Motor cluster remote monitoring and energy efficiency optimization method and system based on cloud platform

Abstract

The application provides a motor cluster remote monitoring and energy efficiency optimization method and system based on a cloud platform, and relates to the technical field of motor monitoring.The method comprises the following steps: establishing a double-track parallel monitoring framework on a cloud platform, setting monitoring target parameters and double-channel parallel interaction parameters respectively; constructing an algorithm power structure diagram of the cloud platform; monitoring and distributing the algorithm power structure diagram by using the monitoring target parameters and the parallel interaction parameters; collecting motor cluster monitoring data and analyzing the motor state by mapping node algorithm power; and adaptively and evenly optimizing motor energy efficiency state evaluation results according to load data and the motor state of the motor cluster, and determining a motor cluster control optimization strategy.The application can solve the technical problem that the energy efficiency of a motor cluster is difficult to improve effectively in the prior art due to single monitoring dimension and difficulty in coping with load changes in the motor cluster, a double-track parallel monitoring framework is established on a cloud platform, and the operation efficiency of the motor cluster is improved.

Description

Technical Field

[0001] This application relates to the field of motor monitoring technology, specifically to a method and system for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform. Background Technology

[0002] Traditional motor cluster monitoring methods typically employ physical sensors and local control systems. While these methods can collect basic parameters such as motor current, voltage, and temperature in real time, their monitoring dimensions are relatively limited, focusing on a single operating state or local parameters, lacking comprehensive analysis of multi-dimensional data. This presents limitations when facing dynamic factors such as load fluctuations and changes in operating conditions within the motor cluster. Because load conditions frequently change during motor cluster operation, significantly impacting motor efficiency and stability, current monitoring methods often lack the ability to dynamically adapt to load variations and cannot make adaptive adjustments based on real-time data. Consequently, the motor cluster fails to maintain optimal operating conditions during load changes, limiting energy efficiency improvements and hindering precise energy efficiency optimization under different operating conditions. Furthermore, it fails to achieve optimal load distribution and operating parameter adjustments under varying load conditions.

[0003] In summary, existing technologies suffer from the problem that due to the limited monitoring dimensions, it is difficult to cope with load changes in motor clusters, which makes it difficult to effectively improve the energy efficiency of motor clusters and further affects their operating efficiency. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform, in order to solve the technical problem in the prior art that the single monitoring dimension makes it difficult to cope with load changes in the motor cluster, which makes it difficult to effectively improve the energy efficiency of the motor cluster and further affects the operating efficiency of the motor cluster.

[0005] To achieve the above objectives, this application provides a method and system for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform.

[0006] Firstly, this application provides a method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform. This method is implemented through a cloud-based system for remote monitoring and energy efficiency optimization of motor clusters. The method includes: establishing a dual-track parallel monitoring framework on the cloud platform, setting monitoring target parameters for the first and second tracks, and setting dual-channel parallel interaction parameters; constructing a computing power structure diagram of the cloud platform, including the cloud-edge topology and the computing power relationship of each topology node; using the monitoring target parameters and parallel interaction parameters to monitor and allocate the computing power structure diagram, establishing a computing power mapping relationship; collecting motor cluster monitoring data based on the computing power mapping relationship, and performing motor status analysis through the computing power of the mapping nodes to obtain the motor energy efficiency status evaluation results of each node; and performing adaptive equilibrium optimization on the motor energy efficiency status evaluation results based on load data and the motor status of the motor cluster to determine the motor cluster control optimization strategy.

[0007] Optionally, with energy efficiency status as the target, a rapid monitoring track and a learning monitoring track are established. The monitoring target parameters of the first track and the second track are set according to the monitoring time period of the rapid monitoring track and the learning monitoring track. Based on the gradient relationship of the monitoring time period of the first track and the second track, an interactive learning relationship of the monitoring target parameters of the first track and the second track is established. The interactive learning process is supported by parameter parsing to obtain parallel interactive parameters.

[0008] Optionally, based on the gradient relationship of the monitoring time periods of the first and second tracks, a time-cumulative analysis of the motor energy efficiency status is performed on historical sample data to obtain the cumulative impact time-series relationship of the gradient periods; based on the cumulative impact time-series relationship, the guidance parameters of the period nodes of the first track are analyzed based on the time-series relationship of the second track, and optimization feedback is performed on the first track; based on the cumulative impact time-series relationship, the influence relationship coefficient of the first track on the second track is analyzed, and learning prediction is performed on the second track; parallel interaction parameters are generated based on the guidance parameters of the first track and the learning influence relationship coefficient of the second track.

[0009] Optionally, the guidance parameters for the first track include at least one of the following: dynamic threshold parameters, feature weight parameters, and strategy control parameters. The dynamic threshold parameters are used to replace the original fixed alarm or warning thresholds in the first track; the feature weight parameters are used to adjust the feature importance weights of the energy efficiency status evaluation algorithm in the first track; and the strategy control parameters are used to define or modify the control strategy executed by the first track under a predefined set of operating states.

[0010] Optionally, taking energy efficiency status as the target event, monitoring data is correlated and parsed based on motor operation scenario data to establish a parsing relationship tree; based on the parsing relationship tree, layer-by-layer aggregation is performed to obtain an energy efficiency status monitoring relationship set; according to the monitoring time period of the first track and the second track, the collection frequency of the energy efficiency status monitoring relationship set is matched to obtain the monitoring target parameters of the first track and the second track.

[0011] Optionally, based on the parsed relation tree, according to the labeled topological relationship and causal edge, feature aggregation is performed layer by layer from bottom to top according to the same causal attribute to obtain the causal attribute monitoring influence relationship; according to the influence relationship between the causal attribute and the energy efficiency status evaluation, the causal attribute monitoring influence relationship is merged, and the monitoring parameters are extracted according to the monitoring parameters and correlation relationships involved in the merged relationship to obtain the energy efficiency status monitoring relationship set.

[0012] Optionally, a first track window and a second track window are set according to the monitoring time period of the first track and the second track, respectively; the monitoring quantities with different sampling periods are resampled and time-aligned according to the first track window and the second track window; the monitoring quantities with a sampling period no greater than that of the first track window are assigned to the first track, and the monitoring quantities with a sampling period greater than that of the first track window but not greater than that of the second track window are assigned to the second track; the periodic sampling monitoring quantity that each track can support is used as a constraint to select and match monitoring parameters from the energy efficiency status monitoring relationship set as the monitoring target parameters of the corresponding track.

[0013] Optionally, the monitoring time periods of the first and second tracks are analyzed to determine the corresponding maximum allowable response time and computation duration. Based on the computing power structure diagram, with the maximum allowable response time as the first constraint, the monitoring task of the first track is mapped to the edge computing power node that meets the maximum end-to-end latency requirement and the task computation duration, establishing a first mapping relationship. With the computation duration as the second constraint, based on the computing power structure diagram, the monitoring and analysis task of the second track is mapped to the cloud platform central computing power node with sufficient batch computing resources, establishing a second mapping relationship. According to the data flow and instruction flow relationship defined by the parallel interaction parameters, the first and second mapping relationships are collaboratively optimized and adjusted to obtain a globally optimized computing power mapping relationship. The collaborative optimization and adjustment is based on satisfying the first and second constraints, with the goal of minimizing cloud-edge communication overhead and computing resource conflicts.

[0014] Optionally, by integrating the high-frequency real-time energy efficiency status data of the first track with the long-term energy efficiency trend evaluation results generated by the second track, rolling time-domain predictions of the future energy efficiency status of each appliance in the cluster are performed to construct a multi-scale energy efficiency status prediction data chain. Adjustment coefficients are set according to the motor status of the motor cluster. Based on the multi-scale energy efficiency status prediction data chain, with the goal of maximizing multi-scale time-series cumulative energy efficiency and using the load data as a constraint, the operating parameters of each motor are optimized based on the adjustment coefficients to obtain the optimal load allocation sequence of each motor in the multi-scale time domain, thus obtaining the motor cluster control optimization strategy.

