is abnormal;

[0049] In the case where the operation data of the monitoring point is abnormal, a fault monitoring point closest to the monitoring point is determined, and the monitoring point is determined to be affected by the fault monitoring point closest to the monitoring point.

[0050] In a second aspect, the present application provides an electronic device, including:

[0051] a memory configured to store instructions; and

[0052] The processor is configured to call the instructions from the memory and to implement the above-mentioned online intelligent cloud control method for motor energy saving and consumption reduction management when executing the instructions.

[0053] In a third aspect, the present application provides a cloud control platform, including a database, wherein the database stores a no-load characteristic curve and a load characteristic curve of a monitoring point, wherein the monitoring point includes a permanent magnet synchronous motor and a three-phase asynchronous motor, and the cloud control platform further includes:

[0054] An acquisition module, used to acquire real-time operating parameters of all the monitoring points in the electric power supply system;

[0055] A remaining monitoring point determination module is used to input the real-time operating parameters into a pre-built fault prediction model to obtain a fault monitoring point, a fault type and a fault time, and to filter out the fault monitoring point from all the monitoring points to obtain remaining monitoring points, wherein each of the remaining monitoring points is connected to at least one load device;

[0056] A power supply scenario set construction module, used to obtain historical operating parameters of all the monitoring points in the electric power supply system, and construct a power supply scenario set according to the historical operating parameters of all the monitoring points, wherein the power supply scenario set includes at least one power supply scenario;

[0057] a target characteristic curve determination module, configured to obtain, for each of the remaining monitoring points in each of the power supply scenarios, the operating currents of all the load devices connected to the remaining monitoring points within a preset time period, and determine, in the database, a target characteristic curve matching the remaining monitoring points based on the operating currents;

[0058] A load demand determination module, configured to determine the load demand of each of the remaining monitoring points in each of the power supply scenarios according to the load characteristic curve in each of the power supply scenarios when the target characteristic curve is the load characteristic curve;

[0059] An optimal load scheduling result determination module is used to transmit the load demand of each of the remaining monitoring points in each of the power supply scenarios to the edge node, and the edge node is used to perform load optimization calculation according to the load demand to obtain the optimal load scheduling result that maximizes the energy utilization efficiency of the power supply system in each of the power supply scenarios, and upload the optimal load scheduling result in each of the power supply scenarios to the cloud control platform;

[0060] A current power supply scenario determination module, configured to determine a current power supply scenario in the power supply scenario set according to the real-time operating parameters;

[0061] The load scheduling module is used to use the optimal load scheduling result in the current power supply scenario as the target load scheduling result, and schedule the loads of the remaining monitoring points according to the target load scheduling result.

[0062] Through the above technical scheme, by obtaining the real-time operating parameters of all monitoring points in the power supply system and inputting the real-time operating parameters into the pre-built fault prediction model, the fault monitoring points can be identified and screened in advance to effectively avoid energy waste and unnecessary high-power operation caused by faults; the load demand is transmitted to the edge node for load optimization calculation to obtain the best load scheduling result under each power supply scenario, which can ensure that the power supply system can operate with the highest energy utilization efficiency under different power supply scenarios; the current power supply scenario is determined according to the real-time operating parameters, and the best load scheduling result under the current power supply scenario is implemented as the target load scheduling result to adapt to the real-time load demand and avoid unnecessary high-power operation, thereby improving energy utilization efficiency.

[0063] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 A flow chart of an online intelligent cloud control method for motor energy saving and consumption reduction management provided in an embodiment of the present application;

[0065] Figure 2 A schematic diagram of the structure of a cloud control platform provided in an embodiment of the present application. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0067] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0068] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0069] Figure 1 The flowchart of an online intelligent cloud control method for motor energy saving and consumption reduction management according to an embodiment of the present application is schematically shown. Figure 1 As shown, an embodiment of the present application provides an online intelligent cloud control method for motor energy saving and consumption reduction management, which is applied to a cloud control platform. The cloud control platform includes a database, which stores no-load characteristic curves and load characteristic curves of monitoring points. The monitoring points include permanent magnet synchronous motors and three-phase asynchronous motors. The method may include the following steps.

[0070] S110, obtaining real-time operating parameters of all monitoring points in the power supply system;

[0071] S120, inputting the real-time operating parameters into a pre-built fault prediction model to obtain a fault monitoring point, a fault type, and a fault time, and filtering out the fault monitoring point from all monitoring points to obtain remaining monitoring points, wherein each remaining monitoring point is connected to at least one load device;

[0072] S130, obtaining historical operating parameters of all monitoring points in the electric power supply system, and constructing a power supply scenario set according to the historical operating parameters of all monitoring points, the power supply scenario set including at least one power supply scenario;

[0073] S140, for each remaining monitoring point in each power supply scenario, obtaining the operating current of all load devices connected to the remaining monitoring point within a preset time period, and determining a target characteristic curve matching the remaining monitoring point in a database according to the operating current;

[0074] S150, when the target characteristic curve is a load characteristic curve, determining the load requirement of each remaining monitoring point in each power supply scenario according to the load characteristic curve in each power supply scenario;

[0075] S160, transmitting the load demand of each remaining monitoring point in each power supply scenario to the edge node, the edge node is used to perform load optimization calculation according to the load demand, obtain the best load scheduling result that maximizes the energy utilization efficiency of the power supply system in each power supply scenario, and upload the best load scheduling result in each power supply scenario to the cloud control platform;

[0076] S170, determining a current power supply scenario in a power supply scenario set according to real-time operating parameters;

[0077] S180: Taking the optimal load scheduling result under the current power supply scenario as the target load scheduling result, and scheduling the loads of the remaining monitoring points according to the target load scheduling result.

[0078] In this embodiment, the power supply system includes multiple monitoring points, which are permanent magnet synchronous motors and three-phase asynchronous motors. Real-time operating parameters of permanent magnet synchronous motors and three-phase asynchronous motors (hereinafter referred to as motors) can be obtained through sensors, and the real-time operating parameters of the motors include but are not limited to voltage, current, speed, temperature, vibration, power factor, etc.

[0079] The sensor can transmit the collected real-time operating parameters to the cloud control platform in real time. After receiving the real-time operating parameters, the cloud control platform first performs data cleaning and preprocessing to remove noise and outliers.

[0080] After obtaining the real-time operating parameters of all monitoring points, the real-time operating parameters are input into the pre-built fault prediction model. The fault prediction model is built based on a machine learning algorithm, which is trained with historical operating parameters to identify and predict possible equipment failures.

[0081] Specifically, the fault prediction model uses a deep learning network, which can predict possible fault monitoring points, fault types, and fault times based on the input real-time operating parameters. For example, if the temperature and vibration data of a motor suddenly increase abnormally, the fault prediction model predicts that the motor is at risk of overheating or mechanical failure. In this embodiment, after the fault monitoring points are identified, the fault monitoring points are screened out to obtain the remaining normal operating monitoring points, i.e., the remaining monitoring points.

[0082] Obtain the historical operating parameters of all monitoring points in the power supply system. Historical operating parameters usually include operating records of the past few months or even years, covering operating conditions in different seasons, different loads and different working conditions, such as voltage, current, power factor, load rate and other parameters of the monitoring point. In addition to historical operating parameters, environmental parameters corresponding to historical operating parameters, such as temperature, humidity, etc., can also be included. By analyzing the historical operating parameters, the typical operating mode of the power supply system under different conditions can be identified. According to the typical operating mode, a power supply scenario set is constructed. Each power supply scenario represents a specific operating state, such as a weekday peak scenario, a summer high load scenario, and a winter low load scenario. Specifically, data mining techniques such as cluster analysis can be used to construct a power supply scenario set to classify similar operating modes into the same scenario.

