Dual-motor servo press energy efficiency optimization method based on edge calculation

Abstract

The invention relates to the technical field of motor servo control, and discloses a dual-motor servo press energy efficiency optimization method based on edge calculation. According to the method, operation parameters of a main driving motor and an auxiliary motor are collected in real time through edge calculation nodes deployed on a dual-motor servo press body; dynamically constructing a dual-motor cooperative load state matrix based on the operation parameters, and calculating a real-time comprehensive energy efficiency index by using an energy efficiency evaluation model in an edge calculation node; according to a deviation value between the index and a target energy efficiency threshold value, determining a dynamic energy efficiency adjustment parameter of the double motors; sending an instruction to the controller by the edge computing node, and synchronously updating a power distribution strategy; and monitoring and feedback adjustment are executed circularly until the dual-motor cooperative load state reaches a preset energy efficiency stable interval. According to the method, real-time monitoring and dynamic optimization of the dual-motor cooperative operation state are realized, and the energy efficiency stability of equipment operation is improved.

Description

Technical Field

[0001] The present invention relates to the field of motor servo control technology, and specifically to an energy efficiency optimization method for a dual-motor servo press based on edge computing. Background Art

[0002] As a key piece of equipment in modern industrial production, dual-motor servo presses are widely used in the automotive, aerospace, and precision instrument industries. Their operating energy efficiency directly impacts production costs and environmental performance. With the continuous advancement of industrial automation, the load conditions of dual-motor servo presses are becoming increasingly complex, and the coordinated operation of the main drive motor and auxiliary motor has an increasingly significant impact on overall energy efficiency.

[0003] Currently, traditional energy efficiency optimization methods rely heavily on preset control parameters or offline data analysis, making it difficult to respond to dynamic load changes in real time. For example, during the stamping process, fluctuations in material thickness and hardness can cause instantaneous changes in motor load. Traditional control systems, often based on fixed power allocation strategies, are unable to adapt to dynamic load fluctuations in a timely manner, resulting in energy efficiency losses. When dual motors operate in tandem, the rationality of load distribution between the main drive motor and the auxiliary motor directly impacts overall energy efficiency. An unbalanced load distribution can cause one motor to operate in an inefficient range for extended periods, or even become overloaded. This not only reduces energy efficiency but also shortens the lifespan of the equipment. Furthermore, traditional energy efficiency assessments are often based on the operating parameters of a single motor, lacking a comprehensive consideration of the coordinated state of the dual motors. This makes it difficult to accurately reflect the overall energy efficiency level of the equipment, thus impacting the effectiveness of optimization strategies.

[0004] With the development of Industrial Internet of Things (IIoT) technology, real-time data collection and analysis have become a crucial approach to improving equipment energy efficiency. However, traditional cloud computing models suffer from data transmission delays and high bandwidth usage when processing data with high real-time requirements, making it difficult to meet the real-time dynamic adjustment requirements of dual-motor servo presses. Therefore, achieving dual-motor collaborative energy efficiency assessment and dynamic optimization based on real-time data has become a pressing issue. Summary of the Invention

[0005] The purpose of the present invention is to provide an energy efficiency optimization method for a dual-motor servo press based on edge computing to solve the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention provides an energy efficiency optimization method for a dual-motor servo press based on edge computing, the method comprising: The operating parameters of the main drive motor and auxiliary motor are collected in real time through the edge computing node deployed on the dual-motor servo press body; Dynamically build a dual-motor collaborative load state matrix based on operating parameters, and calculate real-time comprehensive energy efficiency indicators through the energy efficiency evaluation model in the edge computing node; Determine the dynamic energy efficiency adjustment parameters of the main drive motor and the auxiliary motor according to the deviation between the real-time comprehensive energy efficiency index and the target energy efficiency threshold; Send dynamic energy efficiency adjustment parameter instructions to the controller of the dual-motor servo press through the edge computing node, and synchronously update the power allocation strategy of the main drive motor and auxiliary motor; The real-time comprehensive energy efficiency index monitoring and dynamic energy efficiency adjustment parameter feedback adjustment are executed cyclically until the dual-motor collaborative load state reaches the preset energy efficiency stability range.

[0007] Preferably, the operating parameters include the real-time output torque of the main drive motor, the instantaneous speed of the auxiliary motor, the current phase difference when the two motors work together, and the displacement acceleration of the press slide.

[0008] Preferably, the dynamic construction of the dual-motor collaborative load state matrix includes: Extracting the torque fluctuation coefficient of the main drive motor and the speed fluctuation coefficient of the auxiliary motor within a preset time window; Calculate the coupling influence factor between the dual motor current phase difference and the slider displacement acceleration; The torque fluctuation coefficient, speed fluctuation coefficient and coupling influencing factors are integrated into a multidimensional state matrix according to time series.

[0009] Preferably, the calculation process of the real-time comprehensive energy efficiency index includes: A multi-dimensional state matrix is ​​input into the energy efficiency evaluation model, and the load characteristics of the main drive motor and the auxiliary motor are separated through a spatiotemporal feature extraction network. Fusion of load characteristics and preset motor power loss baseline values ​​generates dynamic energy efficiency parameters when dual motors work together; The dynamic energy efficiency parameters are normalized and weighted to output a real-time comprehensive energy efficiency index in the range of 0-1.

[0010] Preferably, the determining of the dynamic energy efficiency adjustment parameters of the main drive motor and the auxiliary motor includes: Establish an energy efficiency change curve with real-time comprehensive energy efficiency index as the vertical axis and timestamp as the horizontal axis; Identify the period in the energy efficiency change curve that is continuously lower than the target energy efficiency threshold, and extract the maximum torque fluctuation coefficient and minimum speed fluctuation coefficient corresponding to the period; Generating a torque compensation amount for the main drive motor according to a proportional relationship between the maximum torque fluctuation coefficient and the standard torque threshold; The speed compensation amount of the auxiliary motor is generated according to the difference between the minimum speed fluctuation coefficient and the reference speed.

[0011] Preferably, the updating of the power allocation strategy includes: Adding the torque compensation amount to the current output torque command value of the main drive motor; Embed the speed compensation into the closed-loop control feedback loop of the auxiliary motor; The edge computing node is used to verify whether the updated phase difference of the dual motor currents meets the preset coordination tolerance range.

[0012] Preferably, the feedback adjustment process includes: After the power allocation strategy is updated, the displacement acceleration of the slider during the update period is recollected; Calculate the rate of change of the variance of the slider displacement acceleration before and after the update; If the variance change rate exceeds a preset threshold, the adjustment range of the torque compensation amount and the speed compensation amount is reduced proportionally.

