Energy-optimized failure machine power management method

Abstract

The application discloses a fault machine tool power management method based on energy optimization, and relates to the technical field of machine tool power management.The steps of the method comprise the following steps: three-axis vibration sensors and current and voltage sensors are simultaneously installed on the spindle and motor drive side of each machine tool to collect vibration signals and electrical signals of the machine tool in real time; a fault alarm strategy is set; after the fault is triggered, a load redistribution process is started; a multi-criteria decision model is established to dynamically allocate priority weights to each machine tool, and a load distribution algorithm is established to distribute power to each machine tool; and the motor drive voltage and frequency are dynamically adjusted according to the actual load distribution.The application solves the problem of constant voltage and frequency ratio method which only adjusts voltage and frequency according to single machine load, cannot automatically reduce total power supply output when a machine tool is stopped, and leads to the problems of energy waste in no-load operation and machine tool level energy consumption prediction.

Description

Technical Field

[0001] This invention relates to the field of machine tool power management technology, specifically to a fault-based machine tool power management method based on energy optimization. Background Technology

[0002] Traditional constant voltage-frequency ratio methods only adjust voltage and frequency based on the load of a single machine, and cannot automatically reduce the total power output when a machine tool is stopped, resulting in wasted energy during idle operation and machine tool-level energy consumption prediction. Existing research has used machine learning to predict the energy consumption curve of a single CNC machine tool in order to optimize cutting parameters, but it has not incorporated load fluctuations and fault shutdowns into the overall power management and lacks a system-level coordination mechanism. Although fault diagnosis technologies based on autoencoders or deep learning can achieve online monitoring and location, they are usually only used as a reference for maintenance decisions and do not provide an interface for energy consumption scheduling after a fault occurs. While existing technologies have achieved great results in single-machine energy-saving control, fault detection and energy consumption prediction, they lack a solution for "multi-machine system-level energy redistribution when a single machine fails and stops," and have key problems such as response lag, resource waste and lack of multi-objective collaborative optimization. Therefore, this invention provides a fault-based machine tool power management method based on energy optimization. Summary of the Invention

[0003] The purpose of this invention is to provide a fault-based machine tool power management method based on energy optimization, so as to solve the existing problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a fault-based machine tool power management method based on energy optimization, comprising the following steps: S1. Install a three-axis vibration sensor and a current and voltage sensor on the spindle and motor drive side of each machine tool to collect the vibration signal and electrical signal of the machine tool in real time. S2. Set up a fault alarm policy. After a fault is triggered, start the load redistribution process. S3. Establish a multi-criteria decision-making model to dynamically assign priority weights to each machine tool, and establish a load allocation algorithm to allocate power to each machine tool; S4. Dynamically adjust the motor drive voltage and frequency according to the actual load distribution.

[0005] A further improvement of this invention lies in that the fault alarm strategy constructs an encoder by batch normalizing the output of the preprocessing step S1 through continuous convolutional layers and then downsampling it to the latent spatial dimension using an activation function. The symmetric transposed convolutional layer is batch normalized and then reconstructed using an activation function to obtain an image with the same dimension as the original image. Training was performed using fault-free time-frequency images, followed by training with the Adam optimizer for 50-100 rounds until the reconstruction error on the validation set converged; this was then applied to newly acquired time-frequency images. Calculate the reconstructed new time-frequency graph Calculate the reconstruction error ; Reconstruction error of health dataset And calculate the average health data. with standard deviation The fault alarm threshold is k represents the set variation coefficient. When the reconstruction error is greater than the fault alarm threshold, it is judged as a machine fault and a fault alarm is triggered. Subsequently, the PLC receives the shutdown signal, marks the machine tool status as stopped, and synchronizes the status update to the SCADA / IoT platform through the industrial Ethernet to start the load redistribution process.