[0015] Secondly, this application also provides a cloud-based remote monitoring and energy efficiency optimization system for motor clusters, used to execute the cloud-based remote monitoring and energy efficiency optimization method for motor clusters as described in the first aspect. The cloud-based remote monitoring and energy efficiency optimization system for motor clusters includes: a monitoring framework establishment module, used to establish a dual-track parallel monitoring framework on the cloud platform, setting monitoring target parameters for the first track and the second track, and dual-channel parallel interaction parameters respectively; a computing power structure graph construction module, used to construct a computing power structure graph of the cloud platform, including the cloud-edge topology and the computing power relationship of each topology node; a mapping relationship establishment module, used to use the monitoring target parameters and parallel interaction parameters to monitor and allocate the computing power structure graph, establishing a computing power mapping relationship; a motor status analysis module, used to collect motor cluster monitoring data based on the computing power mapping relationship, and perform motor status analysis through the computing power of the mapping nodes to obtain the motor energy efficiency status evaluation results of each node; and a balancing optimization module, used to perform adaptive balancing optimization on the motor energy efficiency status evaluation results based on load data and the motor status of the motor cluster, determining the motor cluster control optimization strategy.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0017] By establishing a dual-track parallel monitoring framework on a cloud platform, monitoring target parameters for the first and second tracks, as well as dual-channel parallel interaction parameters, are set respectively. A computing power structure diagram of the cloud platform is constructed, including the cloud-edge topology and the computing power relationships of each topology node. The computing power structure diagram is monitored and allocated using the monitoring target parameters and parallel interaction parameters to establish a computing power mapping relationship. Based on this mapping relationship, monitoring data of the motor cluster is collected, and motor status is analyzed using the computing power of the mapped nodes to obtain the motor energy efficiency status evaluation results for each node. Based on the load data and the motor status of the motor cluster, adaptive equilibrium optimization is performed on the motor energy efficiency status evaluation results to determine the motor cluster control optimization strategy. In other words, by establishing a dual-track parallel monitoring framework on the cloud platform, the motor control strategy is dynamically adjusted according to the real-time operating status and load changes of the motor cluster, maximizing energy efficiency, avoiding energy waste caused by uneven load or improper status, and significantly improving the overall energy efficiency and operating efficiency of the motor cluster.

[0018] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the cloud-based method for remote monitoring and energy efficiency optimization of motor clusters.

[0021] Figure 2 This is a schematic diagram of the structure of the cloud-based motor cluster remote monitoring and energy efficiency optimization system of this application.

[0022] Explanation of reference numerals in the attached diagram: 11 Monitoring framework establishment module, 12 Computing power structure diagram construction module, 13 Mapping relationship establishment module, 14 Motor status analysis module, 15 Balancing optimization module. Detailed Implementation

[0023] This application provides a cloud-based method and system for remote monitoring and energy efficiency optimization of motor clusters. It addresses the technical problem in existing technologies where the single monitoring dimension makes it difficult to address load variations within the motor cluster, hindering effective energy efficiency improvement and further impacting overall motor cluster operating efficiency. By establishing a dual-track parallel monitoring framework on the cloud platform, the control strategy of the motors is dynamically adjusted based on the real-time operating status and load changes of the motor cluster, maximizing energy efficiency and avoiding energy waste caused by uneven load or improper conditions. This significantly improves the overall energy efficiency and operating efficiency of the motor cluster.

[0024] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0025] Example 1, please refer to the appendix. Figure 1 This application provides a method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform. The method is applied to a cloud-based remote monitoring and energy efficiency optimization system for motor clusters, and specifically includes the following steps:

[0026] Establish a dual-track parallel monitoring framework on the cloud platform, and set the monitoring target parameters for the first track and the second track, as well as the dual-channel parallel interaction parameters.

[0027] Furthermore, this application also includes the following steps: taking energy efficiency status as the target, establishing a rapid monitoring track and a learning monitoring track, setting the monitoring target parameters of the first track and the second track according to the monitoring time period of the rapid monitoring track and the learning monitoring track; based on the gradient relationship of the monitoring time period of the first track and the second track, establishing an interactive learning relationship of the monitoring target parameters of the first track and the second track, performing support parameter parsing on the interactive learning process, and obtaining parallel interactive parameters.

[0028] Furthermore, this application also includes the following steps: based on the gradient relationship of the monitoring time periods of the first track and the second track, performing time-cumulative analysis of the motor energy efficiency status on historical sample data to obtain the cumulative impact time-series relationship of the gradient period; based on the cumulative impact time-series relationship, analyzing the guidance parameters of the period nodes of the first track based on the time-series relationship of the second track, and performing optimization feedback on the first track; based on the cumulative impact time-series relationship, analyzing the impact relationship coefficient of the first track on the second track, and performing learning prediction on the second track; generating parallel interaction parameters based on the guidance parameters of the first track and the learning impact relationship coefficient of the second track.

[0029] Furthermore, this application also includes the following steps: the guidance parameters of the first track include at least one of the following: dynamic threshold parameters, feature weight parameters, and strategy control parameters, wherein the dynamic threshold parameters are used to replace the original fixed alarm or warning thresholds in the first track; the feature weight parameters are used to adjust the feature importance weights of the energy efficiency status evaluation algorithm in the first track; and the strategy control parameters are used to define or modify the control strategy executed by the first track under a predefined set of operating states.

[0030] Specifically, a dual-track parallel monitoring framework is established on the cloud platform. This framework enhances the monitoring accuracy and energy efficiency optimization capabilities of the motor cluster by setting up separate fast monitoring and learning monitoring tracks. The dual-track parallel monitoring framework refers to running two monitoring methods simultaneously on the cloud platform: a fast monitoring track and a learning monitoring track. The fast monitoring track is used for real-time monitoring of the motor cluster's operating status and has a shorter monitoring cycle; while the learning monitoring track is used for long-term trend analysis and optimization and has a longer monitoring cycle. In simpler terms, it gradually learns from the results given by the fast monitoring channel for long-term monitoring. The two work together, using different sampling frequencies to achieve more refined motor energy efficiency management.

[0031] Energy efficiency status refers to the energy efficiency performance of a motor under current load and operating conditions. It is typically measured by parameters such as motor power consumption and efficiency, and is determined by monitoring the ratio of the motor's output power to its power consumption. Based on monitoring requirements, the monitoring time periods for the fast monitoring track and the learning monitoring track are set. The fast monitoring track (track 1) is used for real-time monitoring of the motor cluster's operating status, with a short monitoring cycle, typically a few seconds or tens of seconds, enabling rapid response to any sudden motor failures or performance changes. The learning monitoring track (track 2) is used to analyze the long-term trends of the motor cluster, with a longer monitoring cycle, typically one hour.

[0032] Based on the monitoring time cycles of the rapid monitoring track and the learning monitoring track, the monitoring target parameters for the first and second tracks are set. These are the key indicators that need to be monitored and measured during the two channels, such as motor current, voltage, temperature, and load. For example, the monitoring target parameters for the rapid monitoring track (first track) mainly focus on real-time current, voltage, and temperature to quickly respond to any abnormalities that occur during motor operation. The monitoring target parameters for the learning monitoring track (second track) are set according to the long-term trends of motor operation, such as long-term power consumption data, load balancing status, and energy efficiency assessment results.

[0033] Based on the monitoring time periods of the first and second tracks, a gradient relationship is established, representing the difference in sampling frequency between the rapid monitoring track and the learning monitoring track. This reflects the hierarchical difference between short-cycle and long-cycle monitoring data. Historical sample data refers to motor operation data collected within a certain time range (e.g., one week, one month), including operating parameters such as motor current, voltage, temperature, and power, reflecting the motor's performance under different operating conditions. Based on the gradient relationship of the monitoring time periods, a time-cumulative distribution is performed on the historical data, accumulating the motor's energy efficiency status within different time windows. For example, power and efficiency fluctuations of the motor are calculated in the short term (e.g., every 10 seconds or every minute), and these fluctuations are then gradually summarized into a long-term energy efficiency change trend. Time-cumulative analysis refers to the accumulation and summarization of motor operation data based on different time windows (e.g., hours, days, weeks). By accumulating historical data over time windows, the performance trends of the motor at different time scales can be revealed, such as short-term energy efficiency fluctuations and long-term energy efficiency decline trends.

