Numerical control machine precision machining control system and method based on intelligent sensing

Abstract

The invention discloses a numerical control machine precise machining control system and method based on intelligent sensing, and aims at solving the problems that in existing numerical control machine machining, parameters depend on manual work, sensing data are prone to being interfered, and machining deviation cannot be corrected in real time. The system comprises an intelligent sensing collection layer, an edge calculation layer, an intelligent decision-making layer, an execution layer, a feedback layer and a local database, the intelligent sensing collection layer collects processing data through an anti-electromagnetic interference packaging structure and multiple sensors, the edge calculation layer achieves data noise reduction and feature extraction, the intelligent decision-making layer optimizes processing parameters through a fusion algorithm, and the execution layer achieves feature extraction. And the feedback layer and the execution layer form closed-loop correction. The method comprises the steps of data acquisition and preprocessing, feature recognition and evaluation, parameter adaptive optimization, real-time deviation correction and data precipitation. The method improves the accuracy of sensing data and the machining precision, gets rid of the dependence of artificial experience, prejudges the service life of the cutter, reduces the machining cost, is suitable for high-precision part machining, and has high practical value and industrial application prospects.

Description

Technical Field

[0001] The present invention relates to the field of precise machining of numerical control machines, specifically to a precise machining control system and method for numerical control machines based on intelligent sensing. Background Art

[0002] In the field of numerical control machine machining, the processing requirements for high-precision parts are increasing day by day. Especially in industries such as aerospace and medical devices, the requirements for workpiece machining accuracy, machining efficiency, and machining stability are constantly rising. In current numerical control machine machining technologies, most machining systems rely on manual experience to configure machining parameters, resulting in problems such as poor parameter adaptability and large tolerance fluctuations, making it difficult to meet the high-precision machining requirements. At the same time, in the sensing and acquisition link of existing numerical control machine machining systems, it is often affected by power frequency interference and servo motor pulse width modulation signals in industrial scenarios, resulting in deviations in the data collected by piezoelectric force sensors, temperature sensors, etc., and being unable to accurately reflect the machining state. In addition, the traditional system lacks a complete closed-loop control mechanism, and machining deviations cannot be detected and corrected in real time, easily causing workpiece scrapping. Although some systems in the prior art introduce sensing technologies, they have defects such as high data processing delay, no self-learning ability, and inability to predict tool wear, resulting in increased machining costs and low production efficiency, and being difficult to adapt to the precise machining requirements under complex working conditions. Therefore, researching and developing a precise machining control system and method for numerical control machines based on intelligent sensing that can solve the above problems has become an urgent need in the current industry. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to overcome the above-mentioned technical defects and provide a precise machining control system and method for numerical control machines based on intelligent sensing.

[0004] To address the aforementioned technical problems, the present invention provides a CNC precision machining control system and method based on intelligent sensing: The CNC precision machining control system based on intelligent sensing includes an intelligent sensing acquisition layer, an edge computing layer, an intelligent decision-making layer, an execution layer, and a feedback layer. These layers are sequentially electrically connected to form a closed-loop control architecture. The intelligent sensing acquisition layer integrates a piezoelectric force sensor, a high-precision temperature sensor, an acoustic emission sensor, and a laser displacement sensor. The intelligent sensing acquisition layer is equipped with an anti-electromagnetic interference packaging structure for... The system isolates interference signals in industrial scenarios. The edge computing layer is equipped with a lightweight chip for real-time noise reduction, feature extraction, and local inference of data collected by the intelligent sensing layer. The intelligent decision-making layer features an adaptive decision engine that employs a fusion algorithm to dynamically optimize cutting parameters based on the machining material and tool wear condition. The execution layer includes a high-precision servo drive module and an intelligent tool adjustment mechanism to respond to parameter commands output by the intelligent decision-making layer. The feedback layer, through a combination of visual inspection and digital twin simulation, detects machining deviations in real time and feeds them back to the intelligent decision-making layer.

[0005] As an improvement, the anti-electromagnetic interference packaging structure is used to solve the problem of power frequency interference and interference of servo motor pulse width modulation signal on small signals in industrial scenarios. The sensors in the intelligent sensing and acquisition layer synchronously collect various data during the processing.

[0006] As an improvement, the lightweight chip of the edge computing layer can control the data processing and decision-making delays within a preset range, avoid processing deviations caused by cloud transmission lag, and achieve local rapid data processing.

[0007] As an improvement, the fusion algorithm of the intelligent decision layer is a fusion algorithm of long short-term memory network and deep Q network, which can overcome the deficiency of insufficient generalization ability of single algorithm and realize dynamic adaptive adjustment of processing parameters.

[0008] As an improvement, the system also includes a local database, which is electrically connected to the intelligent decision-making layer and is used to store processing parameters, sensor data and deviation data, providing parameter support for subsequent processing of similar workpieces and enabling the system to learn and iterate.