[0083] In this embodiment, the K-means algorithm is used, and the number of clusters K is set to 5. The feature vector includes the average load rate L, the average power factor PF, and the average ambient temperature T. Each data point can be expressed as (L, PF, T). The algorithm minimizes the objective function through iterative optimization:

[0084] J=∑(||x i -μ k || 2 );

[0085] where x i is the data point, μ k is the cluster center. The clustering results are different power supply scenarios. By constructing power supply scenarios, we can capture the periodic patterns and characteristics of power system operation, providing an important reference for subsequent load optimization. By analyzing the power consumption characteristics in different scenarios, we can formulate more targeted optimization strategies. For example, in the high-load power supply scenario in summer, we need to pay more attention to load balancing and peak reduction; while in the low-load scenario in winter, we should pay more attention to how to improve the operating efficiency of equipment.

[0086] For each remaining monitoring point in each power supply scenario, the operating current of all load devices connected to the remaining monitoring point within a preset time period is obtained, and the target characteristic curve matching the remaining monitoring point is determined in the database based on the operating current of all load devices. The preset time period may be 15 minutes or 30 minutes, depending on the response requirements of the power supply system. The acquisition of operating current is achieved through smart meters or current sensors, and data is collected at fixed intervals (such as every minute).

[0087] For each load device, calculate its average current, maximum current and minimum current within a preset time period. Then, compare the above current data with the no-load characteristic curve and load characteristic curve stored in the database. The comparison process uses a curve fitting algorithm, such as the least squares method. Assume that the characteristic curve equation stored in the database is y=ax 2 +bx+c, where y represents the current and x represents the load rate. Substitute the actual measured current data into the equation and calculate the sum of squares of the error:

[0088]

[0089] By minimizing E, the best matching characteristic curve parameters a, b and c can be determined. If the fitting error is less than a preset threshold, it is considered that a matching characteristic curve has been found.

[0090] The above process is performed for each monitoring point separately, and finally the target characteristic curve of each monitoring point in the current power supply scenario is determined, so as to accurately reflect the characteristics of different load devices in actual operation and provide a more accurate basis for subsequent load optimization. By matching the most suitable characteristic curve, the energy consumption under different load conditions can be more accurately predicted, and more effective energy-saving strategies can be formulated. For example, for a three-phase asynchronous motor, if the load characteristic curve is matched, the energy consumption under different load rates can be more accurately calculated in the subsequent optimization, so as to find the best operating point.

[0091] When the target characteristic curve is a load characteristic curve, the load requirement of each remaining monitoring point in each power supply scenario is determined according to the load characteristic curve in each power supply scenario. The load characteristic curve is usually expressed as a function of the relationship between current or power and load rate, such as P=f(L), where P is power and L is load rate.

[0092] First, it is necessary to determine the expected load rate range under the current power supply scenario. This can be obtained by analyzing the load pattern of the power supply scenario in the historical operating parameters, for example, using time series analysis methods such as ARIMA models for prediction. Assume that the predicted load rate range is [Lmin, Lmax]. Then, sample the load rate with a fixed step size (such as 1%) within this range, and calculate the corresponding power value for each sampling point. The specific calculation process is as follows: For the load rate Li, substitute the load characteristic curve equation to calculate the power Pi = f (Li); consider the motor efficiency η (Li), calculate the input power Pin = Pi / η (Li); consider the power factor cosφ (Li), calculate the apparent power Si = Pin / cosφ (Li); calculate the current Ii = Si / U, where U is the rated voltage.

[0093] Through the above calculation, a set of (Li, Ii) point pairs can be obtained. These point pairs constitute the load demand curve of the monitoring point in the current power supply scenario. Because it takes into account the characteristics of motor efficiency and power factor changing with load rate, it more accurately reflects the actual load demand. For example, for a motor with a rated power of 100kW, when the load rate is 80%, if its efficiency is 0.92 and the power factor is 0.85, then the calculated input power is about 87kW and the apparent power is about 102kVA. In this way, the optimal operating load rate range can be identified, which can not only meet production needs but also achieve higher energy efficiency.

[0094] After the load demand is determined, the data is transmitted to the edge node. The edge node is a computing device deployed close to the data source. After the edge node receives the load demand data, it performs load optimization calculations. The goal of the load optimization calculation is to find the best load scheduling result for each power supply scenario to maximize the energy utilization efficiency of the power supply system. In this embodiment, an optimization algorithm can be used to comprehensively determine the load demand, power supply capacity and system constraints of each monitoring point to determine the optimal load scheduling solution. Finally, the optimal load scheduling result for each power supply scenario is uploaded to the cloud control platform to realize intelligent scheduling and optimization of the load and improve energy utilization efficiency.

[0095] During the actual operation, the real-time monitoring parameters are continuously updated. Based on the real-time monitoring parameters, the current power supply scenario can be determined in the power supply scenario set. Specifically, in a possible embodiment, the current power supply scenario can be matched by comparing the real-time monitoring parameters with the typical operation mode in the power supply scenario set. For example, if the current load and ambient temperature match the typical mode of a high-load scenario, it can be determined that the current power supply scenario is a high-load scenario.

[0096] After determining the current power supply scenario, the optimal load scheduling result under the power supply scenario is used as the target load scheduling result. According to the target load scheduling result, the loads of the remaining monitoring points are scheduled. Specifically, the operating parameters of the motor, such as speed and voltage, can be adjusted to meet the target load scheduling result. For example, in a low-load scenario, the speed and voltage of the motor can be reduced to reduce energy consumption. In summary, the dynamic optimization of the power supply system can be achieved, energy utilization efficiency can be improved, and operating costs can be reduced.

[0097] This embodiment obtains the real-time operating parameters of all monitoring points in the electric power supply system and inputs the real-time operating parameters into a pre-built fault prediction model, so as to identify and screen out faulty monitoring points in advance, so as to effectively avoid energy waste and unnecessary high-power operation caused by faults; transmits the load demand to the edge node for load optimization calculation to obtain the optimal load scheduling result under each power supply scenario, so as to ensure that the electric power supply system can operate with the highest energy utilization efficiency under different power supply scenarios; determines the current power supply scenario according to the real-time operating parameters, and implements the optimal load scheduling result under the current power supply scenario as the target load scheduling result to adapt to the real-time load demand and avoid unnecessary high-power operation, thereby improving energy utilization efficiency.

[0098] In one implementation of this embodiment, building a power supply scenario set according to the historical operating parameters of all monitoring points includes the following steps:

[0099] S210, extracting features from historical operating parameters of all monitoring points to obtain at least one target feature for describing a power supply scenario;

[0100] S220, clustering all target features to obtain at least one clustering result, wherein each clustering result is used to characterize a power supply scenario, and a set of all clustering results is a power supply scenario set;

[0101] Each power supply scenario includes power supply characteristics, power supply time, load level, and total available power of the power supply system.

[0102] First, the historical operating parameters of all monitoring points are cleaned and preprocessed, including removing outliers, processing missing data and standardization. Outliers can be identified by box plots or Z-score methods, and missing data can be processed by mean filling or multiple interpolation. Standardization uses the Z-score standardization method, that is, (x-μ) / σ, where x is the original value, μ is the mean, and σ is the standard deviation, to ensure that features of different dimensions can be compared.

[0103] Next, dimensionality reduction techniques such as principal component analysis (PCA) are applied to reduce the data dimension and extract the main features. The principle of principal component analysis is to project the original high-dimensional data to the direction with the largest variance and retain the most important information. The specific calculation steps include: calculating the covariance matrix, solving the eigenvalues ​​and eigenvectors, and selecting the eigenvectors corresponding to the largest k eigenvalues ​​as the principal components. For example, for a data set containing 10 original features such as voltage, current, and power factor, it can be reduced to 3-4 main features through principal component analysis.

[0104] In addition, feature selection algorithms, such as recursive feature elimination, can be applied to further filter out the most relevant features. It works by repeatedly building a model (such as a random forest), removing the least important features each time, until a preset number of features is reached. For example, you can start with 10 initial features and remove 1 feature each time, eventually retaining the 5 most important features.