[0034] The first and second tracks are mutually analyzed and optimized based on the cumulative impact time-series relationship. The second track is responsible for long-term energy efficiency trend analysis. For example, by analyzing the energy efficiency data of a motor cluster over the past hour, it identifies certain motors with poor energy efficiency under long-term low-load conditions, and provides feedback to the first track for optimization. Suppose the second track detects frequent efficiency drops in motors under low load conditions (i.e., load rate below 20% of the rated value), it sends guidance parameters to the first track to adjust the alarm threshold, raising the current alarm threshold from 100A to 115A to avoid false alarms and reminding the operator to pay attention to the load status of the motor. The first track adjusts its monitoring strategy in real time based on the feedback guidance parameters. The time-series relationship of the second track (i.e., the learning monitoring track) is the relationship formed by data analysis over a longer time period. The guidance parameters for each cycle node of the first track are generated from the analysis results of the second track, designed to guide the operation of the first track (i.e., the rapid monitoring track) at each monitoring cycle node.

[0035] Based on the cumulative impact time series relationship, the influence coefficient of the first track on the second track is analyzed. Short-term sampling data of the first track will affect the long-term trend of the second track. According to the cumulative impact time series relationship, the influence coefficient of the first track data on the second track is analyzed. For example, assuming that a short-term overload event with a current exceeding 120A has a 5% negative impact on the energy efficiency status of the second track, the influence coefficient is 0.5, indicating that the overload event of the first track contributes 50% to the energy efficiency decline of the second track. Based on the influence coefficient, the second track is learned and predicted to determine the specific impact of the first track data on the second track, thereby achieving more accurate long-term energy efficiency prediction and control.

[0036] Parallel interaction parameters are generated based on the guidance parameters of the first track and the learning influence coefficients of the second track. The guidance parameters of the first track include at least one of the following: dynamic threshold parameters, feature weight parameters, and strategy control parameters. The dynamic threshold parameters are alarm thresholds dynamically adjusted based on real-time data during monitoring, rather than using fixed thresholds, and are used to replace the original fixed alarm or warning thresholds in the first track. The feature weight parameters assign different weights to different features (such as temperature, current, power, etc.) when evaluating motor energy efficiency status, and are used to adjust the feature importance weights of the energy efficiency status evaluation algorithm in the first track. The strategy control parameters are used to define or modify the control strategy executed by the first track under a predefined set of operating states. The predefined set of operating states includes at least one of the following: the motor's start / stop transient process, overload operation state, light load operation state with a load rate lower than 20% of the rated value, and low-efficiency operation ranges mined from historical data.

[0037] Parallel interaction parameters are parameters that influence and coordinate the optimization between the first and second tracks, including dynamic thresholds, feature weights, and control strategies. Based on these parameters, the two monitoring tracks can work collaboratively at different time scales, improving the operational efficiency of the motor cluster.

[0038] Furthermore, this application also includes the following steps: taking energy efficiency status as the target event, performing monitoring data association analysis based on motor operation scenario data to establish an analytical relationship tree; performing layer-by-layer aggregation based on the analytical relationship tree to obtain an energy efficiency status monitoring relationship set; and performing sampling frequency matching on the energy efficiency status monitoring relationship set according to the monitoring time period of the first track and the second track to obtain the monitoring target parameters of the first track and the second track.

[0039] Furthermore, this application also includes the following steps: based on the parsed relation tree, according to the labeled topological relationship and causal edge, perform layer-by-layer feature aggregation from bottom to top according to the same causal attribute to obtain the causal attribute monitoring influence relationship; according to the influence relationship between the causal attribute and the energy efficiency status evaluation, merge the causal attribute monitoring influence relationships, extract the monitoring parameters according to the monitoring parameters and correlation relationships involved in the merged relationship, and obtain the energy efficiency status monitoring relationship set.

[0040] Furthermore, this application also includes the following steps: setting a first track window and a second track window respectively according to the monitoring time period of the first track and the second track; resampling and time-aligning the monitoring quantities with different sampling periods according to the first track window and the second track window; assigning the monitoring quantities with a sampling period no greater than that of the first track window to the first track, and assigning the monitoring quantities with a sampling period greater than that of the first track window but not greater than that of the second track window to the second track; using the periodic sampling monitoring quantity that each track can support as a constraint, selecting matching monitoring parameters from the energy efficiency status monitoring relationship set as the monitoring target parameters of the corresponding track.

[0041] Specifically, this involves acquiring motor operating scenario data, including various environmental and state data of the motor during actual operation, such as load, speed, current, voltage, and temperature. Energy efficiency status is used as the target event, and monitoring data correlation analysis is performed using the motor operating scenario data. Monitoring data correlation analysis refers to identifying the correlations between different data points by analyzing the relationships between motor operating data. For example, there is usually a high correlation between current and load, because changes in the motor's load cause fluctuations in current. Based on the data correlations, a relational tree is built. The relational tree displays the causal relationships between data through a hierarchical structure. For example, the root node represents the motor's energy efficiency status, i.e., the final target event; intermediate nodes represent various monitoring data points, such as current, temperature, and load, which directly or indirectly affect the motor's energy efficiency status; leaf nodes represent specific monitoring indicators or data points, such as real-time current values ​​and real-time load values.

[0042] Based on the established analytical relationship tree, feature aggregation is performed layer by layer on the monitoring data. The bottom-up feature aggregation process means starting from the bottom of the tree and gradually summarizing features from each layer. For example, starting with current data, the relationship between it and temperature and load is analyzed, and then upper-layer data is aggregated based on this information, gradually extracting the influence degree of each monitoring data point and their influence coefficients on the motor's energy efficiency status. According to the labeled topological relationships and causal edges in the analytical relationship tree, feature aggregation is performed layer by layer from bottom to top according to the same causal attributes. In the analytical relationship tree, each node is labeled, and the hierarchical structure between these nodes is defined. The connection relationships between nodes at each layer are labeled as causal edges. Causal attributes refer to monitoring parameters that can directly or indirectly affect the motor's status during the motor energy efficiency assessment process; the same causal attribute refers to monitoring parameters that affect the same parameters. Bottom-up feature aggregation means starting from the bottom layer, gradually merging various features in the motor operating data, and calculating their impact on energy efficiency status. In the first layer, the direct impact of basic parameters on motor energy efficiency is analyzed, and the preliminary influence coefficient of each bottom-level monitoring parameter on energy efficiency status is obtained. In the second and higher layers, the influence results of the previous layer are gradually aggregated.

[0043] Based on the causal edges and influence coefficients in the parse tree, the impact of each causal attribute on the energy efficiency status is quantified, resulting in the calculation of the causal attribute influence relationship. Causal attribute monitoring and influence relationship assessment refers to identifying the degree of influence of the causal attributes of each node in the parse tree on the motor's energy efficiency status.

[0044] Based on the influence relationship between causal attributes and energy efficiency status evaluation, that is, by analyzing the relationship between causal attributes and motor energy efficiency status, the degree of influence of each causal attribute on motor energy efficiency status is assessed. The monitoring influence relationships of each causal attribute are merged to form a more comprehensive monitoring relationship. Based on the merged influence relationship, the most critical monitoring parameters, such as current, load, and temperature, are extracted to form an energy efficiency status monitoring relationship set, which includes all monitoring parameters that have a significant impact on motor energy efficiency and their causal relationships, for real-time monitoring of the motor's operating status.