[0009] A precision machining control method for CNC machines based on intelligent sensing includes the following steps: First, sensor data acquisition and preprocessing: Sensors in the intelligent sensing acquisition layer synchronously acquire data on cutting force, spindle vibration, machining temperature, tool wear, and workpiece dimensional deviation. Wavelet threshold noise reduction algorithm is used to process the acquired data, eliminating electromagnetic interference and mechanical noise while retaining minor faults and deviations. Second, feature recognition and status assessment: Core features such as tool wear, spindle operating status, and workpiece dimensional deviation are extracted through an edge computing layer. An improved long short-term memory network is used to predict the remaining tool life. Third, adaptive parameter optimization, combined with workpiece... The system employs a material database and real-time machining status, dynamically adjusting machining parameters such as feed rate, spindle speed, and depth of cut using a deep Q-network algorithm to address tolerance fluctuations caused by manual experience-based parameters. The fourth step involves real-time execution and deviation correction. The optimized machining parameters are executed through the execution layer, and workpiece dimensional deviations are detected in real-time by a laser displacement sensor. This deviation data is fed back to the intelligent decision layer, which generates compensation commands to drive the intelligent tool adjustment mechanism for online correction. The fifth step involves full lifecycle data accumulation. Machining parameters, sensor data, and deviation data are stored in a local database to provide parameter support for subsequent machining of similar workpieces, enabling the system to self-learn and iterate.

[0010] As an improvement, the wavelet threshold denoising algorithm in the first step is an original design that can accurately remove various interference signals, ensuring the accuracy and effectiveness of the extracted processing data.

[0011] As an improvement, the feature recognition process in the second step is completed at the edge computing layer, eliminating the need to transmit the raw data to the cloud and shortening the response time for feature extraction and state assessment.

[0012] As an improvement, the online correction process in the fourth step achieves seamless integration of data collection, analysis, decision-making, and correction, ensuring that processing deviations can be eliminated in real time and improving processing accuracy.

[0013] As an improvement, the data accumulation process in the fifth step can continuously optimize the processing parameters, enabling the system to adapt to the processing needs of workpieces of different materials and specifications, and improving the system's generalization ability.

[0014] The advantages of this invention compared to existing technologies are as follows: First, it improves the accuracy and reliability of sensor data acquisition. The anti-electromagnetic interference encapsulation structure of the intelligent sensor acquisition layer can isolate various interference signals, and the wavelet threshold noise reduction algorithm of the edge computing layer removes noise, ensuring the accuracy of data such as cutting force and machining temperature, providing reliable support for subsequent processing. Second, it achieves adaptive optimization of machining parameters. The adaptive decision engine of the intelligent decision layer dynamically adjusts machining parameters by combining machining status and historical data through a fusion algorithm, eliminating reliance on manual experience and reducing tolerance fluctuations. Third, it constructs a closed-loop control system. The feedback layer detects and feeds back workpiece dimensional deviations in real time, and the intelligent decision layer generates compensation commands. The intelligent tool adjustment mechanism corrects the deviations online, significantly improving machining accuracy and reducing defect rates. Fourth, it accurately predicts the remaining tool life, detects abnormal wear in advance, extends tool life, reduces unplanned downtime and workpiece scrap, and lowers machining costs. Fifth, it has self-learning and iterative capabilities. The local database stores various machining data, and the system optimizes algorithms and parameters through data analysis, improving the adaptability of the system to different workpieces. Sixth, the edge computing layer enables local data processing, reduces transmission latency, improves system response speed, ensures real-time parameter optimization and deviation correction, and balances processing accuracy and efficiency. Attached Figure Description

[0015] Figure 1 This is the overall system architecture diagram of the CNC precision machining control system based on intelligent sensing of this invention.

[0016] Figure 2 This is a diagram showing the composition of the intelligent sensing acquisition layer of the CNC precision machining control system based on intelligent sensing, as described in this invention.

[0017] Figure 3 This is a functional diagram of the edge computing layer of the CNC precision machining control system based on intelligent sensing, as described in this invention.

[0018] Figure 4 This is a functional diagram of the intelligent decision-making layer of the CNC precision machining control system based on intelligent sensing, as described in this invention.

[0019] Figure 5 This is a diagram showing the execution layer composition of the CNC precision machining control system based on intelligent sensing according to the present invention.

[0020] Figure 6 This is a functional diagram of the feedback layer of the CNC precision machining control system based on intelligent sensing, as described in this invention.

[0021] Figure 7 This is an overall flowchart of the control method for precision machining of CNC machines based on intelligent sensing, as described in this invention.