[0105] Through the above steps, we can obtain the target features that describe the power supply scenario, such as average load rate, peak-to-valley difference, power factor change rate, etc. These features can effectively capture the operation mode and characteristics of the power supply system. For example, the average load rate reflects the overall power consumption level, the peak-to-valley difference reflects the power consumption fluctuation, and the power factor change rate reflects the change trend of reactive power.

[0106] In summary, the most representative and discriminative features can be extracted from massive historical operating parameters, which greatly reduces the computational complexity of subsequent analysis while retaining key information to more accurately describe and distinguish different power supply scenarios.

[0107] After feature extraction, cluster analysis is performed on the target features. Cluster analysis is used to classify operating states with similar features into the same power supply scenario. Specifically, the K-means clustering algorithm can be used. The K-means algorithm is a commonly used unsupervised learning algorithm that iteratively updates the cluster center and assigns the data points to the closest cluster center.

[0108] In the specific implementation, first determine the number of clusters K, and then randomly initialize K cluster centers. Next, assign each target feature to the nearest cluster center and update the location of the cluster center. Repeat the above steps until the cluster center no longer changes significantly, so that multiple clustering results can be obtained, each clustering result represents a power supply scenario. The collection of all clustering results is the power supply scenario set.

[0109] Specifically, in the power supply scenario, the power supply characteristics may include average voltage, current fluctuation, power factor, etc.; the power supply time refers to the time distribution of the target feature in each clustering result; the load level can be calculated from the target feature in the clustering result, representing the typical load condition of the power supply scenario; the total available power refers to the maximum power supply capacity of the power supply system under the current power supply scenario.

[0110] For example, a high-load scenario in summer could be:

[0111] Power supply characteristics (cluster center): high average voltage, large current fluctuation, and low power factor;

[0112] Power supply time: 13:00-17:00 on weekdays;

[0113] Load level: peak load (e.g. 85% of the power supply system capacity);

[0114] Total available power: The maximum power supply capability based on the current grid configuration.

[0115] This implementation method can construct a set of power supply scenarios that describe different operating states of the power supply system by extracting features and performing cluster analysis on the historical operating parameters of all monitoring points. Cluster analysis can classify similar operating states into the same power supply scenario, simplifying the management and optimization of the power supply system. Each power supply scenario includes multiple key features, which can comprehensively describe the characteristics of the power supply system under different operating states, provide basic data for subsequent load optimization, improve the energy utilization efficiency of the power supply system, and reduce operating costs.

[0116] In one implementation of this embodiment, determining the load requirement of each remaining monitoring point in each power supply scenario according to the load characteristic curve in each power supply scenario includes the following steps:

[0117] S310, determining the operating efficiency and output power of each remaining monitoring point in each power supply scenario according to the load characteristic curve in each power supply scenario;

[0118] S320. For each remaining monitoring point, calculate the input power of the remaining monitoring point in each power supply scenario based on the operating efficiency and the output power. The input power is used to characterize the load demand of the remaining monitoring point in each power supply scenario, wherein the total value of the load demand of all the remaining monitoring points in each power supply scenario is less than the total available power.

[0119] This embodiment can determine the operating efficiency and output power of each remaining monitoring point in each power supply scenario based on the load characteristic curve in each power supply scenario. The load characteristic curve is usually expressed as the relationship between the motor output power and the load rate, which can be expressed by a polynomial function or a piecewise function. For example, a typical load characteristic curve may be expressed as:

[0120] P o =a*L 3 +b*L 2 +c*L+d;

[0121] Where P o is the output power, L is the load factor, and a, b, c, and d are fitting coefficients.

[0122] First, it is necessary to determine the possible load rate range according to the characteristics of each power supply scenario (such as average load rate, peak-to-valley difference, etc.). For example, in the "summer high load" scenario, the load rate range may be 70%-100%. Then, sampling is performed in small steps (such as 1%) within this range, and the corresponding output power is calculated for each sampling point.

[0123] Next, calculate the operating efficiency. The efficiency of the motor is usually also a function of the load factor, which can be expressed as: η = f(L).

[0124] The above function can be a lookup table obtained from experiments or a fitted polynomial function.

[0125] For example, η = e*L 2 +f*L+g, where e, f, and g are fitting coefficients. For each load rate sampling point, the corresponding efficiency must be calculated.

[0126] In specific implementation, the following steps can be adopted: for each monitoring point, obtain the parameters of its load characteristic curve and efficiency curve; determine the load rate range according to the power supply scenario, such as [Lmin, Lmax]; sample in the range of [Lmin, Lmax] with a step size of 1%; for each sampling point Li:

[0127] Calculate the output power:

[0128] Computational efficiency:

[0129] Generate (L i , P oi , η i ) dataset, which describes the possible operating states of the remaining monitoring points in this power supply scenario.

[0130] For example, for a motor with a rated power of 100kW, at an 80% load rate, its output power may be 80kW and its efficiency may reach 93%. At a 50% load rate, the output power is 50kW and the efficiency may drop to 89%. Through this detailed analysis, accurate basic data can be provided for subsequent optimization decisions, helping to find the best balance between energy efficiency and production needs.

[0131] For each remaining monitoring point, the input power of the remaining monitoring point in each power supply scenario is calculated based on the operating efficiency and the output power. The input power is used to characterize the load demand of the remaining monitoring point in each power supply scenario.

[0132] First, the input power is calculated based on the output power and efficiency data obtained in step S310. The calculation formula for the input power is:

[0133] Pi = Po / η;

[0134] Where Pi is the input power, Po is the output power, and η is the efficiency. For example, if the output power of a motor at a certain load point is 80kW and the efficiency is 92%, then the input power is 80kW / 0.92≈87kW.

[0135] Then, consider the effect of power factor. Power factor is usually a function of load factor and can be expressed as:

[0136] PF=h*L 2 +i*L+j;

[0137] Where h, i, j are fitting coefficients. After considering the power factor, the apparent power can be calculated:

[0138] S=P i / PF;

[0139] Apparent power can more accurately reflect the actual load of the power grid.

[0140] Next, all remaining monitoring points in each power supply scenario are aggregated to calculate the total load demand. It is necessary to ensure that the total load demand does not exceed the total available power defined in the power supply scenario.

[0141] The specific implementation steps are as follows:

[0142] For each (L generated in S310 i , P oi , η i ) dataset:

[0143] Calculate input power: P i =P oi / η i ;

[0144] Calculate the power factor:

[0145] Calculate the apparent power: S i =P i / PF i ;

[0146] Generate (L i , Pi , S i ) dataset is used to describe the load demand characteristics of the remaining monitoring points in this power supply scenario.

[0147] For each power supply scenario, S of all monitoring points are summed up. i , and get the total load demand St.

[0148] Determine whether St is less than the total available power Pa. If St>Pa, load optimization or adjustment of the power supply strategy is required.

[0149] For example, consider a power supply scenario with 10 monitoring points, each with an apparent power demand ranging from 50kVA to 200kVA. By aggregating these demands, the total load demand may be 1500kVA. If the total available power is 2000kVA, then the load demand in this scenario can be met. However, if the total available power is only 1400kVA, then load optimization or increased power supply capacity needs to be considered.

[0150] This implementation can not only reflect the operating status of a single remaining monitoring point, but also evaluate the load of the entire power supply system. Through load demand analysis, the operating status of the power system can be better understood and predicted, providing an accurate basis for load optimization and energy efficiency management. It not only helps to improve energy efficiency, but also prevents possible power shortages and ensures the stable operation of the power system.