[0045] For example, suppose there is a motor cluster containing 5 motors, each with a rated power of 75kW. Real-time data collection is performed on current, load, and temperature. The motor current fluctuates between 70A and 100A; the motor load varies between 60% and 85%; and the motor temperature varies from 30℃ to 70℃. The correlation between current and load is 0.9, between current and temperature is 0.75, and between load and temperature is 0.6. The bottom-level nodes of the relational tree are current and load, the middle nodes are temperature and power, and the root node is energy efficiency status. First, the influence coefficients of current and load are calculated. The influence coefficient of current on temperature is 0.7, and the influence coefficient of load on temperature is 0.5. The influence coefficient of temperature on energy efficiency status is 0.8, thus deriving the influence relationships of current, load, and temperature on energy efficiency status. Combining these relationships yields a comprehensive influence model. Through this combination, it is found that current and load are the main factors affecting motor energy efficiency. Current, load, and temperature were extracted as key monitoring parameters and used for real-time monitoring and optimization. An energy efficiency status monitoring relationship set was generated, including parameters such as current, load, and temperature, as well as the causal relationships between them, for real-time evaluation of the motor's energy efficiency status.

[0046] The first and second track windows are set according to the sampling period of each track. The first track window corresponds to the first track and is usually set to a shorter time period, such as 10 seconds or 1 minute, to collect and analyze short-term operating data of the motor in real time and capture rapid changes in the motor's state. The second track window corresponds to the second track and is usually set to a longer time period, such as 10 minutes or 1 hour, to analyze the long-term operating trend of the motor and identify long-term trends such as declining energy efficiency.

[0047] Because the data sampling periods for the first and second tracks differ, resampling and time alignment are performed on the monitoring data with different sampling periods according to the first and second track windows. Through interpolation or other algorithms, data with longer sampling periods are converted into a format aligned with shorter period data. For example, if the first track samples every 10 seconds and the second track samples every 10 minutes, the data from the second track needs to be resampled to a 10-second sampling period to align with the data from the first track. Resampling refers to converting data into a uniform time interval, ensuring that data from different periods can be compared and analyzed at the same point in time. For data with different sampling periods, it is ensured that they can be compared and analyzed at the same time point. Time alignment ensures that all monitoring data are arranged in the correct chronological order within each time window.

[0048] Quantities with sampling periods no greater than the first track window are assigned to the first track. For example, if the first track window is 10 seconds, all data with sampling periods of 10 seconds or less (such as current and voltage) will be assigned to the first track. Quantities with sampling periods greater than the first track window but not greater than the second track window are assigned to the second track. For example, if the second track window is 10 minutes, data with sampling periods from 10 seconds to 10 minutes (such as temperature and power) will be assigned to the second track.

[0049] Constrained by the periodic sampling monitoring volume supported by each track, matching monitoring parameters are selected from the energy efficiency status monitoring relationship set as the monitoring target parameters for each track. For the first track, monitoring parameters with short sampling periods and significant impact on motor energy efficiency are selected. For example, parameters such as current, load, and temperature typically have short sampling periods and are used as the monitoring target parameters for the first track. For the second track, monitoring parameters with longer sampling periods that can reflect the long-term operating trend of the motor are selected, such as average temperature, power consumption, and long-term load.

[0050] Monitoring is conducted through two dimensions: rapid monitoring tracks and learning-based monitoring tracks. The cloud platform's computing power and processing characteristics are used for matching to ensure effective monitoring and optimization. Monitoring target parameters refer to the key indicators monitored and analyzed for each track. For the first track, these are typically data with significant short-term fluctuations, while for the second track, the monitoring target is usually a long-term trend indicator. Dual-channel parallel interaction parameters refer to the data or information exchanged between the first and second tracks, helping the two tracks coordinate their work. This allows the first track to provide real-time feedback to the second track, and vice versa.

[0051] Construct a computing power structure diagram for the cloud platform, including the cloud-edge topology and the computing power relationships between each topology node.

[0052] Specifically, a cloud platform is an internet-based computing platform used in motor cluster monitoring to centrally store, process, and analyze data collected from various motors, providing powerful computing capabilities for data analysis. A computing power structure diagram is a visual representation of the computing resources (computing power) of each node in the cloud platform and the relationships between them, showing the interconnections between different computing resources, data flow, and resource allocation methods. For example, the cloud may include multiple computing nodes, database nodes, and data storage nodes. Computing nodes perform data analysis tasks, such as real-time data processing, trend prediction, and model training. Edge nodes are local computing units connecting the sensors and monitoring equipment of the motor cluster. Edge computing nodes are typically responsible for real-time data acquisition and preliminary preprocessing, such as data cleaning and preliminary analysis. Edge nodes have relatively weaker computing power but can reduce data transmission latency and improve system response speed.

[0053] By arranging and connecting different computing resources, a cloud-edge topology is constructed, including an edge layer, a communication layer, and a cloud layer. The edge layer includes monitoring devices, sensors, and local gateways for each motor; data acquisition and initial data processing are completed at this layer. The communication layer connects the edge layer to the cloud platform via communication networks, including LANs and WANs, ensuring data transmission from the motor cluster to the cloud. The cloud layer includes the computing nodes, storage nodes, and database of the cloud computing platform. The cloud platform is responsible for processing the large-scale data transmitted from the edge layer, performing advanced analysis and storage, and optimizing the operation of the motor cluster based on the analysis results.

[0054] Each node represents a computing, storage, or communication resource. For example, edge nodes include motor monitoring equipment, sensors, gateways, etc., and are responsible for collecting data from the motor and transmitting it to the cloud. Cloud computing nodes are used to process data from the edge layer and perform tasks such as analysis, calculation, and prediction. Storage nodes are responsible for storing real-time and historical data for subsequent analysis, backup, and query. Database nodes are used to store and manage motor performance indicators, maintenance records, and prediction results.

[0055] The computing power relationship determines the computing tasks and resource allocation for each topology node. For example, edge nodes typically have lower computing power, and their tasks mainly involve data acquisition, preliminary analysis, and transmission. Edge nodes may reduce data transmission latency through local caching. Cloud computing nodes, on the other hand, require powerful computing capabilities to handle tasks such as large-scale data analysis, model training, and algorithm optimization. Cloud platform computing resources are allocated on demand and may involve multiple virtual machines, containers, and computing clusters. To ensure high availability and stability, cloud computing resources are typically dynamically allocated using load balancing algorithms. For instance, when a computing node is overloaded, tasks are automatically distributed to other nodes with lighter loads.

[0056] Data is first collected from the motor's sensors and monitoring equipment, undergoing initial preprocessing at the edge layer. After data cleaning and compression, the data is transmitted to the cloud platform via the communication layer. The cloud performs in-depth analysis of the data, extracting information such as motor performance indicators and energy efficiency assessments, and returns optimization suggestions. Edge nodes collect and perform preliminary analysis of real-time data. Due to limited resources, edge nodes have relatively simple task allocation, primarily handling low-complexity operations. Cloud computing nodes handle more complex tasks, such as big data analysis, model training, and algorithm optimization. Based on task priority and the availability of computing resources, computing tasks are automatically assigned to appropriate nodes.

[0057] Cloud platform computing resources are typically elastic and scalable, dynamically adjusting resources based on the monitoring needs of motor clusters. For example, when motor operating data increases over a period of time, the cloud platform can automatically increase the number of computing nodes to handle high-concurrency tasks. To prevent some computing nodes from overloading, the cloud platform uses load balancing mechanisms to ensure that computing tasks are reasonably distributed among the nodes, thereby optimizing overall computing efficiency and avoiding bottlenecks on individual nodes.

[0058] The cloud-edge topology refers to the connection structure between the cloud platform and various endpoints, such as local sensors, gateways, and computing nodes. It describes how data flows from the monitoring equipment of the motor cluster to the cloud platform for processing and storage via the network. The computing power relationship between the topology nodes refers to the allocation and utilization of computing resources among different topology nodes, describing how different nodes collaborate to complete tasks, such as distributed data processing, load balancing, and the allocation of computing tasks. By constructing the computing power structure diagram of the cloud platform and rationally configuring the cloud-edge topology, real-time and long-term processing and analysis of large amounts of data from the motor cluster can be performed, thereby improving the accuracy and response speed of monitoring.