[0022] Figure 8This is a detailed diagram of the first step in the CNC precision machining control method based on intelligent sensing of the present invention: sensor data acquisition and preprocessing.

[0023] Figure 9 This is a detailed diagram of the second step of the CNC precision machining control method based on intelligent sensing in this invention: feature recognition and state assessment.

[0024] Figure 10 This is a detailed diagram of the third step: adaptive parameter optimization, in the CNC precision machining control method based on intelligent sensing of this invention.

[0025] Figure 11 This is a detailed diagram of the fourth step of the CNC precision machining control method based on intelligent sensing in this invention: real-time execution and deviation correction.

[0026] Figure 12 This is a detailed diagram of the fifth step in the CNC precision machining control method based on intelligent sensing of the present invention: full life cycle data accumulation. Detailed Implementation

[0027] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings, which illustrate embodiments of the present application. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of this application will be thorough and complete.

[0028] Referring to the accompanying drawings, a precision machining control system and method for CNC machines based on intelligent sensing is described. The precision machining control system for CNC machines based on intelligent sensing includes an intelligent sensing acquisition layer, an edge computing layer, an intelligent decision-making layer, an execution layer, and a feedback layer. These layers are electrically connected sequentially to form a closed-loop control architecture. The intelligent sensing acquisition layer integrates a piezoelectric force sensor, a high-precision temperature sensor, an acoustic emission sensor, and a laser displacement sensor. It also features an anti-electromagnetic interference packaging structure to isolate interference signals in industrial environments. The edge computing layer carries a lightweight chip for real-time noise reduction, feature extraction, and local inference of the data acquired by the intelligent sensing acquisition layer. The intelligent decision-making layer has an adaptive decision engine that uses a fusion algorithm to dynamically optimize cutting parameters based on the machining material and tool wear state. The execution layer includes a high-precision servo drive module and an intelligent tool adjustment mechanism to respond to parameter commands output by the intelligent decision-making layer. The feedback layer uses visual inspection and digital twin simulation in conjunction to detect machining deviations in real time and feed them back to the intelligent decision-making layer.

[0029] The electromagnetic interference-resistant packaging structure is used to solve the problem of power frequency interference and servo motor pulse width modulation signal interference to small signals in industrial scenarios. The sensors in the intelligent sensing and acquisition layer synchronously collect various data during the processing.

[0030] The lightweight chip in the edge computing layer can control data processing and decision-making delays within a preset range, avoiding processing deviations caused by cloud transmission lag, and enabling rapid local data processing.

[0031] The fusion algorithm of the intelligent decision layer is a fusion algorithm of long short-term memory network and deep Q network, which can overcome the deficiency of insufficient generalization ability of single algorithm and realize dynamic adaptive adjustment of processing parameters.

[0032] The system also includes a local database, which is electrically connected to the intelligent decision-making layer and is used to store processing parameters, sensor data, and deviation data, providing parameter support for the processing of similar workpieces in the future and enabling the system to learn and iterate.

[0033] A precision machining control method for CNC machines based on intelligent sensing includes the following steps: First, sensor data acquisition and preprocessing: Sensors in the intelligent sensing acquisition layer synchronously acquire data on cutting force, spindle vibration, machining temperature, tool wear, and workpiece dimensional deviation. Wavelet threshold noise reduction algorithm is used to process the acquired data, eliminating electromagnetic interference and mechanical noise while retaining minor faults and deviations. Second, feature recognition and status assessment: Core features such as tool wear, spindle operating status, and workpiece dimensional deviation are extracted through an edge computing layer. An improved long short-term memory network is used to predict the remaining tool life. Third, adaptive parameter optimization, combined with workpiece... The system employs a material database and real-time machining status, dynamically adjusting machining parameters such as feed rate, spindle speed, and depth of cut using a deep Q-network algorithm to address tolerance fluctuations caused by manual experience-based parameters. The fourth step involves real-time execution and deviation correction. The optimized machining parameters are executed through the execution layer, and workpiece dimensional deviations are detected in real-time by a laser displacement sensor. This deviation data is fed back to the intelligent decision layer, which generates compensation commands to drive the intelligent tool adjustment mechanism for online correction. The fifth step involves full lifecycle data accumulation. Machining parameters, sensor data, and deviation data are stored in a local database to provide parameter support for subsequent machining of similar workpieces, enabling the system to self-learn and iterate.

[0034] The wavelet threshold denoising algorithm in the first step is an original design that can accurately remove various interference signals, ensuring the accuracy and effectiveness of the extracted processing data.

[0035] The feature recognition process in the second step is completed at the edge computing layer, eliminating the need to transmit the raw data to the cloud and shortening the response time for feature extraction and status assessment.

[0036] The online correction process in the fourth step achieves seamless integration of data collection, analysis, decision-making, and correction, ensuring that processing deviations can be eliminated in real time and improving processing accuracy.