[0013] Preferably, the determination of the energy efficiency stability range includes: Continuously monitor the real-time comprehensive energy efficiency indicators for three consecutive time windows; When the standard deviation of the real-time comprehensive energy efficiency indicators in the three time windows is less than the preset tolerance value and all are higher than 90% of the target energy efficiency threshold, it is determined to have entered the energy efficiency stability range.

[0014] Preferably, when performing energy efficiency optimization, the edge computing node simultaneously compresses the dynamic energy efficiency adjustment parameters and the corresponding operating parameters into an encrypted data packet, and transmits them to the cloud server through the industrial Internet of Things protocol for long-term performance analysis.

[0015] Preferably, the continuous monitoring process of the energy efficiency stable range includes: When it is detected that the standard deviation of the real-time comprehensive energy efficiency index exceeds the preset tolerance value, the previous data of the current dual-motor cooperative load state matrix is ​​retrieved from the historical database of the edge computing node; Extract the historical records that match the current torque fluctuation coefficient and speed fluctuation coefficient from the previous data, and calculate the mean variance of the slider displacement acceleration corresponding to the historical records; The current slider displacement acceleration variance is compared with the historical variance mean. If the deviation exceeds the threshold, the initialization reset of the dynamic energy efficiency adjustment parameters is triggered.

[0016] Compared with the prior art, the present invention has the following beneficial effects: By deploying edge computing nodes within the dual-motor servo press, real-time acquisition of the operating parameters of the main drive motor and auxiliary motor is achieved. This eliminates the data transmission latency limitations of traditional cloud computing models and enables rapid response to changes in the equipment's operating status. A dynamic dual-motor collaborative load state matrix is ​​constructed based on the real-time acquired operating parameters, incorporating the operating status of both motors into a unified analysis framework. Compared to traditional evaluation methods that focus solely on single motor parameters, this more comprehensively reflects the load distribution of the dual motors when working together, providing a more realistic basis for subsequent energy efficiency evaluations.

[0017] The energy efficiency evaluation model integrated into the edge computing node calculates a real-time comprehensive energy efficiency index based on the collaborative load state matrix. This index comprehensively considers the operating characteristics and collaborative relationship of the dual motors, making the energy efficiency evaluation results more accurate and comprehensive. Dynamic energy efficiency adjustment parameters are determined based on the deviation between the real-time comprehensive energy efficiency index and the target energy efficiency threshold, allowing the adjustment strategy to accurately match the current energy efficiency status of the device, avoiding the blindness of traditional fixed parameter adjustment methods. By sending adjustment parameter commands directly to the controller via the edge computing node, the adjustment commands are quickly transmitted and executed, ensuring the timely dynamic energy efficiency adjustment and enabling immediate intervention in the event of fluctuations in the equipment's operating status, thereby maintaining the high efficiency of the dual-motor operation. This cyclical monitoring and feedback adjustment process forms a closed loop of continuous optimization, ensuring that the coordinated load state of the dual-motor system remains stable within the preset energy efficiency range. This avoids energy efficiency fluctuations caused by external factors or changes in the equipment's internal state, and ensures the long-term energy efficiency stability of the equipment. The dynamically updated power allocation strategy flexibly adjusts the power output of both motors based on their real-time load conditions, ensuring that the main drive motor and auxiliary motor always operate within their respective high-efficiency ranges. This reduces energy efficiency losses caused by improper load distribution and the likelihood of a single motor being in an inefficient or overloaded state for extended periods, helping to extend the life of the equipment. This dynamic optimization model, based on real-time data, can adapt to different load conditions, maintaining high energy efficiency performance in both stable and dynamic load environments, enhancing the adaptability of dual-motor servo presses in complex industrial scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a timing diagram of the energy efficiency optimization method of a dual-motor servo press based on edge computing according to the present invention; Figure 2 Flowchart constructed for dual motor collaborative load state matrix; Figure 3 Flowchart for determining dynamic energy efficiency adjustment parameters; Figure 4 Flowchart adjusted for feedback; Figure 5 Flowchart for continuous monitoring of energy efficiency stability range. DETAILED DESCRIPTION

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] See also Figure 1 The present invention provides an energy efficiency optimization method for a dual-motor servo press based on edge computing, the method comprising: By deploying edge computing nodes on the dual-motor servo press, the operating parameters of the main drive motor and auxiliary motor are collected in real time. A dual-motor collaborative load state matrix is ​​dynamically constructed, and a real-time comprehensive energy efficiency index is calculated based on an energy efficiency evaluation model. Based on the deviation between the real-time comprehensive energy efficiency index and the target energy efficiency threshold, dynamic energy efficiency adjustment parameters for the main drive motor and auxiliary motor are determined. Dynamic energy efficiency adjustment parameter instructions are sent to the controller of the dual-motor servo press via the edge computing node, synchronously updating the power allocation strategy for the main drive motor and auxiliary motor. The real-time comprehensive energy efficiency index monitoring and feedback adjustment of the dynamic energy efficiency adjustment parameters are executed cyclically until the dual-motor collaborative load state reaches the preset energy efficiency stability range.

[0021] Example 1: See Figure 2Optimizing the energy efficiency of a dual-motor servo press relies on the precise collection and analysis of operating parameters. The real-time output torque of the main drive motor is acquired using a high-precision torque sensor. This sensor, integrated into the mechanical connection between the motor output shaft and the driven component, directly measures the torsional stress during transmission. The torque signal is sampled at a frequency of at least one thousand times per second to ensure rapid dynamic changes are captured. The instantaneous speed of the auxiliary motor is measured in real time using a photoelectric encoder mounted on the motor rotor shaft. This high-resolution encoder outputs a differential pulse train. The edge computing node captures the pulse edges using a high-speed counter and, combined with a precise clock reference, calculates the instantaneous angular velocity, achieving measurement accuracy within ±0.01 revolution per minute. The current phase difference when the dual motors operate in coordination requires the simultaneous acquisition of the three-phase stator current signals of the two motors. The current signals are converted into voltage signals suitable for sampling by isolated current transformers and then input into a multi-channel synchronous sampling analog-to-digital converter. The edge computing node performs a fast Fourier transform on the collected current waveform data to accurately calculate the phase angle of the fundamental current, thereby deriving the real-time phase difference between the current signals of the two motors. The press slide's displacement acceleration is acquired using a triaxial piezoelectric accelerometer rigidly connected to the slide body. The sensor's output signal is conditioned by a charge amplifier and then fed into an analog acquisition module. The sampling frequency is strictly synchronized with the motor parameter acquisition to ensure consistent temporal and spatial data correlation.