[0059] The computing power structure diagram is monitored and allocated using the monitoring target parameters and parallel interaction parameters to establish a computing power mapping relationship.

[0060] Furthermore, this application also includes the following steps: analyzing the monitoring time periods of the first track and the second track to determine the corresponding maximum allowable response time and computation duration; based on the computing power structure diagram, with the maximum allowable response time as the first constraint, mapping the monitoring task of the first track to an edge computing power node that meets the maximum end-to-end latency requirement and task computation duration, and establishing a first mapping relationship; with the computation duration as the second constraint, based on the computing power structure diagram, mapping the monitoring and analysis task of the second track to a cloud platform central computing power node with sufficient batch computing resources, and establishing a second mapping relationship; according to the data flow and instruction flow relationship defined by the parallel interaction parameters, co-optimizing and adjusting the first mapping relationship and the second mapping relationship to obtain a globally optimized computing power mapping relationship, wherein the co-optimization and adjustment is based on satisfying the first constraint and the second constraint, and aims to minimize cloud-edge communication overhead and computing resource conflicts.

[0061] Specifically, based on the monitoring time periods of the first and second tracks, the corresponding maximum allowable response time and computation duration are determined. The maximum allowable response time refers to the maximum time interval from the occurrence of an event to the monitoring system's response, and the computation duration refers to the time required to complete the computation task. Using the maximum allowable response time as the first constraint, the monitoring task of the first track is mapped to edge computing nodes. That is, according to the task requirements of the first track, the monitoring task is assigned to the edge node of the most suitable data source. This node needs to meet the maximum end-to-end latency requirement and the task computation duration.

[0062] The maximum allowable response time sets a time limit for the tasks in the first track, requiring each monitoring task to complete within this maximum time. In addition to the maximum response time, the actual computation time of the task must also be considered. Task computation time refers to the computation time required to complete each monitoring task. Based on the maximum allowable response time and task computation time, a suitable edge computing node is selected to handle the monitoring tasks in the first track. The edge computing node needs to have sufficient computing power to complete task processing within the remaining time and meet the response time requirement. The maximum end-to-end latency requirement ensures that the total latency from task acquisition to processing does not exceed the maximum response time. For example, if data acquisition takes 2 seconds and computation time is 2 seconds, then the remaining time can be used for data transmission, such as 6 seconds. The computing power of the edge computing node needs to meet the processing requirements of real-time tasks. If the task is relatively simple, fewer computing resources on the edge node are sufficient. If the task is more complex, a more powerful edge node is required.

[0063] Based on constraints, and considering the task's computational requirements, response time requirements, and the computing capabilities of edge nodes, the monitoring tasks of the first track are mapped to suitable edge computing nodes. For example, a current monitoring task with a maximum response time of 10 seconds and a computation time of 2 seconds is available. Available edge computing nodes include edge node A (high computing power, low latency) and edge node B (low computing power, high latency). Based on this information, the current monitoring task is assigned to edge node A because it meets the 10-second maximum response time requirement and has sufficient computing power. The first mapping relationship refers to assigning the monitoring tasks of the first track to computing nodes that meet the maximum allowable response time requirement, taking into account the task's latency requirements, computation time, and the availability of computing resources.

[0064] Using computation duration as a second constraint, the monitoring and analysis tasks of the second track are mapped to the cloud platform's central computing nodes. These nodes possess strong computing capabilities, suitable for complex batch computing tasks. Computation duration refers to the time required for the second track task from data acquisition and processing to outputting analysis results. The computing resources of the cloud platform's central computing nodes must meet the needs of these long-duration computing tasks. For example, the second track task may require 10 minutes or more to complete data processing and analysis. Appropriate cloud platform computing nodes are selected based on the task's computation duration to ensure tasks are completed within a reasonable timeframe.

[0065] Based on the computation duration of the second track task, a cloud platform central node with strong computing power capable of handling large-scale batch computing tasks is selected. Cloud platform nodes are typically high-performance computing servers in data centers, capable of handling complex batch computing tasks. The cloud platform center provides elastic computing resources, dynamically adjusting resource allocation according to task requirements to ensure successful task completion. The computing resources of the cloud platform node are suitable for long-duration data analysis tasks, are not limited by latency, and can perform continuous computation.

[0066] After selecting appropriate cloud platform central computing nodes, the monitoring and analysis tasks of the second track are mapped to these computing nodes. For example, if the task type is a long-term temperature analysis task with a calculation duration of 10 minutes, a high-performance computing node is selected on the cloud platform to support the 10-minute analysis task, while also having sufficient storage resources to process large amounts of data.

[0067] After identifying the central computing nodes of the cloud platform and selecting appropriate computing resources, the monitoring and analysis tasks of the second track are mapped to these cloud computing nodes. The mapping relationship is primarily based on the computation duration, ensuring that nodes with sufficient computing resources are selected to meet the computational needs of the tasks. It is ensured that the computation duration of the tasks does not exceed the computing capacity of the selected cloud platform nodes. For example, if an analysis task takes 10 minutes to complete, a cloud platform node capable of completing the computation within that time is selected. The cloud platform's computing nodes can provide sufficient batch computing resources for the second track tasks, such as multi-core processors, high-speed storage, and large memory, thereby ensuring the efficient execution of the second track tasks. The second mapping relationship is the process of mapping the second track monitoring and analysis tasks to the cloud platform's computing nodes, requiring that the tasks can be completed over a relatively long computation duration and can process large amounts of batch data.

[0068] The data flow relationship defined by the parallel interaction parameters is the entire data transmission process from the motor sensor to the edge computing node, and then to the cloud platform computing node. The data collected by the first track needs to be transmitted and processed quickly, while the second track needs to analyze and compute large-scale historical data. The command flow relationship defined by the parallel interaction parameters is the transmission of control commands between the first and second tracks. For example, the first track requires the second track to adjust its analysis model based on real-time monitoring results, or the analysis results of the second track are fed back to the first track to adjust the monitoring strategy.

[0069] To minimize cloud-to-edge communication overhead and computing resource conflicts, the first and second mapping relationships are jointly optimized. The optimization objectives are to minimize cloud-to-edge communication overhead and avoid computing resource conflicts. Minimizing cloud-to-edge communication overhead means reducing data transmission latency as much as possible, ensuring efficient data flow between edge computing nodes and the cloud platform without consuming excessive bandwidth; avoiding computing resource conflicts means rationally allocating computing tasks to avoid resource contention between edge computing nodes and cloud platform computing nodes. For example, if the real-time task in the first track requires more computing resources, adjustments are made automatically to avoid overloading the cloud platform's computing nodes.

[0070] The data and instruction flows between the first and second tracks need to be scheduled based on priority and task type. For example, the analysis results from the second track adjust the monitoring thresholds of the first track, and the real-time data from the first track influences the long-term trend predictions of the second track. Resource allocation is dynamically adjusted based on load and task response time requirements. For instance, when the computation duration of the second track is long, the cloud platform allocates more computing resources to the second track while maintaining the response time requirements of the first track. Data compression and the use of efficient transmission protocols such as MQTT and WebSocket reduce data transmission latency and ensure that real-time data can be quickly transmitted to the cloud platform for further analysis. Through optimization, it is ensured that tasks in both the first and second tracks are completed as expected, meeting response time requirements while avoiding conflicts in computing resources.

[0071] Based on satisfying the first and second constraints, a globally optimized computing power mapping relationship is obtained through collaborative optimization. This ensures that tasks on both tracks can be executed efficiently on appropriate computing resource nodes, computing resources are rationally allocated, and resource conflicts and overload are avoided. In other words, a collaborative optimization mechanism is established between the first and second tracks to ensure that data transmission and computing resources do not conflict between them. Based on real-time data streams and analysis task streams, task allocation strategies are adjusted to ensure that computing resources on the first and second tracks do not interfere with each other. For computing tasks, resource allocation is optimized according to a load balancing algorithm to ensure that response time is not delayed and that computing tasks can be completed within the predetermined time.