[0006] A further improvement of this invention is that the multi-criteria decision-making model includes a hierarchical sub-model, which comprises an objective layer, a criterion layer, and a solution layer. The objective layer is used to prioritize the power supply to machine tools, the criterion layer is used to obtain process criticality, safety risks, and delivery timeliness, and the solution layer is used to create a machine tool dataset. , Let S represent the nth machine tool, where n represents the total number of machine tools; perform eigenvalue decomposition of the covariance matrix on the index data to obtain the eigenvalues ​​of the covariance matrix S. With feature vectors Select the top M principal components, and use the loadings of each criterion layer index on the top M principal components as the contribution; then the loading of criterion layer index j on the kth principal component. Then the weight of index j is expressed as , where c represents the number of principal components with a cumulative contribution rate of 85%, and m represents the number of criteria layer indicators.

[0007] A further improvement of this invention is that the multi-criteria decision-making model further includes a factor set composed of the indicators of the criterion layer. And their corresponding weights are sent into the weight set. And set evaluation levels. The membership degree of each factor j to each level i is calculated using the trapezoidal membership function. The membership matrix is ​​obtained. ; Perform a weighted average of the weight set and the membership matrix. , , obtained the Table machine tool rating The rating set GR for each machine tool is obtained by sorting them from largest to smallest.

[0008] A further improvement of this invention is that the load distribution algorithm includes defining the optimization objective as the first... Minimize the total energy consumption of the machine tool and set the constraint that the sum of the squares of the machine tool's energy consumption is less than or equal to the set maximum energy consumption.

[0009] A further improvement of this invention is that the load allocation algorithm further includes power allocation command control, specifically including allocating the power load of each machine tool through an MPC model. Load target Using the optimization objective as input to the algorithm, a control sequence with the total time domain length of future time control is obtained. and execute the first control input. The prediction cycle restarts at the next time step; the prediction step size is set. With control step size At each control time, g solves for the optimization objective. , Indicates the first The state vector for each prediction step; then new measurements are acquired to update the prediction.

[0010] A further improvement of this invention is that the allocation instruction control also includes incorporating machine tool failures and load fluctuations into the algorithm parameters, monitoring the real-time power request of the faulty machine tool, and when the power demand of the faulty machine tool decreases beyond a set threshold within a set short time t, the system allocates the excess power to other machine tools in real time. Specifically, this is manifested in extracting the change in the power request of the faulty machine tool. The proportion of each machine tool's power allocation command in the total power is extracted as the allocation weight of each machine tool. This weight is then used to calculate the allocation weight of each machine tool. The power distribution command control adjustment value is obtained by multiplying the power distribution command and the power distribution command control adjustment value. Based on the power distribution command control, the power of each machine tool is adjusted again.

[0011] A further improvement of the present invention is that step S4 includes installing a voltage sensor in the power supply circuit of each machine tool motor to collect the voltage of each machine tool in real time and setting the rated frequency of the motor. With rated phase voltage Extract the first Power of the machine tool after adjustment ; Calculate the first Table machine tool dynamic frequency and dynamic voltage .

[0012] A further improvement of the present invention is that step S1 includes transmitting the original signal to the edge computing unit for denoising, extracting time-domain and frequency-domain features, and converting the vibration signal into a time-frequency image through synchronous compression transformation.

[0013] On the other hand, the present invention provides a fault-based machine tool power management method system based on energy optimization, comprising: The data acquisition module is used to simultaneously install a three-axis vibration sensor and current and voltage sensors on the spindle and motor drive side of each machine tool to collect the vibration and electrical signals of the machine tool in real time. The fault alarm module is used to set fault alarm policies. After a fault is triggered, the load redistribution process is started. The load redistribution module is used to establish a multi-criteria decision model, dynamically assign priority weights to each machine tool, and establish a load distribution algorithm to allocate power to each machine tool. The dynamic adjustment module is used to dynamically adjust the motor drive voltage and frequency according to the actual load distribution.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention firstly uses online monitoring of reconstruction error by an autoencoder to promptly identify faulty machine tools and trigger a reassignment process; 2. Through load distribution algorithms, critical machine tools can maintain high-priority power supply even during fault switching, avoiding production line shutdowns due to single-point failures; 3. By combining online fault detection, multi-criteria decision-making, model predictive control and dynamic adjustment, this invention can instantly capture faults and quickly redistribute energy resources when machine tools fail or loads change suddenly. At the same time, by using technologies such as dynamic voltage-frequency adjustment and energy storage feedback, it can achieve multi-machine collaboration and on-demand energy supply, significantly reduce energy consumption fluctuations and idle waste, and improve system reliability and production continuity. Attached Figure Description

[0015] Figure 1 This is a flowchart of the energy-optimized fault machine tool power management method of the present invention; Figure 2 This is a framework diagram of the energy-optimized fault machine tool power management system of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0017] The term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three cases: A exists alone, A and B exist simultaneously, and B exists alone.