[0022] The dynamic construction of the dual-motor coordinated load state matrix involves a multidimensional feature extraction and integration process. The system sets a preset time window length, which must cover an integer multiple of the press's typical duty cycle, for example, 200 milliseconds. Within this time window, the torque fluctuation coefficient of the main drive motor is determined by calculation. This process involves statistically analyzing all torque sampling points within the window, calculating their standard deviation, and then dividing this standard deviation by the average torque value within the window. The resulting ratio is defined as the torque fluctuation coefficient. This coefficient objectively reflects the smoothness of the torque output. The speed fluctuation coefficient of the auxiliary motor uses a similar calculation logic. First, the standard deviation of all speed sampling values ​​within the time window is calculated, then divided by the average speed value for the corresponding time window. The resulting ratio is the speed fluctuation coefficient, which represents the stability of speed control. There is an inherent energy coupling relationship between the dual-motor current phase difference and the slider's displacement acceleration. Calculation of this coupling influencing factor requires comprehensive consideration of both sets of parameters. A multivariate regression analysis method is used to establish a mathematical model between the change in current phase difference and the change in slider displacement acceleration. The regression coefficient of the model is the coupling influence factor, which quantitatively describes the combined impact of these two changes on the overall energy efficiency of the system. The three characteristic quantities mentioned above—the torque fluctuation coefficient, the speed fluctuation coefficient, and the coupling influence factor—must be integrated according to a strict time series. This integration method constructs a multidimensional state matrix with fixed dimensions. Each row of the matrix corresponds to a sampling time point, and the number of rows is equal to the total number of sampling points within the preset time window. The matrix contains three columns of data, storing the calculated values ​​of the torque fluctuation coefficient, the speed fluctuation coefficient, and the coupling influence factor corresponding to that time point. This matrix is ​​implemented as a ring buffer in memory and updated in real time.

[0023] The accuracy of operating parameter acquisition directly impacts the reliability of the load state matrix. The torque sensor should be calibrated using a standard weight lever assembly for static calibration and dynamic frequency response testing to ensure that the amplitude error and phase offset within the motor's operating frequency band meet technical specifications. The photoelectric encoder must be installed to ensure strict coaxiality with the motor shaft and eliminate gap errors in the connection device. The encoder signal interface circuit design incorporates anti-interference measures, such as twisted-pair shielded cable transmission and Schmitt trigger shaping at the receiver, to reduce the impact of environmental electromagnetic noise. The core material selection and turns ratio design of the current transformer must balance bandwidth and accuracy, with particular attention to nonlinear distortion characteristics under high dynamic loads. Current signal sampling utilizes a multi-channel synchronous hold analog-to-digital converter with a resolution of at least 16 bits and a conversion speed sufficient to accurately restore the fundamental frequency. Piezoelectric accelerometers should be selected based on the measurement frequency band and range. High-strength bolts should be used for mounting to ensure the sensor's fundamental frequency remains well above the upper limit of the measured signal. The charge amplifier should be equipped with a high-pass filter with an appropriate time constant to eliminate zero drift and an appropriate sensitivity range to accommodate variations in acceleration amplitude under varying operating conditions. All these sensor signals undergo necessary anti-aliasing low-pass filters for front-end conditioning before entering the central processing unit of the edge computing node.

[0024] The edge computing node performs preliminary preprocessing on the collected raw sampled data. Raw torque and speed data may contain sudden interference glitches, so a sliding median filter algorithm is used for smoothing. The filter window width is set to a certain number of sampling points, which must be greater than the typical interference pulse width. Analyzing the spectral characteristics of the slider displacement acceleration signal is crucial. To this end, a fast Fourier transform (FFT) is performed within the edge node to analyze the distribution of the signal's primary frequency components. This information helps to subsequently determine the mechanical resonance points and load characteristics of the equipment. Operating parameters are timestamped using the edge node's built-in real-time clock chip and the operating system's Precision Time Protocol service, ensuring time alignment of multi-source data streams with microsecond accuracy. After preprocessing, key parameters are categorized and stored in database partitions within the edge node. This database is organized in a time-series data table structure, with each parameter allocated independent data channel storage space based on its physical characteristics and update frequency. The database provides an efficient query interface that supports retrieving historical data records by time range or specific operating condition tags. After the data storage period expires, older records exceeding a certain age limit are automatically archived to non-volatile storage media or deleted according to policy to manage storage resources.

[0025] The multidimensional state matrix construction process is dynamic and continuous. The edge computing node's central processing unit continuously schedules a dedicated matrix generation task. This task first loads preprocessed raw data of all relevant parameters within a set time window from the database. This loading operation strictly ensures data temporal integrity. The task then initiates the torque fluctuation coefficient calculation module. This module loads the array of raw torque samples of the main drive motor within the time window, calculates the arithmetic mean of the array elements as a reference, and then calculates the sum of squared deviations of the torque value at each sampling point relative to this mean. The sum of squared deviations is divided by the number of sampling points minus one to obtain the variance. The square root of the variance is taken to obtain the torque standard deviation. Finally, the torque standard deviation is divided by the torque mean to output the torque fluctuation coefficient. The speed fluctuation coefficient calculation uses the same algorithm logic to independently process the auxiliary motor speed data and output the speed fluctuation coefficient value. The coupling influence factor calculation process is more complex: this module loads the synchronously collected current phase difference time series data and the slider displacement acceleration time series data. Covariance analysis is performed on the two data sets to calculate their covariance values. The variance values ​​of the current phase difference data and the slider displacement acceleration data are calculated separately. Finally, the covariance value is divided by the product of the standard deviation of the current phase difference and the standard deviation of the slider displacement acceleration to obtain the Pearson correlation coefficient, which represents the degree of linear correlation between the two parameter sets. This coefficient serves as the core component of the coupling influence factor. Within a single processing task cycle, the system sequentially populates the current row of the multidimensional state matrix with the three values ​​of the torque fluctuation coefficient, speed fluctuation coefficient, and coupling influence factor calculated in real time based on the current time index. The matrix row index increments with time. At the end of the processing task cycle, a complete multidimensional state matrix containing recent load dynamic characteristics is constructed, which can be used by downstream energy efficiency evaluation models. The matrix construction cycle is a system-configurable parameter, typically matching the typical operating rhythm of the press. The matrix data is temporarily stored in a high-speed shared memory area and simultaneously written to a historical state database on a non-volatile storage device for persistent storage.