[0151] In one implementation of this embodiment, the edge node is used to perform load optimization calculation according to the load demand to obtain the best load scheduling result that maximizes the energy utilization efficiency of the power supply system in each power supply scenario, including the following steps:

[0152] S410, in response to receiving the transmission instruction of the load demand, the edge node obtains the voltage, current, speed of each remaining monitoring point and the flow and wind pressure of each load device connected to the remaining monitoring point through the cloud control platform, and the transmission instruction is used to transmit the load demand of each remaining monitoring point in each power supply scenario to the edge node;

[0153] S420, constructing a load optimization model based on the voltage, current, and rotation speed of each remaining monitoring point and the flow and wind pressure of each load device connected to the remaining monitoring point, wherein the optimization goal of the load optimization model is to maximize energy utilization efficiency, and the constraint condition is that the total value of the load demand of all the remaining monitoring points in each power supply scenario is less than the total available power;

[0154] S430, using a preset optimization algorithm to solve a load optimization model and obtain at least one load scheduling result;

[0155] S440. Calculate the actual working efficiency of each remaining monitoring point in each power supply scenario according to the load scheduling result;

[0156] S450: Taking the load scheduling result with the highest actual working efficiency as the best load scheduling result.

[0157] After receiving the transmission instruction of the load demand, the edge node immediately obtains the key operating parameters of each remaining monitoring point through the cloud control platform. These parameters include voltage, current, speed, and flow and wind pressure of the load equipment connected to the monitoring point. The data acquisition process uses a secure communication protocol, such as HTTPS or MQTT, to ensure the security and reliability of data transmission. After receiving the request, the cloud control platform extracts the required information from the real-time database. These data are usually collected and stored at a high frequency (such as once per second) and come from smart sensors installed on motors and load equipment. For example, voltage data can be three-phase voltage values, current is measured by current transformers, and speed can be measured using Hall sensors or encoders. Flow and wind pressure data come from flow meters and pressure sensors installed in pipes or ducts. During the data acquisition process, NTP (Network Time Protocol) is used to ensure data time synchronization between different devices. The cloud control platform packages the collected data and sends it back to the edge node. The data packet may contain detailed parameter information of multiple monitoring points. After receiving the data, the edge node stores it in the local cache to prepare for subsequent optimization calculations.

[0158] To build a load optimization model, we first need to clarify the optimization objectives and constraints. The optimization goal is to maximize energy efficiency while ensuring that the total load demand does not exceed the available power, that is, to minimize energy waste while meeting all load demands. This can be achieved by optimizing power distribution so that the power supply system can achieve the highest efficiency when working.

[0159] In order to achieve the above goals, first of all, energy utilization efficiency is defined as the ratio of useful power output to input power. For multiple monitoring points, the overall efficiency of the power supply system can be expressed as:

[0160] η=∑(P oi ) / ∑(P i ), where P oi and P i are the output and input power of the ith monitoring point respectively. The objective function is set to maximize η. The constraints include:

[0161] (1) Load rate limit: 0≤Li≤1; (2) Total load demand limit: ∑(P i )≤Pa; (3) Flow and wind pressure limits: Fmin≤Fi≤Fmax, Pmin≤Pi≤Pmax. In addition, consider the relationship between the motor and the load device, such as the fan law:

[0162] P∝n 3 , Q∝n, H∝n 2 ;

[0163] The power and efficiency functions in the load optimization model can be obtained by polynomial fitting, for example:

[0164] Poi=ai*Li 3 +bi*Li 2 +ci*Li+di;

[0165] In summary, the load optimization model takes into account motor efficiency, load characteristics and system constraints, and can find the maximum overall energy utilization efficiency while satisfying all constraints.

[0166] After the load optimization model is constructed, the preset optimization algorithm can be used to solve the load optimization model to obtain the best load scheduling result. Considering the nonlinear characteristics and multiple constraints of the model, advanced global optimization algorithms such as genetic algorithms are usually selected.

[0167] The steps of using genetic algorithm include: (1) randomly generating a set of feasible load scheduling schemes; (2) evaluating the fitness of each load scheduling scheme, that is, the energy utilization efficiency; (3) selecting excellent individuals to enter the next generation; (4) performing crossover operations to generate new offspring; (5) performing mutations with a small probability to increase population diversity; (6) processing constraints to ensure that the newly generated individuals meet all constraints; (7) repeating steps 2-6 until the termination condition is reached.

[0168] According to the load scheduling results, the actual working efficiency of each remaining monitoring point in each power supply scenario is calculated to evaluate the actual effect of the load scheduling scheme obtained by the optimization algorithm. The calculation process is as follows: (1) Extract the preset load rate Lf of each monitoring point; (2) Calculate the input power: P i =P*L f , where P is the rated power of the motor; (3) Considering the motor efficiency curve, calculate the output power: P o =P i *η(L f ), where η(L f ) represents the motor efficiency under load rate Lf; (4) Calculate the efficiency of the load equipment (water pump): η = f(Q, H), where Q is the flow rate and H is the pressure head; (5) Calculate the overall efficiency of the power supply system: ηs = η(L f )*η; (6) Calculate energy efficiency: P oa =P o *η,η e =P oa / P i .

[0169] The load scheduling result with the highest actual working efficiency is selected as the optimal load scheduling result. The specific implementation process is as follows: (1) Summarize the actual working efficiency of all remaining monitoring points under each power supply scenario calculated in step S440; (2) For each load scheduling scheme, calculate its average actual working efficiency under each power supply scenario; (3) Compare the average actual working efficiency of all schemes and select the scheme with the highest efficiency as the optimal load scheduling result.

[0170] In practical applications, an efficiency improvement threshold can be set. Only when the efficiency improvement of the new load scheduling scheme exceeds this threshold compared with the current operating scheme, will we consider switching to the new scheme to avoid system instability caused by frequent adjustments.

[0171] This implementation can significantly improve the energy efficiency of the power supply system, reduce energy waste, and reduce operating costs. In addition, optimized load scheduling can also improve the stability and reliability of the power supply system, avoid failures and power outages caused by overload, thereby improving power supply quality and user satisfaction.

[0172] In one implementation of this embodiment, after taking the load scheduling result with the highest actual working efficiency as the optimal load scheduling result, the following steps are further included:

[0173] S510. Calculate the power saving rate in each power supply scenario according to the operating efficiency and actual working efficiency of each remaining monitoring point in each power supply scenario, and upload the power saving rate to the cloud control platform.

[0174] After determining the optimal load scheduling result, the power saving rate in each power supply scenario is calculated based on the operating efficiency and actual working efficiency of each remaining monitoring point in each power supply scenario. Specifically, the power saving rate represents the proportion of power consumption saved after optimization relative to that before optimization. The calculation of the power saving rate needs to consider the operating efficiency (before optimization) and actual working efficiency (after optimization) of each remaining monitoring point.

[0175] The calculation process can be divided into the following sub-steps: 1. Obtain the operating efficiency and actual work efficiency data of each remaining monitoring point. Among them, the operating efficiency data comes from historical records or benchmarks, while the actual work efficiency data is calculated through the optimal load scheduling results. 2. For each remaining monitoring point, use the formula: single-point power saving rate = (1-operating efficiency / actual work efficiency) * 100%. For example, if the operating efficiency of a monitoring point is 80% and the actual work efficiency is 85%, the single-point power saving rate is (1-80% / 85%)*100%≈5.88%. 3. Calculate the scene power saving rate, scene power saving rate = ∑(single-point power saving rate i*weight i) / ∑(weight i), where the weight of each single-point power saving rate is preset. After calculating the scene power saving rate, the calculation result is rounded up and down, usually retaining two decimal places. 4. Package the scene power saving rate and related information (such as scene ID, calculation time, number of monitoring points involved in the calculation, etc.) into a specified format, such as JSON or XML, and upload the data package to the cloud control platform using a secure network protocol (such as HTTPS or MQTT). During the upload process, data encryption and identity authentication are required to ensure the security of data transmission.

[0176] For example, suppose a power supply scenario has three monitoring points, and the data are as follows: Monitoring point A: operating efficiency 82%, actual work efficiency 88%, weight 0.5; Monitoring point B: operating efficiency 75%, actual work efficiency 80%, weight 0.3; Monitoring point C: operating efficiency 70%, actual work efficiency 78%, weight 0.2; the calculation process is as follows:

[0177] The power saving rate of monitoring point A: (1-82% / 88%)*100%≈6.82%; the power saving rate of monitoring point B: (1-75% / 80%)*100%=6.25%; the power saving rate of monitoring point C: (1-70% / 78%)*100%≈10.26%; the scene power saving rate: (6.82%*0.5+6.25%*0.3+10.26%*0.2) / (0.5+0.3+0.2)≈7.34%.