[0018] Example 1 Figure 1The flowchart of the energy-optimized fault-based machine tool power management method disclosed in this embodiment is shown, and the steps are as follows: S1. Install a three-axis vibration sensor and a current and voltage sensor on the spindle and motor drive side of each machine tool to collect the vibration signal and electrical signal of the machine tool in real time. The original signal is transmitted to the edge computing unit for denoising, time-domain and frequency-domain feature extraction, and the vibration signal is converted into a time-frequency image through synchronous squeezing transformation.

[0019] A three-axis accelerometer is installed on the machine tool spindle, with a sampling frequency typically selected between 5-10 kHz to cover the main fault frequency band. A short-time Fourier transform is then performed, and the amplitude spectrum is calculated for each time window. Normalized and scaled to a fixed size (e.g., 128×128 pixels), it serves as input to the fault alarm strategy.

[0020] S2. Set up a fault alarm policy. After a fault is triggered, start the load redistribution process. The fault alarm strategy constructs an encoder by batch normalizing the output of the preprocessed step S1 through continuous convolutional layers and then downsampling it to the latent spatial dimension using an activation function. It then reconstructs an image with the same dimension as the original image by batch normalizing the symmetric transposed convolutional layer and using an activation function. Training was performed using fault-free time-frequency images, followed by training with the Adam optimizer for 50-100 rounds until the reconstruction error on the validation set converged; this was then applied to newly acquired time-frequency images. Calculate the reconstructed new time-frequency graph Calculate the reconstruction error ; Reconstruction error of health dataset And calculate the average health data. with standard deviation The fault alarm threshold is 'k' represents the set variation coefficient. When the reconstruction error exceeds the fault alarm threshold, it is determined to be a machine fault, and a fault alarm is triggered. Subsequently, the PLC receives a stop signal, marks the machine tool status as stopped, and maps the "reconstruction error exceeds threshold" event to a Boolean input flag F, which is defined internally in the PLC as an input register such as I:0 / 1. When the threshold detection result is true, the output flag Fault_Flag is set. Using HART standard commands: Command 0: Read Primary Variable, used to display the machine tool's operating load in real time; Command 1: Read dynamic variables (such as reconstruction error, threshold); In HART command 15 (Read General Disposal Code), the Fault_Flag is packaged into diagnostic bytes through a custom device description (DD) to enable remote visibility of the fault status. If WirelessHART is used, the device uses the IEC62591 standard self-organizing mesh routing to send the message to the SCADA / IIoT platform through the gateway and start the load redistribution process.