[0026] State matrix management includes essential lifecycle control. Newly constructed matrices overwrite the oldest historical matrix copies, enabling recycling of storage space. The system generates a unique index identifier and detailed metadata for each matrix construction round, including the construction timestamp, the time window covered, the total number of collected data points, and an overview of eigenvalue statistics. In non-volatile storage, the state matrix is ​​structured and stored as a binary large object (BLO) data block accompanied by metadata records. The edge computing node's external data service interface supports querying and downloading load state matrix data for a specific time period, facilitating offline analysis or remote diagnosis. The matrix data compression algorithm uses a highly efficient lossless compression method, such as dictionary-based compression, to conserve storage space and network bandwidth while maintaining data accuracy. When edge nodes need to store massive amounts of matrix data for a long time, a time-sharing archiving strategy is used to migrate older data to larger storage media. This dynamic construction, efficient storage, and recycling mechanism ensures the data resource management required for long-term stable system operation.

[0027] Example 2: See Figure 3 After the multidimensional state matrix is ​​generated, it enters the real-time comprehensive energy efficiency index calculation process. The core of the energy efficiency evaluation model is the spatiotemporal feature extraction network. This network utilizes a parallel processing architecture. The input layer receives multidimensional state matrix data organized in time series. The first part of the network is a one-dimensional convolutional layer, which is configured with multiple convolution kernels of different scales that slide along the time axis. These convolution kernels capture local variations in the torque fluctuation coefficient, speed fluctuation coefficient, and coupling influencing factors in the state matrix, such as sudden load changes or periodic oscillations within a short period of time. The convolution output is processed by a nonlinear activation function and then fed into a pooling layer. The pooling layer selects the most significant feature responses, retaining their positional information while reducing the data dimension. The second part of the network consists of a long-short-term memory (LSTM) structure. LSTM units receive the feature sequence from the convolutional layer and use their internal gating mechanism to learn the long-term temporal dependencies of state changes. The forget gate determines the degree of retention of previously remembered states, the input gate adjusts the fusion weight of new features, and the output gate controls the release of information from the memory unit to subsequent network layers. The output of the LSTM network carries the temporal evolution characteristics of the load state.

[0028] The network uses a branching structure to separate the load characteristics of the main drive motor and the auxiliary motor. After the long short-term memory layer, the network bifurcates into two independent sub-paths. Each path contains a fully connected layer for deep feature reorganization. Through weight configuration, one path focuses on the load correlation pattern of the main drive motor, while the other path focuses on the load characteristics of the auxiliary motor. Each path outputs a high-dimensional vector containing specific load characteristics. This vector abstractly represents the energy conversion characteristics of the corresponding motor in its current operating state. The fully connected layer uses different weight parameter initialization schemes to encourage the two paths to learn differentiated feature representations.

[0029] These load eigenvectors are fused with preset motor power loss baseline values. The power loss baseline values ​​are derived from the motor design specifications and offline test data and include three components: copper loss, iron loss, and mechanical loss. Copper loss is calculated based on the DC resistance and rated current of the motor winding; iron loss includes eddy current loss and hysteresis loss, determined by the stator core material properties and the operating magnetic flux density; and mechanical loss includes the measured values ​​of bearing friction loss and windage loss. The fusion operation is implemented using feature concatenation, which concatenates the load eigenvector and the three loss component values ​​into a single extended eigenvector. This fusion process enables the model to relate real-time load dynamics to inherent energy consumption characteristics.

[0030] The expanded feature vector is fed into another set of fully connected layers for feature mapping. This layer learns the nonlinear relationship between load dynamics and loss components, outputting a transient power loss parameter. This parameter represents the system's real-time power conversion efficiency under specific load conditions. The model further normalizes this transient parameter to a range between 0 and 1. The normalized reference coordinate system is dynamically constructed based on historical operating data, with the minimum value being the minimum transient power loss detected within a recent window, and the maximum value being the maximum loss within that window. The transient power loss parameter is compressed into this range through a linear transformation. Because the energy consumption contributions of the main drive motor and auxiliary motor vary during different operating phases of the press, the system sets dynamic weighting coefficients for the final synthesis of the comprehensive energy efficiency index. The weighting coefficients are automatically adjusted based on the stroke position signal from the press controller, for example, increasing the weight of the auxiliary motor during the acceleration phase and the main drive motor during the stamping phase. Finally, the real-time comprehensive energy efficiency index is calculated using a weighted average, with values ​​closer to 1 indicating higher overall system energy efficiency.

[0031] The process of determining dynamic energy efficiency adjustment parameters begins with analyzing the energy efficiency change curve. The system continuously records timestamps and corresponding real-time comprehensive energy efficiency index values, maintaining a fixed-length energy efficiency data sequence in memory. This sequence forms a time trajectory chart of energy efficiency changes. The system scans this curve using a sliding window detection algorithm, with the window length set to the duration of a typical abnormal state. The detection logic identifies whether all energy efficiency indicator data points in the window are below the preset target energy efficiency threshold. If three or more consecutive sampling points are detected below the threshold, the system determines that there is an energy efficiency anomaly period and locks that time period for in-depth analysis.

[0032] During this abnormal period, the system performs eigenvalue extraction. From the slice of the multidimensional state matrix corresponding to this period, it retrieves all data points for the main drive motor's torque fluctuation coefficient. By scanning these points, the maximum value is found and recorded as the maximum torque fluctuation coefficient. Similarly, the auxiliary motor's speed fluctuation coefficient data sequence for this period is retrieved. After applying a sliding average filter to eliminate random fluctuation interference, the minimum speed fluctuation coefficient is determined. These extreme values ​​reflect the most significant load imbalance during the abnormal period.

[0033] Determining the torque compensation amount relies on a pre-calibrated relationship model. The system maintains a pre-calibrated torque-energy efficiency response surface data table, which describes the sensitivity of energy efficiency indicators to changes under different torque fluctuation coefficients. The ratio of the current maximum torque fluctuation coefficient to the standard torque threshold is calculated. The standard torque threshold is preset based on the press's rated load and motor characteristics. The response surface is queried to determine the energy efficiency loss ratio corresponding to this ratio, and the required torque compensation amount is derived from this. A positive compensation amount indicates that the torque output ratio of the main drive motor needs to be increased, while a negative compensation amount indicates that the torque output ratio needs to be reduced.