[0072] Based on the computing power mapping relationship, the monitoring data of the motor cluster is collected, and the motor status is analyzed through the computing power of the mapping nodes to obtain the motor energy efficiency status evaluation results of each node.

[0073] Specifically, the computing power mapping relationship is the process of allocating the monitoring and computing tasks of the motor cluster to the cloud platform and edge computing nodes according to different constraints. This ensures that tasks can be executed on appropriate nodes and meet the requirements of response time and computing power. Based on the computing power mapping relationship, data acquisition tasks are assigned to edge computing nodes or the cloud platform to guarantee real-time data acquisition and transmission. The motor cluster monitors data; each motor in the cluster continuously collects various monitoring data through sensors, such as current, voltage, load, temperature, and power.

[0074] Based on the computing power mapping relationship, tasks are assigned to different computing nodes. For example, tasks in the first track might be assigned to edge computing nodes to collect data such as current and load in real time and perform preliminary analysis; while tasks in the second track might be assigned to cloud platform computing nodes to perform temperature trend analysis and long-term energy efficiency assessment. Mapping node computing power refers to the computing capability of the computing node to which the motor monitoring task is assigned. Based on the task's computing requirements, monitoring data and analysis tasks are mapped to cloud or edge computing nodes to ensure that the execution of each task meets time and computing requirements. Motor status analysis refers to analyzing the collected monitoring data to evaluate the motor's operating status and energy efficiency. Through motor status analysis, it is possible to identify whether the motor has faults, whether it is in its optimal energy efficiency state, and whether operating parameters need adjustment.

[0075] The power status of motor cluster monitoring data is analyzed to evaluate the energy efficiency of the motors according to set standards and identify any anomalies. For example, a large difference between the input and output power of a motor indicates low energy efficiency or a malfunction. Based on the motor status analysis results, an energy efficiency evaluation result is generated, typically expressed as the energy efficiency ratio (EER), which is the ratio of output power to input power. Output power represents the actual power provided by the motor, while input power represents the electrical energy consumed by the motor. A high EER indicates efficient motor operation; a low EER suggests potential load imbalance, overload, or malfunction; abnormally high power consumption, or excessive temperature or load, may indicate a motor malfunction or the need for maintenance.

[0076] Based on the motor energy efficiency status evaluation results, corresponding optimization measures are proposed. If the load on some motors is too high or too low, the task load is adjusted to optimize the motor's working efficiency. If the motor energy efficiency is low, the control strategy is adjusted, such as adjusting the motor speed, load range, or cooling system. For example, there is a motor cluster containing 100 motors, each with a rated power of 100kW, and data such as current, voltage, load, and temperature are collected in real time. The current data of each motor is sampled every 10 seconds, the load data is sampled every 10 seconds, and the temperature data is sampled every minute. Through the real-time collection of current and load data via the first track, if the current of a motor exceeds 100A and the load exceeds the rated load (100kW), the motor is marked as overloaded. Through the analysis of the motor's temperature and power consumption data via the second track, if the temperature of a motor continuously exceeds 80℃ and the energy efficiency ratio is lower than 0.85, the motor is marked as having low energy efficiency and requires maintenance. One motor has an input power of 95kW and an output power of 90kW, with an efficiency ratio (ERR) of 0.947 (high efficiency); another motor has an input power of 100kW and an output power of 75kW, with an EER of 0.75 (low efficiency). Based on these evaluation results, provide optimization solutions, such as adjusting the load or adjusting the motor's operating range.

[0077] Based on the load data and the motor status of the motor cluster, adaptive equilibrium optimization is performed on the motor energy efficiency status evaluation results to determine the motor cluster control optimization strategy.

[0078] Furthermore, this application also includes the following steps: integrating the high-frequency real-time energy efficiency status data of the first track with the long-term energy efficiency trend evaluation results generated by the second track, performing rolling time-domain prediction of the future energy efficiency status of each appliance in the cluster, and constructing a multi-scale energy efficiency status prediction data chain; setting adjustment coefficients according to the motor status of the motor cluster, based on the multi-scale energy efficiency status prediction data chain, with the goal of maximizing multi-scale time-series cumulative energy efficiency, and using the load data as a constraint, optimizing the operating parameters of each motor based on the adjustment coefficients, obtaining the optimal load allocation sequence of each motor in the multi-scale time domain, and obtaining the motor cluster control optimization strategy.

[0079] Specifically, the high-frequency real-time energy efficiency status data from the first track reflects the motor's immediate operating status and is used to assess the motor's energy efficiency status in real time. The long-term energy efficiency trend evaluation results generated by the second track are used to determine the future energy efficiency change trend of the motor. By fusing the high-frequency real-time energy efficiency status data from the first track with the long-term energy efficiency trend evaluation results from the second track, the real-time data can help capture short-term changes in the motor's status, while the long-term trend analysis can help predict the motor's future operating performance.

[0080] Rolling time-domain forecasting refers to predicting the energy efficiency status of a motor cluster over a future period based on historical and current data. Rolling forecasts are typically dynamic, constantly updating as new data flows in, forming a rolling time series. A multi-scale energy efficiency status forecasting data chain combines short-term and long-term data streams to create a comprehensive forecasting system that reflects both short-term fluctuations and long-term trends. Features are extracted from real-time data on the first track and combined with long-term trend data from the second track to generate rolling time-domain forecasts of future energy efficiency status. This not only describes the future energy efficiency of the motor cluster but also reflects potential changes in motor operation over a future period.

[0081] Rolling time-domain forecasting refers to predicting the state of a motor cluster over a future period based on historical and current real-time data, and continuously updating the forecast results as new data arrives. For motor clusters, rolling time-domain forecasting can help dynamically predict future energy efficiency changes in motors, allowing for timely adjustments to operating strategies. Adjustments are made based on the future energy efficiency status of the motor cluster according to rolling time-domain forecasts. For example, if a forecast indicates that a motor will be overloaded or its efficiency will decrease in the future, the load distribution of that motor can be adjusted in advance; if a motor's temperature is expected to remain high for an extended period, the load can be reduced in advance through the control system to extend the motor's lifespan.

[0082] Based on the health status of each motor, an adjustment factor is set for it. For example, if a motor is in poor health, the adjustment factor is set to a smaller value to reduce its load; if a motor is in good health, the adjustment factor is larger, allowing it to bear more load. By setting the adjustment factor, the operating status of the motors is balanced, extending their lifespan.

[0083] By predicting energy efficiency at multiple scales, the energy efficiency of the motor cluster is maximized across different time scales. Load allocation is optimized based on motor health status and real-time data. Operating parameters are optimized based on constraints such as adjustment coefficients and load data to ensure optimal load allocation for each motor. For example, the load on motors in poor health is reduced to avoid excessive energy consumption or motor damage; the load on motors in good health is increased to improve overall energy efficiency.

[0084] Based on the optimization objective (i.e., maximizing multi-scale time-series cumulative energy efficiency), an optimal load distribution sequence is calculated for each motor to control the operation of the motor cluster, ensuring efficient and healthy motor operation. The motor cluster control optimization strategy not only focuses on short-term energy efficiency but also considers the long-term operational stability and health of the motors. Through reasonable load distribution and optimized control, the lifespan of the motors is extended, and unnecessary energy consumption is reduced. Maximizing multi-scale time-series cumulative energy efficiency refers to maximizing the energy efficiency of the motor cluster at different time scales through multi-scale time series analysis. By optimizing the load distribution of the motor cluster, the overall energy efficiency of the cluster is maximized, and the lifespan of the motors is extended.

[0085] Motor operating parameter optimization refers to finding the optimal operating parameters by adjusting and optimizing the motor's operating state. For example, adjusting the motor's load, speed, and temperature allows the motor to operate under optimal conditions, achieving optimal energy efficiency. Motor cluster control optimization strategies, based on the goal of maximizing multi-scale time-series cumulative energy efficiency, optimize the load distribution and other operating parameters of the motors to maximize energy efficiency and extend motor lifespan.