[0021] S3. Establish a multi-criteria decision-making model to dynamically assign priority weights to each machine tool, and establish a load allocation algorithm to allocate power to each machine tool; The multi-criteria decision-making model includes a hierarchical sub-model, which comprises an objective layer, a criterion layer, and a solution layer. The objective layer is used to prioritize machine tool power supply, the criterion layer is used to obtain process criticality, safety risks, and delivery timeliness, and the solution layer is used to create a machine tool dataset. , Let S represent the nth machine tool, where n represents the total number of machine tools; perform eigenvalue decomposition of the covariance matrix on the index data to obtain the eigenvalues ​​of the covariance matrix S. With feature vectors Select the top M principal components, and use the loadings of each criterion layer index on the top M principal components as the contribution; then the loading of criterion layer index j on the kth principal component. Then the weight of index j is expressed as , where c represents the number of principal components with a cumulative contribution rate of 85%, and m represents the number of indicators in the criterion layer; The data sources for the process criticality are: analyzing the machine tool's position in the entire production process and its upstream and downstream dependencies through BOM (Bill of Materials) and process flow diagram analysis; obtaining a process dependency matrix by statistically analyzing how many subsequent processes directly depend on the machine tool's output; identifying equipment with high utilization rates and frequent bottlenecks through capacity utilization rate and bottleneck analysis reports; and then using the... The proportion of subsequent processes of a machine tool to the total number of subsequent processes of all machine tools is obtained. Taiwan machine tool dependence index; through the first Machine tool downtime accounts for the largest share of all machine tool downtime-related production losses. The bottleneck contribution of the machine tool is obtained by weighting and summing the dependence index and the bottleneck contribution. Key process parameters of the machine tool; The safety risks are reported through FMEA (Failure Mode and Effects Analysis), including potential failure modes, impact severity, frequency of occurrence, and detectability of each machine tool; obtaining the number of accidents, downtime, and injury level for each machine tool; and classifying machine tools according to different hazard levels based on CE, ISO 13849, etc., assigning a score (1-10 points) to each hazard level from highest to lowest; calculating the risk priority number RPN for each failure mode: RPN = S × O × D, where S represents the severity of the failure impact, O represents the frequency of failure occurrence, and D represents the undetectability; calculating the first... Weighted average of machine tool accident losses in Taiwan ,in, Indicates the number of historical accidents. Indicates the total downtime. This represents the cumulative score based on the damage level; the score is obtained after normalizing the accident losses. Calculate security risks , Indicates weight; The delivery timeframe indicates the countdown to the remaining delivery time.

[0022] The multi-criteria decision-making model also includes a factor set composed of indicators from the criterion layer. And their corresponding weights are sent into the weight set. And set evaluation levels. The membership degree of each factor j to each level i is calculated using the trapezoidal membership function. The membership matrix is ​​obtained. ; Perform a weighted average of the weight set and the membership matrix. , , obtained the Table machine tool rating The rating set GR for each machine tool is obtained by sorting them from largest to smallest.

[0023] This embodiment combines production progress and order urgency, and updates the weight distribution in real time through the SCADA system to ensure that critical processes receive the highest power supply priority during fault switching, and the corresponding level is the machine tool level assignment.

[0024] The load allocation algorithm includes defining the optimization objective as the first... The machine tool is designed to minimize its total energy consumption. Minimizing total energy consumption is expressed as the power output within time t within the total control time domain length T. The constraint condition is that the sum of the squares of the machine tool's energy consumption is less than or equal to the set maximum energy consumption value.

[0025] In model predictive control, control is performed over a finite prediction time window, where each time step represents a future prediction point of a fixed duration (e.g., 1 second or 100 ms). Therefore, the summation in this embodiment is the accumulation of all control time steps within the future prediction time domain.

[0026] If the control system samples once per second, and N=10, it means that the model considers power scheduling for the next 10 seconds in each optimization. N is one of the hyperparameters in model predictive control. A larger N indicates a longer time range being considered, which usually leads to higher accuracy, but also higher computational cost. This represents the sum of the squares of energy consumption, with the goal of minimizing total power consumption. This is a penalty term indicating the speed (smoothness) of power adjustment to prevent overly drastic adjustments.

[0027] The load allocation algorithm also includes power allocation command control, specifically including allocating the power load of each machine tool through an MPC model. Load target Using the optimization objective as input to the algorithm, a control sequence with the total time domain length of future time control is obtained. and execute the first control input. The prediction cycle restarts at the next time step; the prediction step size is set. With control step size At each control time, g solves for the optimization objective. , Indicates the first The state vector for each prediction step; then new measurements are acquired to update the prediction.

[0028] In this embodiment, the problem can be transformed into a quadratic programming problem with linear constraints (QP), and solved online using algorithms such as the interior point method or ADMM. Only the first control increment in the optimal sequence is executed, and then the new measurement value is obtained to update the prediction, ensuring robustness. In large-scale workshops, the overall problem can be decomposed, and the local model prediction and control of each region can be co-optimized. The inputs to model predictive control include the current system state, reference target, system model, weight matrix and constraints, and the output is an optimal control sequence. In actual execution, only the first input is taken and updated continuously.