[0034] The speed compensation is calculated using difference processing logic. First, the auxiliary motor's baseline speed value is retrieved from the press process parameter database and bound to the current machining process code. The absolute difference between the detected minimum speed fluctuation coefficient and the baseline speed is calculated. Because the speed fluctuation coefficient is a relative ratio rather than a physical speed value, it must be converted into the corresponding physical speed standard deviation offset. This conversion method multiplies the speed fluctuation coefficient by the auxiliary motor's average speed during the corresponding time window to obtain the actual speed fluctuation amplitude during the abnormal period. This fluctuation amplitude serves as the compensation reference value. Its application direction is determined by the auxiliary motor's power characteristics: if the auxiliary motor is underloaded, positive compensation is applied to increase the speed; if it is overloaded, negative compensation is applied to reduce the speed demand. The resulting dynamic energy efficiency adjustment parameters include the torque compensation value for the main drive motor and the speed compensation value for the auxiliary motor, along with a timestamp. These parameters are encapsulated as data instruction packets and temporarily stored in the edge node's send buffer. The entire calculation process is completed before the next multidimensional state matrix is ​​generated to meet the system's real-time response requirements.

[0035] Example 3: See Figure 4The transmission of dynamic energy efficiency adjustment parameter instructions adopts the real-time communication interface between the edge computing node and the press controller. This interface is based on the industrial Ethernet protocol, and the transmission cycle is synchronized with the scan cycle of the press control system. The edge node writes the encapsulated adjustment instructions into the specified data area of ​​the shared memory, and the controller reads the data area at the beginning of each control cycle. The instruction data packet contains three core fields: the torque compensation value of the main drive motor, the speed compensation value of the auxiliary motor, and the instruction effective timestamp. The timestamp is used to ensure that the controller is strictly synchronized with other motion control signals when executing instructions. The instruction transmission process implements a hardware-level verification mechanism, such as a cyclic redundancy check code attached to the end of the data packet to prevent data tampering or loss during transmission.

[0036] The application of torque compensation to the main drive motor occurs in two stages. The first stage involves dynamic limiting. The controller obtains the maximum allowable output torque specified by the motor manufacturer and compares the absolute value of the torque compensation with this limit. If the absolute value of the compensation exceeds 50% of the limit, the system reassigns the compensation in the direction of the sign of the 50% limit. The second stage involves torque command reconstruction. The controller extracts the original torque command value of the main drive motor for the current operating cycle and linearly superimposes the limited torque compensation value onto this command value. This superposition process is not a simple arithmetic addition, but rather an interpolation transition based on the press's real-time operating phase. During the press slide's ascending phase, the compensation value is accumulated in five steps over 200 milliseconds. During the slide's descending phase, a three-step accumulation strategy is used to avoid mechanical shock caused by sudden torque changes. The reconstructed torque command value is fed to the motor driver's current loop input port.

[0037] The speed compensation acts on the closed-loop control circuit of the auxiliary motor. The controller's speed regulation module includes a proportional-integral-derivative control link. The compensation is input into the module's set value processing channel. In the standard process, the set value processing channel first performs a ramp transition on the speed setting reference value to generate a smooth command, and then passes the speed compensation through a first-order low-pass filter and superimposes it on the command. The cutoff frequency of the low-pass filter is automatically adjusted according to the mechanical time constant of the auxiliary motor rotor to ensure that the introduction of the compensation does not excite mechanical resonance. The superimposed integrated speed command is input into the input of the proportional-integral-derivative controller. The output of the proportional-integral-derivative controller serves as the set value of the current loop, ultimately driving the auxiliary motor to perform the speed regulation task.

[0038] After the power allocation strategy is updated, the dual-motor coordination performance needs to be verified. The edge computing node re-collects the three-phase current signals of the two motors during the update period. The current signal sampling window length is set to an integer multiple of the press's operating cycle. A fast Fourier transform is performed on each of the two current signals to extract the fundamental component and calculate the phase angle. The phase angle difference is calculated as follows: ; in: is the average phase difference angle after update (degrees), is the number of sampling points, is the sampling time point, and Respectively represent The current phase angles of the main drive motor and auxiliary motor at each sampling point are calculated. This calculated result is compared with a preset coordination tolerance range. This tolerance range is dynamically adjusted based on the press's operating phase: the upper limit is relaxed to 8 degrees during the slide acceleration phase and tightened to 3 degrees during the forming and pressurization phase. If the average phase difference exceeds the tolerance range for the current phase, the system flags a coordination anomaly.

[0039] The feedback adjustment mechanism is automatically activated after the power policy is updated. The edge nodes synchronously collect the slider's displacement acceleration signals before and after the update. The raw acceleration sensor data is normalized to eliminate the influence of different time scales. The ratio of the slider's displacement acceleration variance before and after the update is calculated as follows: System-defined variance change rate parameter ,in Represents the variance of the slider displacement acceleration after the strategy is updated, Represents the original variance value before the update. The preset variance change threshold The value is 0.25.

[0040] When monitoring When the compensation adjustment amplitude is reduced, the compensation adjustment amplitude reduction procedure is triggered. This reduction operation uses a step-by-step iterative strategy: first, the absolute value of the current torque compensation is multiplied by a reduction factor of 0.7, and the absolute value of the speed compensation is multiplied by a reduction factor of 0.6, while maintaining the sign of the original compensation direction. The new reduced compensation value is re-input into the torque command reconstruction and speed closed-loop control loops. The trigger delay of the coordinated tolerance verification step is also adjusted to 400 milliseconds to allow the mechanical system sufficient time to respond to the adjustment.

[0041] The system establishes a closed loop for evaluating the adjustment results. After each reduction in compensation, two sets of data are collected again: the updated current phase difference information and the slider displacement acceleration signal. The phase difference data is substituted into the collaborative verification formula to recalculate If the value is still outside the tolerance range, the secondary reduction process is started. The secondary reduction uses a stronger reduction coefficient, the torque compensation coefficient is reduced to 0.4, and the speed compensation coefficient is reduced to 0.3. The time domain analysis of the slider displacement acceleration variance is performed to detect whether there is still a high-frequency vibration component. After two consecutive reduction operations If the value is still greater than the threshold, it is determined to be an abnormal operating condition, and the system terminates the automatic adjustment process and issues a signal for equipment maintenance.