[0178] According to the above calculation results, it can be known that in this power supply scenario, about 7.34% power saving effect can be achieved after optimization.

[0179] In another embodiment, the power saving rate of each remaining monitoring point can also be calculated by the parameters of the monitoring point. In this embodiment, the power saving rate includes the active power saving rate, the reactive power saving rate and the comprehensive power saving rate, wherein the active power saving rate refers to the proportion of active power (the part that actually does work) reduced by energy-saving measures to the original active power within a period of time. The calculation formula is: active power saving rate = (saved active power / original active power) × 100%; reactive power saving rate refers to the proportion of reactive power (the part that does not do actual work, used to maintain voltage and magnetic field) reduced by energy-saving measures to the original reactive power. The calculation formula is: reactive power saving rate = (saved reactive power / original reactive power) × 100%; the comprehensive power saving rate can comprehensively consider the saving of active and reactive power, and is used to measure the overall power saving effect.

[0180] Specifically, taking the monitoring point as a three-phase asynchronous motor as an example, if in one of the power supply scenarios, the energy utilization efficiency calculated by the load scheduling result still does not meet the energy saving expectations, in another embodiment, the three-phase asynchronous motor can be replaced with a permanent magnet motor to improve the energy utilization efficiency and energy saving effect, that is, the monitoring point is replaced with the target device to achieve energy saving, and the calculated power saving rate is transmitted to the cloud control platform. Specifically, the power saving rate can be calculated by the parameters of the three-phase asynchronous motor. The parameters of the three-phase asynchronous motor can be obtained through the load characteristic curve and actual testing. It is assumed that the parameters of the three-phase asynchronous motor (hereinafter referred to as the old motor) are shown in the following table:

[0181] Table 1

[0182]

[0183] Among them, the power factor of the permanent magnet motor is 0.85, and the average operating current is 470A. The calculation process of the power saving rate after replacing the permanent magnet motor is as follows:

[0184] 1.

[0185] The power consumption of the old motor = 1.732*380*470*0.85 = 263 degrees / hour;

[0186] The actual test shows that the average daily power consumption is 6300 kWh, that is, 6300 / 24=262.5 kWh per hour.

[0187] 2. Operation load rate = operation current / rated current;

[0188] The load factor of the old motor = 470 / 559 = 0.84;

[0189] 3.

[0190] Apparent power of the old motor = power consumption / power factor = 263 / 0.84 = 313.1KW;

[0191] Reactive power of old motor = sqrt(313.1*313.1-263*263) = 169.88 kVA;

[0192] At 84% load rate, the asynchronous motor's load characteristics show that the asynchronous motor's operating efficiency at this load is 90%, while the permanent magnet motor's operating efficiency at 84% load is 97%;

[0193] 4. Actual motor shaft output power = power consumption * operating efficiency;

[0194] The shaft output power of the old motor = 263*0.9 = 236.7KW;

[0195] At the same shaft output power, that is, the same load (236.7KW):

[0196] The power consumption of the permanent magnet motor = 236.7 / 0.97 = 244.02KW;

[0197] 5. Active power saving = power consumption of old motor - power consumption of permanent magnet motor;

[0198] Active power saving of permanent magnet motor = 263-244.02 = 18.98 degrees;

[0199] 6. Active power saving rate = (power consumption of old motor - power consumption of permanent magnet motor) / power consumption of old motor;

[0200] Active power saving rate after replacement = (263-244.02) / 263*100% = 7.22%;

[0201] 7.

[0202] The operating power factor of the permanent magnet motor = 0.95, that is

[0203]

[0204] Reactive power of permanent magnet motor = 244.02 / 0.95*0.3122 = 80.19 kVA;

[0205] 8. Reactive power saving = no power consumption of old motor - no power consumption of permanent magnet motor;

[0206] Reactive power saving of permanent magnet motor = 169.88-80.19 = 80.19 kVA;

[0207] 9. Reactive power saving rate = (no power consumption of old motor - no power consumption of permanent magnet motor) / no power consumption of old motor;

[0208] The reactive power saving rate of permanent magnet motor = (169.88-80.19)169.88*100% = 52.79%;

[0209] 10. Comprehensive power consumption of the motor = power consumption of the motor + reactive economic equivalent * power consumption of the motor;

[0210] Among them, the reactive economic equivalent KQ is 0.03, and the comprehensive power saving = active power saving + reactive economic equivalent * reactive power saving;

[0211] The total power consumption of the old motor = 263 + 0.03 * 169.88 = 268.10KW;

[0212] Permanent magnet motor comprehensive power consumption = 244.02 + 0.03 * 80.19 = 246.43KW;

[0213] Comprehensive power saving rate = (268.10-246.43) / 268.10*100% = 21.67 / 268.10*100% = 8.08%.

[0214] In summary, we can get the following Table 2:

[0215] Table 2

[0216]

[0217] The above content is an energy-saving analysis of a 315KW three-phase asynchronous motor. Similarly, the energy-saving analysis of 280KW, 110KW, 220KW and other motors can be performed according to the above method. The process will not be repeated. The calculated energy-saving data statistics are shown in Table 3 below:

[0218] Table 3

[0219]

[0220]

[0221] This implementation method can identify the operating conditions with the greatest optimization potential by comparing the power saving rates of different power supply scenarios. Uploading data to the cloud platform not only enables centralized management of information, but also provides the possibility for comprehensive analysis across scenarios and time, which helps to identify long-term energy-saving trends and formulate broader energy efficiency strategies.

[0222] In one implementation of this embodiment, determining the current power supply scenario in the power supply scenario set according to the real-time operating parameters includes the following steps:

[0223] S610, extracting features of real-time operating parameters to obtain real-time feature vectors;

[0224] S620, performing time series decomposition on the real-time feature vector to obtain residual features;

[0225] S630, extracting random fluctuation features and mutation features from residual features;

[0226] S640: Determine a current power supply scenario in a power supply scenario set according to the random fluctuation characteristics and the mutation characteristics.

[0227] Perform feature extraction on the real-time operating parameters to obtain real-time feature vectors. Feature extraction refers to converting complex raw data into more representative features. In specific implementation, some key indicators such as voltage, current, frequency, load, etc. can be extracted from the real-time operating parameters.

[0228] For example, assuming that the monitored real-time operating parameters include voltage V, current I, frequency f and load L, a real-time feature vector X=[V, I, f, L] can be obtained by extracting features of the real-time operating parameters.

[0229] This embodiment requires that the real-time feature vector be normalized to eliminate the dimension differences between different parameters so that each parameter can be compared on the same scale. The following formula can be used for normalization:

[0230]

[0231] Among them, X is the original parameter value, μ is the mean value of the parameter, and σ is the standard deviation of the parameter. Through standardization, the real-time feature vector obtained can more accurately reflect the current operating status.

[0232] Next, the real-time feature vector is decomposed into time series to obtain the residual feature. Time series decomposition is to decompose the time series data into different components in order to better analyze and understand the changing patterns of the data.

[0233] Wavelet transform can be used to decompose the time series of real-time feature vectors. Specifically, wavelet transform can decompose the time series into components of different frequencies, thereby extracting the trend and periodicity components in the data. In specific implementation, the real-time feature vector can be subjected to wavelet transform to obtain components of different frequencies, and residual features can be extracted from them. Assuming that the time series of the real-time feature vector X is represented as X(t), through wavelet transform, the high-frequency component X can be obtained. H( t ) and low frequency component X L ( t ) , where the residual feature can be expressed as:

[0234] X R (t) = X(t) - X L (t);

[0235] After obtaining the residual features, it is necessary to extract random fluctuation features and mutation features from the residual features. Random fluctuation features and mutation features are important indicators that reflect data changes and can help identify changes in the current power supply scenario. In specific implementation, the mean and variance of the residual features can be calculated to describe the random fluctuation features, and the mutation detection algorithm can be used to identify the mutation features.