[0029] The allocation command control also includes incorporating machine tool failures and load surges into the algorithm parameters, monitoring the real-time power requests of the faulty machine tool, and when the power demand of the faulty machine tool decreases beyond a set threshold within a set short time t, the system allocates the excess power to other machine tools in real time. Specifically, this is manifested by extracting the change in the power request of the faulty machine tool. The proportion of each machine tool's power allocation command in the total power is extracted as the allocation weight of each machine tool. This weight is then used to calculate the allocation weight of each machine tool. The power distribution command control adjustment value is obtained by multiplying the power distribution command and the power distribution command control adjustment value. Based on the power distribution command control, the power of each machine tool is adjusted again.

[0030] S4. Dynamically adjust the motor drive voltage and frequency according to the actual load distribution. Specifically, this includes installing a voltage sensor in the power supply circuit of each machine tool motor to collect the voltage of each machine tool in real time and setting the rated frequency of the motor. (e.g., 50Hz) and rated phase voltage Extract the first Power of the machine tool after adjustment ; Calculate the first Table machine tool dynamic frequency and dynamic voltage .

[0031] The threshold and weight settings can be set by default according to the present invention, or they can be set by those skilled in the art.

[0032] Example 2 Figure 2 The diagram illustrates the system framework of the energy-optimized fault-tolerant machine tool power management system of the present invention. Based on the same inventive concept as Embodiment 1, the present invention provides an energy-optimized fault-tolerant machine tool power management system, comprising: The data acquisition module is used to simultaneously install a three-axis vibration sensor and current and voltage sensors on the spindle and motor drive side of each machine tool to collect the vibration and electrical signals of the machine tool in real time. The fault alarm module is used to set fault alarm policies. After a fault is triggered, the load redistribution process is started. The load redistribution module is used to establish a multi-criteria decision model, dynamically assign priority weights to each machine tool, and establish a load distribution algorithm to allocate power to each machine tool. The dynamic adjustment module is used to dynamically adjust the motor drive voltage and frequency according to the actual load distribution.

[0033] The load redistribution module includes a multi-criteria decision model construction unit and a load distribution algorithm construction unit; the multi-criteria decision model construction unit is used to dynamically allocate priority weights to each machine tool, including a hierarchical structure sub-unit; the load distribution algorithm construction unit is used to allocate power to each machine tool, including a power distribution command control sub-unit.

[0034] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0035] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0036] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0037] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0038] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A fault-based machine tool power management method based on energy optimization, characterized in that: Includes the following steps: S1. Install a three-axis vibration sensor and a current and voltage sensor on the spindle and motor drive side of each machine tool to collect the vibration signal and electrical signal of the machine tool in real time. S2. Set up a fault alarm policy. After a fault is triggered, start the load redistribution process. S3. Establish a multi-criteria decision-making model to dynamically assign priority weights to each machine tool, and establish a load allocation algorithm to allocate power to each machine tool; S4. Dynamically adjust the motor drive voltage and frequency according to the actual load distribution; The load distribution algorithm includes defining the optimization objective as minimizing the total energy consumption of the machine tool, and setting the constraint that the sum of the squares of the energy consumption of the machine tool is less than or equal to the set maximum energy consumption. The load allocation algorithm also includes power allocation command control, specifically including allocating the power load of each machine tool through an MPC model. Load target Using the optimization objective as input to the algorithm, a control sequence with the total time domain length of future time control is obtained. and execute the first control input. The prediction cycle restarts at the next time step; the prediction step size is set. With control step size At each control time, g solves for the optimization objective. , Indicates the first The state vector for each prediction step; then new measurements are acquired to update the prediction; The allocation command control also includes incorporating machine tool failures and load surges into the algorithm parameters, monitoring the real-time power requests of the faulty machine tool, and when the power demand of the faulty machine tool decreases beyond a set threshold within a set short time t, the system allocates the excess power to other machine tools in real time. Specifically, this is manifested by extracting the change in the power request of the faulty machine tool. The proportion of each machine tool's power allocation command in the total power is extracted as the allocation weight of each machine tool. This weight is then used to calculate the allocation weight of each machine tool. The power distribution command control adjustment value is obtained by multiplying the power distribution command and the power distribution command control adjustment value. Based on the power distribution command control, the power of each machine tool is adjusted again.