[0042] The edge computing node records the complete adjustment process data. This record includes the initial adjustment parameter values, each reduction ratio, phase difference verification results, variance change rate values, and the final adjustment effect code. Data is stored in binary format, and each adjustment event generates a separate data file. The file header contains information such as a timestamp, press model code, and workpiece batch number. These files are periodically uploaded to the central analysis system for long-term operation mode optimization research. In the local storage system, detailed data from the latest three adjustment cycles is stored in high-speed solid-state memory for rapid retrospective analysis. When the local storage space reaches 90% of its capacity, the system automatically clears the oldest 20% of data records to free up space.

[0043] Continuous monitoring of dynamic parameters is performed in the background. The edge node maintains a circular monitoring buffer that continuously tracks the real-time values ​​of three key parameters: the updated real-time comprehensive energy efficiency index, the running average of the current slider displacement acceleration variance, and the mean phase difference of the dual motor currents. The sampling interval for these parameters is 50 milliseconds. The monitoring algorithm calculates the moving standard deviation of the three parameters. When the standard deviation of any parameter exceeds its corresponding threshold for three consecutive sampling periods, the system automatically loads the recorded data of the most recent successful adjustment, resets the torque compensation and speed compensation to the historical values, and backs up the current adjustment parameters before the reset operation. After the reset is complete, the collaborative performance verification process is restarted to form an autonomous recovery mechanism. All reset events generate a detailed operation log, including the reset reason analysis code and the historical parameter matching index.

[0044] Example 4: See Figure 5 The determination of the energy efficiency stability range is performed through a scheduled task on the edge computing node. The system configures a monitoring sequence of three consecutive time windows, with each window length set to twice the standard operating cycle of the press. Window advancement uses a progressive overlapping method, with the start point of each new window separated from the end point of the previous window by one-third of the window length. When the task is initiated, three sets of real-time comprehensive energy efficiency indicator raw data for the current time window and the previous two time windows are loaded. Each set contains the energy efficiency indicator values ​​for all sampling points within that window, with a sampling frequency of 100 times per second. After data loading, preprocessing is performed: a moving average filter is used to smooth random fluctuations, with the filter window width set to 5% of the sampling points. The preprocessed data enters the standard deviation calculation module, which uses an unbiased estimation algorithm to calculate the standard deviation of the energy efficiency indicator values ​​within each window. This calculation first calculates the arithmetic mean of all energy efficiency values ​​within the window, then calculates the squared deviation of each energy efficiency value from this mean. The sum of the squared deviations is divided by the number of data points minus one, and the square root is taken to obtain the standard deviation for that window.

[0045] The dynamic maintenance mechanism of the target energy efficiency threshold runs synchronously. The system associates the press process database and retrieves the historical optimal energy efficiency record based on the material code and thickness parameters of the currently processed workpiece. If the historical database contains more than ten valid records under the same process parameters, eighty-five percent of the highest energy efficiency index value in these records is taken as the temporary threshold; if there is no matching record, the preset default threshold is used. The temporary threshold is combined with the standard deviation calculated from the current window for analysis: when the standard deviation of three consecutive windows is less than five percent of the target energy efficiency threshold, and the arithmetic mean of the energy efficiency index in each window is higher than ninety percent of the threshold value, the system generates an energy efficiency stable state identifier. This identifier is written to the control status register and triggers the control strategy switch at the same time - converting the dynamic adjustment mode to the stable maintenance mode.

[0046] The continuous monitoring process remains active in a stable state. The monitoring frequency is adjusted to four times per minute. The standard deviation of the latest three-window energy efficiency index is recalculated during each monitoring. When it is detected that the standard deviation of any window exceeds 5.5% of the target energy efficiency threshold, the system immediately initiates the abnormal response protocol. The first step of the protocol is to retrieve the dual-motor collaborative load matrix record related to the previous state in the historical database. The search conditions are set as follows: the time range is the last 72 hours, the torque fluctuation coefficient error does not exceed 10% of the current value, and the speed fluctuation coefficient error is controlled within 8% of the current value. The search results list all matching records in reverse chronological order.

[0047] The historical data analysis module processes the search results. For each matching record, the associated slider displacement acceleration data is extracted. This data is time-labeled in the original storage. Acceleration data is grouped: Records with the same process parameters are grouped together. Within each group, the acceleration variance of all records within the equivalent operating time period is calculated, and the weighted average of the group variance values ​​is calculated. Weights are assigned using a time-decay strategy, with the current weight being one and decreasing by 0.05 for each hour preceding the time. See Table 1 for the search results of the data matching process.

[0048] Table 1: Retrieval results of the data matching process

[0049] The current working condition is automatically grouped based on process parameters. The system selects a subset of historical records whose process parameters fully match the current machining task. Within this subset, a quadratic weighted average is calculated for the acceleration variance mean column: the weight is based on the inverse of the time interval between the historical record generation time and the current time interval, with the weight reduced to 0.5 for records with an interval exceeding six hours. The calculated weighted variance mean serves as the historical baseline reference value.

[0050] The abnormality decision module synchronously collects real-time data on the current slider's displacement acceleration. The current acceleration variance is calculated using a window of the same length. The absolute difference between this value and the previously mentioned historical baseline reference value and divided by the baseline reference value yields the deviation rate. When the deviation rate exceeds the 15% threshold, a reset instruction code is written to the control register. The reset operation first backs up all current dynamic energy efficiency adjustment parameters to an isolated storage area. Then, based on matching historical records, the adjustment parameter combination corresponding to the optimal energy efficiency moment is located. If multiple high-efficiency points exist in the matching record, the parameter set with the smallest acceleration variance is prioritized. The parameter initialization process is performed in steps: the torque compensation value is transitioned to the historical value within 0.5 seconds using a three-segment ramp function, and the speed compensation value is gradually transitioned to the target value within 0.8 seconds using a five-segment broken line interpolation. The reset signal also triggers the activation of the energy metering module, which records the power change waveform before and after the initialization operation.

[0051] After parameter initialization, system behavior enters a special monitoring cycle. During the first five operating cycles, energy efficiency metrics are calculated at double the frequency, and the window length for constructing the standard collaborative load state matrix is ​​shortened to 60 percent of its normal value. Real-time comprehensive energy efficiency metric data is marked as being in the initialization recovery state. Its determination results are used only for local policy control and are not included in the long-term database quality metric statistics. Each initialization event is fully documented in the event log: including the trigger deviation rate value, the historical record identifier used, and snapshots of key parameters before and after initialization. Log entries are accompanied by 13-digit timestamps and the device hardware serial number encoding.