[0236] Assume that the residual feature X R(t) The mean value is μ R , the variance is The random fluctuation characteristics can be expressed as:

[0237]

[0238] For mutation features, the cumulative sum algorithm can be used to detect mutation points by calculating the cumulative sum. The specific formula is:

[0239]

[0240] When C(t) exceeds a certain threshold, it can be considered that a mutation has occurred. In this way, random fluctuations and mutation features in the data can be accurately extracted.

[0241] Finally, the current power supply scenario is determined in the power supply scenario set based on the random fluctuation characteristics and mutation characteristics. The power supply scenario set is a predefined set of typical power supply states, each state corresponding to a specific set of feature values. The current power supply scenario can be determined by matching the extracted random fluctuation characteristics and mutation characteristics with the feature values ​​in the power supply scenario set.

[0242] This implementation can realize accurate identification of the current power supply scenario. It can not only monitor the operating status of the power supply system in real time, but also promptly detect abnormal situations and take corresponding measures to ensure the safe and stable operation of the power supply system, thereby significantly improving the operating efficiency of the power supply system, reducing the failure rate, and improving the quality of power supply services.

[0243] In one implementation of this embodiment, determining the current power supply scenario in the power supply scenario set according to the random fluctuation characteristics and the mutation characteristics includes the following steps:

[0244] S710, obtaining historical random fluctuation characteristics and historical mutation characteristics of each power supply scenario in the power supply scenario set;

[0245] S720, calculating a first similarity between the random fluctuation feature and the historical random fluctuation feature;

[0246] S730, calculating a second similarity between the mutation feature and the historical mutation feature;

[0247] S740, calculating a comprehensive similarity according to the first similarity and the second similarity;

[0248] S750: The power supply scenario corresponding to the greatest comprehensive similarity is used as the current power supply scenario.

[0249] The power supply scenario set is a collection of multiple typical power supply states, each of which corresponds to a specific set of historical feature values. The historical random fluctuation features and historical mutation features are obtained based on historical data and are used to represent the performance of each power supply scenario under different circumstances.

[0250] In specific implementation, historical random fluctuation characteristics and historical mutation characteristics can be extracted by statistical analysis and pattern recognition of historical data. For example, assuming that the power supply scenario set contains three typical scenarios: normal power supply, light load and heavy load, the historical data of each power supply scenario includes parameters such as voltage, current, frequency and load. By performing statistical analysis on historical data, the historical random fluctuation characteristics and mutation characteristics of each scenario can be calculated. Assuming that the historical voltage data of a power supply scenario is V = [V1, V2, ..., Vn], the mean value and variance can be calculated as historical random fluctuation characteristics, and the mutation point can be identified by the mutation detection algorithm as a historical mutation feature.

[0251] The first similarity between the random fluctuation feature and the historical random fluctuation feature is calculated. The first similarity calculation is used to measure the similarity between the current feature and the historical feature, so as to determine the current power supply scenario.

[0252] In specific implementation, the Euclidean distance can be used to calculate the similarity. Taking the Euclidean distance as an example, assuming that the current random fluctuation feature is R and the historical random fluctuation feature of a power supply scenario is R h , then the first similarity can be expressed as:

[0253]

[0254] Among them, ||RR h || represents the Euclidean distance.

[0255] Then, the second similarity between the mutation feature and the historical mutation feature is calculated. The mutation feature is an important indicator that reflects the mutation point in the data. By comparing the similarity between the current mutation feature and the historical mutation feature, the current power supply scenario can be more accurately identified. In specific implementation, a mutation point matching algorithm can be used to calculate the similarity. Assume that the current mutation feature is C and the historical mutation feature of a power supply scenario is C. h , then the second similarity can be expressed as:

[0256]

[0257] In summary, the second similarity between the current mutation feature and each historical mutation feature can be calculated.

[0258] Next, the comprehensive similarity is calculated based on the first similarity and the second similarity. The comprehensive similarity is a comprehensive similarity index obtained by combining the similarity of random fluctuation characteristics and mutation characteristics, so as to more accurately identify the current power supply scenario. In specific implementation, the weighted average method can be used to calculate the comprehensive similarity. The weight coefficient is preset to balance the influence of random fluctuation characteristics and mutation characteristics.

[0259] Finally, the power supply scenario corresponding to the largest comprehensive similarity is taken as the current power supply scenario. The greater the comprehensive similarity, the more similar the current feature is to the historical feature of the power supply scenario, so the power supply scenario can be considered to be the current power supply scenario.

[0260] In specific implementation, the comprehensive similarities of all power supply scenarios can be compared, and the scenario with the greatest similarity can be selected as the current power supply scenario.

[0261] This implementation method can more comprehensively reflect the operating status of the power supply system by comprehensively utilizing random fluctuation characteristics and mutation characteristics, thereby achieving more accurate power supply scenario identification.

[0262] In one implementation of this embodiment, after the real-time operating parameters are input into the pre-built fault prediction model to obtain the fault monitoring point, fault type and fault time, the following steps are included:

[0263] S810. Build a digital twin model corresponding to the electric power supply system, where the digital twin model includes a permanent magnet synchronous motor and a three-phase asynchronous motor consistent with the electric power supply system;

[0264] S820, simulating the fault monitoring point, fault type and fault time in the digital twin model, and determining the impact of the fault monitoring point on other monitoring points in the digital twin model;

[0265] S830. According to the influence of the fault monitoring point and the fault time on other monitoring points in the digital twin model, predict the influence scope and display it in the digital twin model.

[0266] After inputting the real-time operating parameters into the pre-built fault prediction model and obtaining the fault monitoring point, fault type and fault time, it is necessary to first build a digital twin model corresponding to the power supply system. The digital twin model is a virtual model that accurately reflects the structure and operating status of the actual power supply system. Among them, the three-phase asynchronous motor is used to adjust the operating frequency and voltage of the motor, so as to achieve precise control of the motor speed and torque. When building a digital twin model, the physical characteristics, operating parameters and control logic of components such as permanent magnet synchronous motors and three-phase asynchronous motors need to be accurately mapped to the virtual model. For example, the model of the permanent magnet synchronous motor can use mathematical equations to describe its electromagnetic and mechanical characteristics, while the model of the three-phase asynchronous motor can be built based on its control algorithm and power electronic characteristics.

[0267] The fault monitoring point, fault type and fault time are simulated in the digital twin model, and the impact of the fault monitoring point on other monitoring points in the digital twin model is determined. The fault monitoring point refers to the permanent magnet synchronous motor or three-phase asynchronous motor that may fail in the power supply system. The fault type can include different types of faults such as short circuit, open circuit, overheating, etc. The fault time refers to the specific time point when the fault occurs.

[0268] When simulating faults in a digital twin model, it is necessary to set the corresponding fault conditions according to the fault monitoring point, fault type, and fault time provided by the fault prediction model. For example, assuming that the fault prediction model predicts that the stator winding of a permanent magnet synchronous motor will have a short circuit fault at a certain moment, the short circuit condition of the stator winding can be set in the digital twin model and its impact on the motor operation state can be observed. In this way, the occurrence process of the fault can be simulated in the digital twin model, and the impact of the fault on other monitoring points can be analyzed. For example, a short circuit fault in the stator winding may cause a sharp increase in the motor current, thereby affecting the output voltage and current of the three-phase asynchronous motor. By simulating the occurrence process of the fault, the impact of the fault on the power supply system can be fully understood, providing a basis for fault diagnosis and processing.