2. The energy-optimized fault-based machine tool power management method according to claim 1, characterized in that: The fault alarm strategy constructs an encoder by batch normalizing the output of the preprocessed step S1 through continuous convolutional layers and then downsampling it to the latent spatial dimension using an activation function. It then reconstructs an image with the same dimension as the original image by batch normalizing the symmetric transposed convolutional layer and using an activation function. Training was performed using fault-free time-frequency images, followed by training with the Adam optimizer for 50-100 rounds until the reconstruction error on the validation set converged; this was then applied to newly acquired time-frequency images. Calculate the reconstructed new time-frequency graph Calculate the reconstruction error ; Reconstruction error of health dataset And statistically analyze the average health data. with standard deviation The fault alarm threshold is k represents the set variation coefficient. When the reconstruction error is greater than the fault alarm threshold, it is judged as a machine fault and a fault alarm is triggered. Subsequently, the PLC receives the shutdown signal, marks the machine tool status as stopped, and synchronizes the status update to the SCADA / IoT platform through the industrial Ethernet to start the load redistribution process.

3. The energy-optimized fault-based machine tool power management method according to claim 1, characterized in that: The multi-criteria decision-making model includes a hierarchical sub-model, which comprises an objective layer, a criterion layer, and a solution layer. The objective layer is used to prioritize machine tool power supply, the criterion layer is used to obtain process criticality, safety risks, and delivery timeliness, and the solution layer is used to create a machine tool dataset. , Let S represent the nth machine tool, where n represents the total number of machine tools; perform eigenvalue decomposition of the covariance matrix on the index data to obtain the eigenvalues ​​of the covariance matrix S. With feature vectors ; Select the first M principal components, and use the loadings of each criterion layer index on the first M principal components as the contribution; then the loading of criterion layer index j on the kth principal component. Then the weight of index j is expressed as , where c represents the number of principal components with a cumulative contribution rate of 85%, and m represents the number of criteria layer indicators.

4. The energy-optimized fault-based machine tool power management method according to claim 3, characterized in that: The multi-criteria decision-making model also includes a factor set composed of indicators from the criterion layer. And their corresponding weights are sent into the weight set. And set evaluation levels. The membership degree of each factor j to each level i is calculated using the trapezoidal membership function. The membership matrix is ​​obtained. ; Perform a weighted average of the weight set and the membership matrix. , , obtained the Table machine tool rating The rating set GR for each machine tool is obtained by sorting them from largest to smallest.

5. The energy-optimized fault-based machine tool power management method according to claim 1, characterized in that: Step S4 includes installing a voltage sensor in the power supply circuit of each machine tool motor to collect the voltage of each machine tool in real time and setting the rated frequency of the motor. With rated phase voltage Extract the first Power of the machine tool after adjustment ; Calculate the first Dynamic frequency of benchtop machine tool and dynamic voltage .

6. The energy-optimized fault-based machine tool power management method according to claim 1, characterized in that: Step S1 includes transmitting the original signal to the edge computing unit for denoising, extracting time-domain and frequency-domain features, and converting the vibration signal into a time-frequency image through synchronous squeezing transformation.

7. An energy-optimized fault-tolerant machine tool power management system, used to execute the energy-optimized fault-tolerant machine tool power management method as described in any one of claims 1-6, characterized in that: include: The data acquisition module is used to simultaneously install a three-axis vibration sensor and current and voltage sensors on the spindle and motor drive side of each machine tool to collect the vibration and electrical signals of the machine tool in real time. The fault alarm module is used to set fault alarm policies. After a fault is triggered, the load redistribution process is started. The load redistribution module is used to establish a multi-criteria decision model, dynamically assign priority weights to each machine tool, and establish a load distribution algorithm to allocate power to each machine tool. The dynamic adjustment module is used to dynamically adjust the motor drive voltage and frequency according to the actual load distribution.

Images