[0052] The local storage system implements lifecycle management for historical load matrices. The currently generated load matrix is ​​marked as prioritized for protection, retaining at least twelve hours of raw data. Matrix data exceeding this limit is losslessly compressed and transferred to a secondary storage area. The compressed package header contains a matrix dimension descriptor and eigenvalue statistical summary. The system performs storage space maintenance tasks daily at dawn: normal matrix data older than seven days is deleted, but all matrix records marked as initialization reference points are retained. These critical records are permanently stored in their original data format in non-volatile memory and simultaneously uploaded to the cloud-based disaster recovery server for dual storage protection.

[0053] Example 5: When executing core energy efficiency optimization tasks, edge computing nodes simultaneously initiate the data archiving and transmission process. Dynamic energy efficiency adjustment parameters and operating parameters for the corresponding acquisition cycle form a structured data packet. The torque compensation values ​​for the main drive motor and the speed compensation values ​​for the auxiliary motor are stored in four-byte floating-point format. The operating parameter set includes the torque sensor raw sample array, the dual current phase difference sequence, and the accelerometer raw data stream for the adjustment cycle. Each data type is accompanied by a microsecond-accurate timestamp. Data packet organization utilizes a hierarchical encapsulation architecture: the base layer is a binary data block, the metadata layer contains sensor calibration coefficients and acquisition frequency information, and the application layer records process parameter codes and equipment serial numbers. After encapsulation, the compression phase begins. The compression engine uses a dictionary-based encoding algorithm to create a dynamic symbol table for recurring patterns. The encoding process prioritizes large floating-point arrays and identifies intervals within consecutive sample values ​​that can be represented by run-length encoding. The maximum size of the compressed data block is limited to 75 percent of the storage buffer capacity.

[0054] Data encryption is performed immediately after compression. The encryption module loads the master key stored in the hardware secure enclave and generates a session encryption key using a key derivation function. An authenticated encryption mode is used, which simultaneously generates ciphertext and an integrity check tag. The encryption operation operates on 128-bit data blocks, with each block undergoing sixteen rounds of substitution-permutation. In each round, logical operations are performed on the data block and the round key, altering data characteristics through nonlinear byte permutations, scrambling row and column shifts, and finally performing a column-by-column permutation. The complete ciphertext output is appended with an eight-byte message authentication code. A version identifier and initialization vector are added to the encrypted packet header, and a cyclic redundancy check sequence is embedded in the tail. The entire encryption process is performed within the edge node's cryptographic coprocessor, with the key material only ephemeral in the secure enclave memory.

[0055] The Industrial IoT transmission protocol utilizes a standardized communication framework. The protocol stack includes transport layer security session protection, and a bidirectional digital certificate authentication mechanism verifies the legitimate identity of cloud servers and edge nodes. The physical link supports dual channels for wired Industrial Ethernet and cellular networks, with the transmission manager dynamically selecting the primary channel based on current network quality indicators. The data fragmentation mechanism breaks large data packets into fragments that meet the maximum transmission unit standard, with each fragment appended with a fragment sequence number and a total packet identifier. The transmission controller maintains a dynamic fragment size table and automatically reduces the fragment size if it detects the loss of three consecutive fragments. The basic transmission cycle is set to a fixed interval of five minutes or an integer multiple of the press completing five work cycles, whichever comes first. Each transmission event triggers the creation of a transmission log in the local storage system, recording the packet hash value, destination server address, and actual transmission time.

[0056] The cloud server receiver is configured with a load balancing mechanism. The ingress gateway parses the packet header and verifies the validity of the message authentication code. If the integrity check fails, the gateway immediately sends a retransmission request to the edge node. If more than three failures occur within three minutes, the node's transmission channel is frozen. The decryption engine uses the corresponding key in the cloud key vault to reverse the decryption process and restore the original packet. The decompression service uses a stream processing mode to restore the original data structure by reconstructing the encoding dictionary. The decompressed data packet is split into two logical parts: the dynamic energy efficiency adjustment parameters are imported into the short-term analysis cache, and the complete operating parameter sequence is written to the time series database.

[0057] The time series database adopts a clustered deployment architecture. Data partitions are distributed to different storage nodes based on the hash value of the first three characters of the device serial number. Each record stores the following elements: an arrival time tag accurate to the millisecond, unparsed binary payload data, and a data packet hash fingerprint. The database automatically creates a multi-level index: the primary index is organized according to the device serial number and time range, and the secondary index marks the process parameter code and the key characteristic value range. The data retention policy sets a two-tier storage mechanism, with recent data stored in the solid-state drive array and historical data older than thirty days transferred to cold storage media. After the compressed parameter data is imported, the label extraction operation is triggered to identify the torque compensation and speed compensation values ​​in the data packet, and establish an association mapping with the metadata in the device registry.

[0058] The long-term performance analysis module initiates an asynchronous data processing pipeline. The analysis engine scans newly added data records every 20 minutes. The data processing phase isolates three sets of features: extracting the time series of torque compensation over three consecutive transmission cycles and calculating its mean and standard deviation; analyzing the gradient of the auxiliary motor speed compensation; and correlating it with maintenance status codes from the equipment operation log. This feature set is fed into the performance evolution evaluation algorithm, which incorporates a sliding time window comparison mechanism. The window width is initialized to 12 hours, sliding by four hours each time. Under the same process parameters, the difference in compensation mean values ​​between the current window and the historical window is compared. If the difference exceeds a preset threshold, a compensation deviation alarm event is generated. The seasonal decomposition algorithm partitions the annual data by quarter and detects whether the compensation parameters exhibit periodic fluctuations related to temperature. The anomaly pattern identifier uses an isolation forest algorithm to identify anomalous mutation points in the compensation change trajectory.

[0059] A separate feedback channel for cloud-based analysis results is established. The feedback data packet contains a unique edge node identifier, a time window locator, and parameter adjustment recommendations. These recommendations may include corrections to the weight coefficients of the energy efficiency evaluation model and updated values ​​for the scaling factors used in the compensation calculation. The transmission controller prioritizes the edge node's currently idle data channel and switches to a lower-rate control channel if the encrypted transmission channel is unavailable. The edge node's receiving end is equipped with a dedicated message parser to isolate the valid command fields in the feedback packet. Parameter updates are performed during device idle time. Weight coefficients for the dynamic energy efficiency evaluation model are loaded using a checksum protection mechanism, with checksums calculated twice before and after the update file is transferred. The compensation scaling factor is updated using a double buffer, with the previous version of the configuration temporarily retained as a fallback recovery point. After each successful update, the node sends an application confirmation signal to the cloud, carrying a snapshot of the parameters before the update and the initial state parameters after execution. Unconfirmed feedback commands are automatically resent after 18 hours and are marked as invalid after five consecutive failures.