[0269] Then, according to the impact of the fault monitoring point and fault time on other monitoring points in the digital twin model, the impact range is predicted and displayed in the digital twin model. The impact range refers to the monitoring points and range affected by the fault after the fault occurs. In order to predict the impact range, it is necessary to analyze the conduction effect of the fault on various parts of the power supply system. For example, assuming that a short circuit fault occurs in the stator winding of a permanent magnet synchronous motor, the motor current may increase sharply, thereby affecting the output voltage and current of the three-phase asynchronous motor. These changes may be further transmitted to other parts of the power supply system, such as the power grid, cables, and loads. By simulating the occurrence process of the fault in the digital twin model, the impact range of the fault on the power supply system can be fully understood. In specific implementation, numerical simulation and visualization technology can be used to display the impact range of the fault in the digital twin model. For example, through color coding and animation effects, the degree and time process of the fault on each monitoring point can be intuitively displayed. In this way, the impact range of the fault can be fully understood, providing a basis for fault diagnosis and processing.

[0270] The construction and application of the digital twin model in this implementation can not only monitor the operating status of the power supply system in real time, but also timely predict and display the impact range of the fault when a fault occurs, providing a reliable basis for fault diagnosis and processing. This can significantly improve the operating efficiency of the power supply system, reduce the failure rate, and improve the quality of power supply services. By comprehensively utilizing the digital twin model and the fault prediction model, the operating status of the power supply system can be more comprehensively reflected, thereby achieving more accurate fault diagnosis and processing.

[0271] In one implementation of this embodiment, determining the impact of a fault monitoring point on other monitoring points in a digital twin model includes the following steps:

[0272] S910, injecting the fault type and fault time into the fault monitoring point in the digital twin model, and running the digital twin model to obtain the operation data of each monitoring point when the digital twin model is running;

[0273] S920, determining whether the operating data of each monitoring point is abnormal;

[0274] S930: When the operation data of the monitoring point is abnormal, determine the fault monitoring point closest to the monitoring point, and determine whether the monitoring point is affected by the fault monitoring point closest to the monitoring point.

[0275] Inject the fault type and fault time into the fault monitoring point in the digital twin model, run the digital twin model, and obtain the operation data of each monitoring point when the digital twin model is running. First of all, the fault monitoring point refers to the motor or three-phase asynchronous motor that may fail in the system. The fault type can be different types of faults such as short circuit, open circuit, overheating, etc., and the fault time refers to the specific time point when the fault occurs. When injecting faults into the digital twin model, it is necessary to set the corresponding fault conditions according to the fault monitoring point, fault type and fault time provided by the fault prediction model.

[0276] For example, if the fault prediction model predicts that the stator winding of a permanent magnet synchronous motor will have a short circuit fault at a certain moment, the short circuit condition of the stator winding can be set in the digital twin model, and the model can be run to observe its impact on the system operation status. The digital twin model simulates the operation process of the actual system and generates operation data for each monitoring point, such as current, voltage, temperature and other parameters.

[0277] In order to determine whether the operating data of the monitoring point is abnormal, it is necessary to set the benchmark data under normal operating conditions. For example, assuming that the normal operating current range of a motor is 10A to 20A, when the actual operating current exceeds this range, it can be judged as abnormal. In addition, machine learning algorithms, such as anomaly detection algorithms, can be used to analyze the operating data in real time to determine whether there are abnormal conditions.

[0278] In the case of abnormal operation data of the monitoring point, determine the fault monitoring point closest to the monitoring point, and determine whether the monitoring point is affected by the fault monitoring point closest to the monitoring point. When the operation data of a monitoring point is found to be abnormal, it is necessary to further analyze its relationship with the fault monitoring point. In specific implementation, distance measurement methods such as Euclidean distance, Manhattan distance, etc. can be used to calculate the distance between the monitoring point and the fault monitoring point. For example, assuming that a short circuit fault occurs in the stator winding of a motor, causing a sharp increase in the motor current, the distance between the motor and the fault monitoring point can be calculated to determine whether it is affected by the short circuit fault. In addition, the causal analysis method can be used to analyze the causal relationship between the monitoring point and the fault monitoring point and determine the fault conduction path. For example, the scope of the fault's impact on other parts of the system can be determined by analyzing the changing trends of parameters such as current and voltage.

[0279] This implementation method can realize a comprehensive analysis of the impact of the fault monitoring point on other monitoring points in the digital twin model. First, by injecting fault conditions into the digital twin model and running the model, the operating data of each monitoring point is obtained; then, by setting the benchmark data under normal operating conditions, it is determined whether the operating data of the monitoring point is abnormal; finally, by calculating the distance and causal relationship between the monitoring point and the fault monitoring point, the impact range of the fault on other parts of the system is determined. Not only can the system operating status be monitored in real time and abnormal conditions be discovered in a timely manner, but the impact of the fault on the system can also be comprehensively analyzed, providing a reliable basis for fault diagnosis and processing. Finally, by comprehensively utilizing the digital twin model and the fault prediction model, the operating efficiency of the system can be significantly improved, the failure rate can be reduced, and the reliability and safety of the system can be improved.

[0280] The present application also provides an electronic device, including:

[0281] a memory configured to store instructions; and

[0282] The processor is configured to call instructions from the memory and implement the above-mentioned online intelligent cloud control method for motor energy saving and consumption reduction management when executing the instructions.

[0283] Reference Figure 2 The embodiment of the present application further provides a cloud control platform, including a database, the database storing no-load characteristic curves and load characteristic curves of monitoring points, the monitoring points including permanent magnet synchronous motors and three-phase asynchronous motors, and the cloud control platform further comprising:

[0284] An acquisition module 10 is used to acquire real-time operating parameters of all monitoring points in the power supply system;

[0285] The remaining monitoring point determination module 20 is used to input the real-time operating parameters into the pre-built fault prediction model to obtain the fault monitoring point, fault type and fault time, and screen out the fault monitoring point from all the monitoring points to obtain the remaining monitoring points, wherein each remaining monitoring point is connected to at least one load device;

[0286] A power supply scenario set building module 30, used to obtain historical operating parameters of all monitoring points in the power supply system, and build a power supply scenario set according to the historical operating parameters of all monitoring points, the power supply scenario set including at least one power supply scenario;

[0287] The target characteristic curve determination module 40 is used to obtain the operating current of all load devices connected to the remaining monitoring points within a preset time period for each remaining monitoring point in each power supply scenario, and determine the target characteristic curve matching the remaining monitoring points in the database according to the operating current;

[0288] A load demand determination module 50, for determining the load demand of each remaining monitoring point in each power supply scenario according to the load characteristic curve in each power supply scenario when the target characteristic curve is a load characteristic curve;

[0289] The optimal load scheduling result determination module 60 is used to transmit the load demand of each remaining monitoring point in each power supply scenario to the edge node, and the edge node is used to perform load optimization calculation according to the load demand, obtain the optimal load scheduling result that maximizes the energy utilization efficiency of the power supply system in each power supply scenario, and upload the optimal load scheduling result in each power supply scenario to the cloud control platform;

[0290] A current power supply scenario determination module 70, configured to determine a current power supply scenario in a power supply scenario set according to real-time operating parameters;

[0291] The load scheduling module 80 is used to take the optimal load scheduling result in the current power supply scenario as the target load scheduling result, and schedule the loads of the remaining monitoring points according to the target load scheduling result.