[0086] In summary, the cloud-based remote monitoring and energy efficiency optimization method for motor clusters provided in this application has the following technical effects:

[0087] By establishing a dual-track parallel monitoring framework on a cloud platform, monitoring target parameters for the first and second tracks, as well as dual-channel parallel interaction parameters, are set respectively. A computing power structure diagram of the cloud platform is constructed, including the cloud-edge topology and the computing power relationships of each topology node. The computing power structure diagram is monitored and allocated using the monitoring target parameters and parallel interaction parameters to establish a computing power mapping relationship. Based on this mapping relationship, monitoring data of the motor cluster is collected, and motor status is analyzed using the computing power of the mapped nodes to obtain the motor energy efficiency status evaluation results for each node. Based on the load data and the motor status of the motor cluster, adaptive equilibrium optimization is performed on the motor energy efficiency status evaluation results to determine the motor cluster control optimization strategy. In other words, by establishing a dual-track parallel monitoring framework on the cloud platform, the motor control strategy is dynamically adjusted according to the real-time operating status and load changes of the motor cluster, maximizing energy efficiency, avoiding energy waste caused by uneven load or improper status, and significantly improving the overall energy efficiency and operating efficiency of the motor cluster.

[0088] Example 2: Based on the same inventive concept as the cloud-based remote monitoring and energy efficiency optimization method for motor clusters in Example 1, this application also provides a cloud-based remote monitoring and energy efficiency optimization system for motor clusters. Please refer to the appendix. Figure 2 The cloud-based motor cluster remote monitoring and energy efficiency optimization system includes:

[0089] The monitoring framework establishment module 11 is used to establish a dual-track parallel monitoring framework on the cloud platform, setting the monitoring target parameters for the first track and the second track, as well as the dual-channel parallel interaction parameters. The computing power structure diagram construction module 12 is used to construct the computing power structure diagram of the cloud platform, including the cloud-edge topology and the computing power relationship of each topology node. The mapping relationship establishment module 13 is used to use the monitoring target parameters and parallel interaction parameters to monitor and allocate the computing power structure diagram and establish a computing power mapping relationship. The motor status analysis module 14 is used to collect motor cluster monitoring data based on the computing power mapping relationship, and perform motor status analysis through the computing power of the mapping nodes to obtain the motor energy efficiency status evaluation results of each node. The balancing optimization module 15 is used to perform adaptive balancing optimization on the motor energy efficiency status evaluation results based on the load data and the motor status of the motor cluster, and determine the motor cluster control optimization strategy.

[0090] Furthermore, the cloud-based motor cluster remote monitoring and energy efficiency optimization system is also used for: establishing a rapid monitoring track and a learning monitoring track with energy efficiency status as the target; setting the monitoring target parameters of the first track and the second track according to the monitoring time period of the rapid monitoring track and the learning monitoring track; establishing an interactive learning relationship of the monitoring target parameters of the first track and the second track based on the gradient relationship of the monitoring time period of the first track and the second track; performing support parameter parsing on the interactive learning process to obtain parallel interactive parameters.

[0091] Furthermore, the cloud-based motor cluster remote monitoring and energy efficiency optimization system is also used for: performing time-cumulative analysis of motor energy efficiency status on historical sample data based on the gradient relationship of the monitoring time periods of the first and second tracks, and obtaining the cumulative impact time-series relationship of the gradient periods; analyzing the guidance parameters of the period nodes of the first track based on the time-series relationship of the second track according to the cumulative impact time-series relationship, and providing optimization feedback for the first track; analyzing the impact coefficient of the first track on the second track based on the cumulative impact time-series relationship, and performing learning prediction for the second track; and generating parallel interaction parameters based on the guidance parameters of the first track and the learning impact coefficient of the second track.

[0092] Furthermore, the cloud-based motor cluster remote monitoring and energy efficiency optimization system is also used for: the guidance parameters of the first track include at least one of the following: dynamic threshold parameters, feature weight parameters, and strategy control parameters, wherein the dynamic threshold parameters are used to replace the original fixed alarm or warning thresholds in the first track; the feature weight parameters are used to adjust the feature importance weights of the energy efficiency status evaluation algorithm in the first track; and the strategy control parameters are used to define or modify the control strategy executed by the first track under a predefined set of operating states.

[0093] Furthermore, the cloud-based motor cluster remote monitoring and energy efficiency optimization system is also used for: taking energy efficiency status as the target event, performing monitoring data association analysis based on motor operation scenario data, and establishing an analytical relationship tree; performing layer-by-layer aggregation based on the analytical relationship tree to obtain an energy efficiency status monitoring relationship set; and matching the collection frequency of the energy efficiency status monitoring relationship set according to the monitoring time period of the first track and the second track to obtain the monitoring target parameters of the first track and the second track.

[0094] Furthermore, the cloud-based motor cluster remote monitoring and energy efficiency optimization system is also used for: based on the parsed relation tree, according to the labeled topological relationship and causal edge, performing layer-by-layer feature aggregation from bottom to top according to the same causal attribute to obtain the causal attribute monitoring influence relationship; according to the influence relationship between the causal attribute and the energy efficiency status evaluation, merging the relationship of each causal attribute monitoring influence relationship, and extracting the monitoring parameters based on the monitoring parameters and correlation relationships involved in the merged relationship to obtain the energy efficiency status monitoring relationship set.

[0095] Furthermore, the cloud-based motor cluster remote monitoring and energy efficiency optimization system is also used for: setting a first track window and a second track window according to the monitoring time period of the first track and the second track, respectively; resampling and time-aligning the monitoring quantities with different sampling periods according to the first track window and the second track window; assigning the monitoring quantities with a sampling period no greater than that of the first track window to the first track, and assigning the monitoring quantities with a sampling period greater than that of the first track window but not greater than that of the second track window to the second track; and selecting and matching monitoring parameters from the energy efficiency status monitoring relationship set as the monitoring target parameters of the corresponding track, based on the periodic sampling monitoring quantity that each track can support.

[0096] Furthermore, the cloud-based motor cluster remote monitoring and energy efficiency optimization system is also used for: analyzing the monitoring time periods of the first track and the second track to determine the corresponding maximum allowable response time and computation duration; based on the computing power structure diagram, with the maximum allowable response time as the first constraint, mapping the monitoring task of the first track to the edge computing power node that meets the maximum end-to-end latency requirement and task computation duration, and establishing a first mapping relationship; with the computation duration as the second constraint, based on the computing power structure diagram, mapping the monitoring and analysis task of the second track to the cloud platform central computing power node with sufficient batch computing resources, and establishing a second mapping relationship; and according to the data flow and instruction flow relationship defined by the parallel interaction parameters, coordinating and optimizing the first mapping relationship and the second mapping relationship to obtain a globally optimized computing power mapping relationship, wherein the coordinating and optimizing adjustment is based on satisfying the first constraint and the second constraint, and aims to minimize cloud-edge communication overhead and computing resource conflicts.

[0097] Furthermore, the cloud-based motor cluster remote monitoring and energy efficiency optimization system is also used to: integrate the high-frequency real-time energy efficiency status data of the first track with the long-term energy efficiency trend evaluation results generated by the second track, perform rolling time-domain prediction of the future energy efficiency status of each appliance in the cluster, and construct a multi-scale energy efficiency status prediction data chain; set adjustment coefficients according to the motor status of the motor cluster, based on the multi-scale energy efficiency status prediction data chain, with the goal of maximizing the multi-scale time-series cumulative energy efficiency, and using the load data as a constraint, optimize the operating parameters of each motor based on the adjustment coefficients, obtain the optimal load allocation sequence of each motor in the multi-scale time domain, and obtain the motor cluster control optimization strategy.

[0098] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Figure 1 The cloud-based motor cluster remote monitoring and energy efficiency optimization method and specific examples in Example 1 are also applicable to the cloud-based motor cluster remote monitoring and energy efficiency optimization system in this example. Through the foregoing detailed description of the cloud-based motor cluster remote monitoring and energy efficiency optimization method, those skilled in the art can clearly understand the cloud-based motor cluster remote monitoring and energy efficiency optimization system in this example. Therefore, for the sake of brevity, it will not be described in detail here.