[0060] The archiving system establishes a closed-loop audit tracking mechanism. At the beginning of the transmission, the edge node generates a transmission event serial number, which is a composite code of the year, month, day, hour, minute, second, and the device serial number. This serial number is transmitted along with the data packet to the cloud database for storage. At each key stage of data processing—after compression, encryption, before transmission, and after cloud decryption—the system records the processor identifier and processing timestamp of the data packet corresponding to the serial number. The tracking log of the entire transmission link is independently stored in the cloud audit database. The complete life cycle trajectory of each transmission link can be reconstructed using the serial number as an index. The edge node locally retains detailed transmission logs for the last 72 hours, and overdue logs are compressed with hash digests and uploaded for archiving.

[0061] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0062] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A dual-motor servo press energy efficiency optimization method based on edge computing, characterized in that: include: The operating parameters of the main drive motor and auxiliary motor are collected in real time through the edge computing node deployed on the dual-motor servo press body; Dynamically build a dual-motor collaborative load state matrix based on operating parameters, and calculate real-time comprehensive energy efficiency indicators through the energy efficiency evaluation model in the edge computing node; Determine the dynamic energy efficiency adjustment parameters of the main drive motor and the auxiliary motor according to the deviation between the real-time comprehensive energy efficiency index and the target energy efficiency threshold; Send dynamic energy efficiency adjustment parameter instructions to the controller of the dual-motor servo press through the edge computing node, and synchronously update the power allocation strategy of the main drive motor and auxiliary motor; The real-time comprehensive energy efficiency index monitoring and dynamic energy efficiency adjustment parameter feedback adjustment are executed cyclically until the dual-motor collaborative load state reaches the preset energy efficiency stability range.

2. The energy efficiency optimization method of a dual-motor servo press based on edge computing according to claim 1 is characterized in that: The operating parameters include the real-time output torque of the main drive motor, the instantaneous speed of the auxiliary motor, the current phase difference when the two motors work together, and the displacement acceleration of the press slide.

3. The energy efficiency optimization method of a dual-motor servo press based on edge computing according to claim 2 is characterized in that: The dynamic construction of the dual-motor collaborative load state matrix includes: Extracting the torque fluctuation coefficient of the main drive motor and the speed fluctuation coefficient of the auxiliary motor within a preset time window; Calculate the coupling influence factor between the dual motor current phase difference and the slider displacement acceleration; The torque fluctuation coefficient, speed fluctuation coefficient and coupling influencing factors are integrated into a multidimensional state matrix according to time series.

4. The energy efficiency optimization method of a dual-motor servo press based on edge computing according to claim 3 is characterized in that: The calculation process of the real-time comprehensive energy efficiency index includes: A multi-dimensional state matrix is ​​input into the energy efficiency evaluation model, and the load characteristics of the main drive motor and the auxiliary motor are separated through a spatiotemporal feature extraction network. Fusion of load characteristics and preset motor power loss baseline values ​​generates dynamic energy efficiency parameters when dual motors work together; The dynamic energy efficiency parameters are normalized and weighted to output a real-time comprehensive energy efficiency index in the range of 0-1.

5. The energy efficiency optimization method of a dual-motor servo press based on edge computing according to claim 4 is characterized in that: Determining the dynamic energy efficiency adjustment parameters of the main drive motor and the auxiliary motor includes: Establish an energy efficiency change curve with real-time comprehensive energy efficiency index as the vertical axis and timestamp as the horizontal axis; Identify the period in the energy efficiency change curve that is continuously lower than the target energy efficiency threshold, and extract the maximum torque fluctuation coefficient and minimum speed fluctuation coefficient corresponding to the period; Generating a torque compensation amount for the main drive motor according to a proportional relationship between the maximum torque fluctuation coefficient and the standard torque threshold; The speed compensation amount of the auxiliary motor is generated according to the difference between the minimum speed fluctuation coefficient and the reference speed.

6. The method for optimizing energy efficiency of a dual-motor servo press based on edge computing according to claim 5 is characterized in that: The updating of the power allocation strategy includes: Adding the torque compensation amount to the current output torque command value of the main drive motor; Embed the speed compensation into the closed-loop control feedback loop of the auxiliary motor; The edge computing node is used to verify whether the updated phase difference of the dual motor currents meets the preset coordination tolerance range.

7. The method for optimizing energy efficiency of a dual-motor servo press based on edge computing according to claim 6 is characterized in that: The feedback adjustment process includes: After the power allocation strategy is updated, the displacement acceleration of the slider during the update period is recollected; Calculate the rate of change of the variance of the slider displacement acceleration before and after the update; If the variance change rate exceeds a preset threshold, the adjustment range of the torque compensation amount and the speed compensation amount is reduced proportionally.

8. The method for optimizing energy efficiency of a dual-motor servo press based on edge computing according to claim 7 is characterized in that: The determination of the energy efficiency stability range includes: Continuously monitor the real-time comprehensive energy efficiency indicators for three consecutive time windows; When the standard deviation of the real-time comprehensive energy efficiency indicators in the three time windows is less than the preset tolerance value and all are higher than 90% of the target energy efficiency threshold, it is determined to have entered the energy efficiency stability range.

9. The method for optimizing energy efficiency of a dual-motor servo press based on edge computing according to claim 1, characterized in that: When performing energy efficiency optimization, the edge computing node simultaneously compresses the dynamic energy efficiency adjustment parameters and the corresponding operating parameters into encrypted data packets, and transmits them to the cloud server through the industrial Internet of Things protocol for long-term performance analysis.

10. The energy efficiency optimization method of a dual-motor servo press based on edge computing according to claim 8 is characterized in that: The continuous monitoring process of the energy efficiency stability range includes: When it is detected that the standard deviation of the real-time comprehensive energy efficiency index exceeds the preset tolerance value, the previous data of the current dual-motor cooperative load state matrix is ​​retrieved from the historical database of the edge computing node; Extract the historical records that match the current torque fluctuation coefficient and speed fluctuation coefficient from the previous data, and calculate the mean variance of the slider displacement acceleration corresponding to the historical records; The current slider displacement acceleration variance is compared with the historical variance mean. If the deviation exceeds the threshold, the initialization reset of the dynamic energy efficiency adjustment parameters is triggered.

Images