[0292] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0293] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0294] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0295] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0296] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0297] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0298] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0299] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0300] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. An online intelligent cloud control method for motor energy saving and consumption reduction management, characterized in that: Applied to a cloud control platform, the cloud control platform includes a database, the database stores no-load characteristic curves and load characteristic curves of monitoring points, the monitoring points include permanent magnet synchronous motors and three-phase asynchronous motors, the method includes: Obtaining real-time operating parameters of all monitoring points in the electric power supply system; Inputting the real-time operating parameters into a pre-built fault prediction model to obtain a fault monitoring point, a fault type and a fault time, and filtering out the fault monitoring point from all the monitoring points to obtain remaining monitoring points, wherein each of the remaining monitoring points is connected to at least one load device; Acquire historical operating parameters of all the monitoring points in the electric power supply system, and construct a power supply scenario set according to the historical operating parameters of all the monitoring points, wherein the power supply scenario set includes at least one power supply scenario; For each of the remaining monitoring points in each of the power supply scenarios, the operating currents of all the load devices connected to the remaining monitoring points within a preset time period are obtained, and according to the operating currents, a target characteristic curve matching the remaining monitoring points is determined in the database; In a case where the target characteristic curve is the load characteristic curve, determining the load requirement of each of the remaining monitoring points in each of the power supply scenarios according to the load characteristic curve in each of the power supply scenarios; The load demand of each of the remaining monitoring points in each of the power supply scenarios is transmitted to the edge node, and the edge node is used to perform load optimization calculation according to the load demand, obtain the best load scheduling result that maximizes the energy utilization efficiency of the power supply system in each of the power supply scenarios, and upload the best load scheduling result in each of the power supply scenarios to the cloud control platform; Determine a current power supply scenario in the power supply scenario set according to the real-time operating parameters; The optimal load scheduling result under the current power supply scenario is used as the target load scheduling result, and the loads of the remaining monitoring points are scheduled according to the target load scheduling result.

2. The method according to claim 1, characterized in that The step of constructing a power supply scenario set according to the historical operating parameters of all the monitoring points includes: Performing feature extraction on the historical operating parameters of all the monitoring points to obtain at least one target feature for describing the power supply scenario; Clustering all the target features to obtain at least one clustering result, wherein each clustering result is used to characterize one of the power supply scenarios, and a set of all the clustering results is the power supply scenario set; Each of the power supply scenarios includes power supply characteristics, power supply time, load level and total available power of the power supply system.

3. The method according to claim 2, characterized in that The determining, according to the load characteristic curve in each power supply scenario, the load requirement of each remaining monitoring point in each power supply scenario includes: Determine the operating efficiency and output power of each of the remaining monitoring points in each of the power supply scenarios according to the load characteristic curve in each of the power supply scenarios; For each of the remaining monitoring points, the input power of the remaining monitoring points in each of the power supply scenarios is calculated based on the operating efficiency and the output power, and the input power is used to characterize the load demand of the remaining monitoring points in each of the power supply scenarios, wherein the total value of the load demand of all the remaining monitoring points in each of the power supply scenarios is less than the total available power.

4. The method according to claim 3, characterized in that The edge node performs load optimization calculation according to the load demand to obtain an optimal load scheduling result that minimizes the energy utilization efficiency of the electric power supply system in each power supply scenario, including: In response to receiving the transmission instruction of the load demand, the edge node obtains the voltage, current, rotation speed of each of the remaining monitoring points and the flow and wind pressure of each load device connected to the remaining monitoring points through the cloud control platform, and the transmission instruction is used to transmit the load demand of each of the remaining monitoring points in each of the power supply scenarios to the edge node; According to the voltage, current, rotation speed of each of the remaining monitoring points and the flow rate and wind pressure of each load device connected to the remaining monitoring points, a load optimization model is constructed, wherein the optimization goal of the load optimization model is to maximize energy utilization efficiency, and the constraint condition is that the total value of the load demand of all the remaining monitoring points in each power supply scenario is less than the total available power; Using a preset optimization algorithm to solve the load optimization model and obtain at least one load scheduling result; According to the load scheduling result, calculating the actual working efficiency of each of the remaining monitoring points in each of the power supply scenarios; The load scheduling result with the highest actual working efficiency is taken as the optimal load scheduling result.

5. The method according to claim 4, characterized in that After taking the load scheduling result with the highest actual working efficiency as the best load scheduling result, the method further includes: According to the operating efficiency and the actual working efficiency of each of the remaining monitoring points in each of the power supply scenarios, the power saving rate in each of the power supply scenarios is calculated, and the power saving rate is uploaded to the cloud control platform.

6. The method according to claim 1, characterized in that Determining the current power supply scenario in the power supply scenario set according to the real-time operating parameters includes: Extracting features of the real-time operating parameters to obtain real-time feature vectors; Performing time series decomposition on the real-time feature vector to obtain residual features; Extracting random fluctuation features and mutation features from the residual features; A current power supply scenario is determined in the power supply scenario set according to the random fluctuation feature and the mutation feature.

7. The method according to claim 6, characterized in that The determining, in the power supply scenario set according to the random fluctuation feature and the mutation feature, a current power supply scenario includes: Acquire historical random fluctuation characteristics and historical mutation characteristics of each of the power supply scenarios in the power supply scenario set; Calculating a first similarity between the random fluctuation feature and the historical random fluctuation feature; Calculating a second similarity between the mutation feature and the historical mutation feature; Calculating a comprehensive similarity based on the first similarity and the second similarity; The power supply scenario corresponding to the largest comprehensive similarity is used as the current power supply scenario.

8. The method according to claim 1, characterized in that After inputting the real-time operating parameters into the pre-built fault prediction model to obtain the fault monitoring point, fault type and fault time, the method further includes: Constructing a digital twin model corresponding to the electric power supply system, wherein the digital twin model includes the permanent magnet synchronous motor and the three-phase asynchronous motor consistent with the electric power supply system; Simulating the fault monitoring point, the fault type and the fault time in the digital twin model, and determining the impact of the fault monitoring point on other monitoring points in the digital twin model; According to the influence of the fault monitoring point and the fault time on other monitoring points in the digital twin model, the influence scope is predicted and displayed in the digital twin model.

9. The method according to claim 8, characterized in that Determining the influence of the fault monitoring point on other monitoring points in the digital twin model includes: Injecting the fault type and the fault time into the fault monitoring point in the digital twin model, and running the digital twin model to obtain operation data of each of the monitoring points when the digital twin model is running; Determining whether the operating data of each of the monitoring points is abnormal; In the case where the operation data of the monitoring point is abnormal, a fault monitoring point closest to the monitoring point is determined, and the monitoring point is determined to be affected by the fault monitoring point closest to the monitoring point.

10. A cloud control platform, characterized in that: The cloud control platform includes a database storing no-load characteristic curves and load characteristic curves of monitoring points, wherein the monitoring points include permanent magnet synchronous motors and three-phase asynchronous motors. The cloud control platform also includes: An acquisition module, used to acquire real-time operating parameters of all the monitoring points in the electric power supply system; A remaining monitoring point determination module is used to input the real-time operating parameters into a pre-built fault prediction model to obtain a fault monitoring point, a fault type and a fault time, and to filter out the fault monitoring point from all the monitoring points to obtain remaining monitoring points, wherein each of the remaining monitoring points is connected to at least one load device; A power supply scenario set construction module, used to obtain historical operating parameters of all the monitoring points in the electric power supply system, and construct a power supply scenario set according to the historical operating parameters of all the monitoring points, wherein the power supply scenario set includes at least one power supply scenario; a target characteristic curve determination module, configured to obtain, for each of the remaining monitoring points in each of the power supply scenarios, the operating currents of all the load devices connected to the remaining monitoring points within a preset time period, and determine, in the database, a target characteristic curve matching the remaining monitoring points based on the operating currents; A load demand determination module, configured to determine the load demand of each of the remaining monitoring points in each of the power supply scenarios according to the load characteristic curve in each of the power supply scenarios when the target characteristic curve is the load characteristic curve; An optimal load scheduling result determination module is used to transmit the load demand of each of the remaining monitoring points in each of the power supply scenarios to the edge node, and the edge node is used to perform load optimization calculation according to the load demand to obtain the optimal load scheduling result that maximizes the energy utilization efficiency of the power supply system in each of the power supply scenarios, and upload the optimal load scheduling result in each of the power supply scenarios to the cloud control platform; A current power supply scenario determination module, configured to determine a current power supply scenario in the power supply scenario set according to the real-time operating parameters; The load scheduling module is used to use the optimal load scheduling result in the current power supply scenario as the target load scheduling result, and schedule the loads of the remaining monitoring points according to the target load scheduling result.

Images