[0099] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0100] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform, characterized in that, include: Establish a dual-track parallel monitoring framework on the cloud platform, and set the monitoring target parameters for the first track and the second track, as well as the dual-channel parallel interaction parameters respectively; Construct a computing power structure diagram for the cloud platform, including the cloud-edge topology and the computing power relationship between each topology node; The computing power structure diagram is monitored and allocated using the monitoring target parameters and parallel interaction parameters to establish a computing power mapping relationship; Based on the computing power mapping relationship, the monitoring data of the motor cluster is collected, and the motor status is analyzed by the computing power of the mapping node to obtain the motor energy efficiency status evaluation results of each node. Based on the load data and the motor status of the motor cluster, adaptive equilibrium optimization is performed on the motor energy efficiency status evaluation results to determine the motor cluster control optimization strategy. Establish a dual-track parallel monitoring framework on the cloud platform, including: With energy efficiency status as the target, establish a rapid monitoring track and a learning monitoring track, and set the monitoring target parameters of the first track and the second track according to the monitoring time cycle of the rapid monitoring track and the learning monitoring track; Based on the gradient relationship of the monitoring time periods of the first and second tracks, an interactive learning relationship of the monitoring target parameters of the first and second tracks is established, and the interactive learning process is supported by parameter parsing to obtain parallel interactive parameters. Set the monitoring target parameters for the first and second tracks, including: Using energy efficiency status as the target event, monitoring data is correlated and analyzed based on motor operating scenario data to establish a relational tree. Based on the parsed relation tree, a layer-by-layer aggregation is performed to obtain the energy efficiency status monitoring relation set; According to the monitoring time period of the first track and the second track, the energy efficiency status monitoring relationship set is sampled and matched at different frequencies to obtain the monitoring target parameters of the first track and the second track.

2. The method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform according to claim 1, characterized in that, Based on the parsed relation tree, a layer-by-layer aggregation is performed to obtain an energy efficiency status monitoring relation set, including: Based on the parsed relation tree, according to the labeled topological relationship and causal edge, feature aggregation is performed layer by layer from bottom to top according to the same causal attribute to obtain the causal attribute monitoring influence relationship; Based on the influence relationship between the causal attributes and the energy efficiency status assessment, the influence relationships of each causal attribute monitoring are merged, and the monitoring parameters are extracted according to the monitoring parameters and correlations involved in the merged relationships to obtain the energy efficiency status monitoring relationship set.

3. The method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform according to claim 2, characterized in that, According to the monitoring time periods of the first and second tracks, the energy efficiency status monitoring relationship set is sampled at different frequencies to obtain the monitoring target parameters of the first and second tracks, including: Based on the monitoring time periods of the first track and the second track, the first track window and the second track window are set respectively; Based on the first and second track windows, monitoring quantities with different sampling periods are resampled and time-aligned. Monitoring quantities with a sampling period no greater than the first track window are assigned to the first track, and monitoring quantities with a sampling period greater than the first track window but not greater than the second track window are assigned to the second track. Constrained by the periodic sampling monitoring volume that each track can support, matching monitoring parameters are selected from the energy efficiency status monitoring relationship set and used as the monitoring target parameters for the corresponding track.

4. The method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform according to claim 1, characterized in that, Based on the gradient relationship of the monitoring time periods of the first and second tracks, an interactive learning relationship of the monitoring target parameters of the first and second tracks is established. Support parameter parsing is performed on the interactive learning process to obtain parallel interactive parameters, including: Based on the gradient relationship of the monitoring time periods of the first and second tracks, a time-cumulative analysis of the motor energy efficiency status is performed on historical sample data to obtain the time-series relationship of the cumulative impact of the gradient period. Based on the cumulative impact time-series relationship, the guidance parameters of the first track periodic nodes are analyzed based on the time-series relationship of the second track, and the first track is optimized and fed back. Based on the cumulative influence time series relationship, the influence relationship coefficient of the first orbit on the second orbit is analyzed, and the second orbit is learned and predicted. Parallel interaction parameters are generated based on the guidance parameters of the first track and the learning influence coefficients of the second track.

5. The method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform according to claim 4, characterized in that, The guidance parameters for the first track include at least one of the following: dynamic threshold parameters, feature weight parameters, and strategy control parameters, wherein the dynamic threshold parameters are used to replace the original fixed alarm or warning thresholds in the first track. The feature weight parameters are used to adjust the feature importance weights of the energy efficiency state evaluation algorithm in the first track; The strategy control parameters are used to define or modify the control strategy executed by the first track under a predefined set of operating states.

6. The method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform according to claim 1, characterized in that, The computing power structure graph is monitored and allocated using the monitoring target parameters and parallel interaction parameters to establish a computing power mapping relationship, including: Analyze the monitoring time periods of the first and second tracks to determine the corresponding maximum allowable response time and calculation duration; Based on the computing power structure diagram, with the maximum allowable response time as the first constraint, the monitoring task of the first track is mapped to the edge computing power node that meets the maximum end-to-end latency requirement and the task calculation time, and the first mapping relationship is established. Using the computation duration as the second constraint, and based on the computing power structure diagram, the monitoring and analysis tasks of the second track are mapped to the cloud platform central computing power node with sufficient batch computing resources, thus establishing a second mapping relationship; Based on the data flow and instruction flow relationship defined by the parallel interaction parameters, the first mapping relationship and the second mapping relationship are collaboratively optimized and adjusted to obtain a globally optimized computing power mapping relationship. The collaborative optimization and adjustment is based on satisfying the first constraint and the second constraint, with the goal of minimizing cloud-to-edge communication overhead and computing resource conflicts.

7. The method for remote monitoring and energy efficiency optimization of motor clusters based on a cloud platform according to claim 1, characterized in that, Based on load data and the motor status of the motor cluster, adaptive equilibrium optimization is performed on the motor energy efficiency status evaluation results to determine the motor cluster control optimization strategy, including: By integrating the high-frequency real-time energy efficiency status data of the first track with the long-term energy efficiency trend evaluation results generated by the second track, the future energy efficiency status of each appliance in the cluster is predicted in a rolling time domain, and a multi-scale energy efficiency status prediction data chain is constructed. Based on the motor status of the motor cluster, the adjustment coefficient is set. Based on the multi-scale energy efficiency status prediction data chain, with the goal of maximizing the multi-scale time-series cumulative energy efficiency and using the load data as a constraint, the operating parameters of each motor are optimized based on the adjustment coefficient to obtain the optimal load allocation sequence of each motor in the multi-scale time domain, thus obtaining the motor cluster control optimization strategy.

8. A cloud-based remote monitoring and energy efficiency optimization system for motor clusters, characterized in that: The steps for implementing the cloud-based remote monitoring and energy efficiency optimization method for motor clusters according to any one of claims 1 to 7, wherein the cloud-based remote monitoring and energy efficiency optimization system for motor clusters comprises: The monitoring framework establishment module is used to establish a dual-track parallel monitoring framework on the cloud platform, and to set the monitoring target parameters for the first track and the second track, as well as the dual-channel parallel interaction parameters. The computing power structure diagram construction module is used to construct the computing power structure diagram of the cloud platform, including the cloud-edge topology and the computing power relationship of each topology node; The mapping relationship establishment module is used to monitor and allocate the computing power structure diagram using the monitoring target parameters and parallel interaction parameters, and establish a computing power mapping relationship. The motor status analysis module is used to collect motor cluster monitoring data based on the computing power mapping relationship, and to analyze the motor status through the computing power of the mapping nodes to obtain the motor energy efficiency status evaluation results of each node. The equalization optimization module is used to perform adaptive equalization optimization on the motor energy efficiency status evaluation results based on load data and the motor status of the motor cluster, and to determine the motor cluster control optimization strategy.

Images