[0037] The data accumulation process in the fifth step enables continuous optimization of processing parameters, allowing the system to adapt to the processing needs of workpieces of different materials and specifications, and improving the system's generalization ability.

[0038] The core objective of this invention is to address the technical pain points in existing CNC machining processes, such as the susceptibility of sensor data to interference, reliance on human experience for machining parameters, inability to correct machining deviations in real time, and weak system adaptability. By constructing a technical system of multimodal sensor fusion, edge intelligent decision-making, and full-process closed-loop control, this invention achieves high precision, high stability, and high efficiency in CNC machining, making it suitable for high-precision parts machining scenarios in fields such as aerospace, medical devices, and precision instruments.

[0039] Specific implementation of a precision machining control system for CNC machines based on intelligent sensing:

[0040] The intelligent sensing-based precision machining control system for CNC machines in this embodiment strictly corresponds to the structure described in claims 1 to 5. Specifically, it includes an intelligent sensing acquisition layer, an edge computing layer, an intelligent decision-making layer, an execution layer, a feedback layer, and a local database. Each layer is electrically connected through wires to form a complete closed-loop control architecture, ensuring the real-time performance and stability of data transmission. The specific implementation details of each part are as follows.

[0041] The intelligent sensing and acquisition layer, as the core of the system, integrates a piezoelectric force sensor, a high-precision temperature sensor, an acoustic emission sensor, and a laser displacement sensor. All four sensors employ a modular design for easy installation, debugging, and replacement, and are positioned at key machining locations on the CNC machine: the piezoelectric force sensor is installed at the connection between the tool holder and the spindle to collect cutting force data during machining; the high-precision temperature sensor is installed near the contact area between the spindle and the workpiece to collect temperature change data during machining; the acoustic emission sensor is installed on the machine bed to collect acoustic emission signals generated by tool wear and workpiece deformation during machining; and the laser displacement sensor is installed on the machine beam, aligned with the machining surface of the workpiece, to collect dimensional deviation data of the workpiece.

[0042] To address the interference issues of power frequency interference and servo motor pulse width modulation signals on minute sensor signals in industrial scenarios, the intelligent sensing acquisition layer is equipped with an anti-electromagnetic interference packaging structure. This anti-electromagnetic interference packaging structure adopts a double-layer shielding design, with an inner metal shielding mesh and an outer insulating protective shell. The metal shielding mesh is made of copper, which can effectively block the penetration of electromagnetic signals. The insulating protective shell is made of high-temperature and corrosion-resistant engineering plastics, which not only provide protection but also further isolate interference signals, ensuring the accuracy of various data collected by the intelligent sensing acquisition layer.

[0043] The edge computing layer is equipped with a lightweight chip, which uses a low-power, high-speed dedicated processing chip to achieve real-time processing of sensor data without transmitting raw data to the cloud, thus avoiding processing deviations caused by cloud transmission lag. The core function of the edge computing layer is to perform noise reduction, feature extraction, and local inference on the raw data collected by the intelligent sensing acquisition layer. The noise reduction process uses a wavelet threshold noise reduction algorithm, which can accurately remove electromagnetic interference and mechanical noise while retaining minor fault and deviation features, providing reliable data support for subsequent state assessment and parameter optimization.

[0044] In the data denoising process, to quantify the denoising effect, a denoising signal-to-noise ratio (SNR) evaluation metric is introduced, and its calculation formula is as follows: ,in, This indicates the noise reduction signal-to-noise ratio, expressed in decibels. This indicates the power of the effective sensing signal after noise reduction; This represents the power of the residual noise after denoising. This formula is used to quantitatively evaluate the denoising effect of the wavelet thresholding denoising algorithm. The larger the value, the better the noise reduction effect and the higher the purity of the effective signal. The wavelet threshold parameter can be adjusted in real time through this formula to ensure the targeted and effective noise reduction processing.

[0045] The intelligent decision-making layer is equipped with an adaptive decision-making engine. This engine employs a fusion algorithm combining Long Short-Term Memory (LSTM) networks and Deep Q-Networks, overcoming the limitations of single-algorithm generalization. It dynamically optimizes cutting parameters based on real-time machining information such as material properties and tool wear. The core logic of the adaptive decision-making engine is to first determine the current machining state using core features extracted from the edge computing layer, then combine this with machining data of similar workpieces stored in the local database, and finally generate optimal machining parameters through a fusion algorithm, achieving adaptive adjustment of the machining parameters.

[0046] To accurately describe the relationship between tool wear and machining parameters, a mapping model between tool wear and cutting parameters is established, and its calculation formula is as follows: ,in, Indicates the amount of tool wear; Indicates the spindle speed; Indicates the feed rate; Indicates the depth of cut; Indicates processing time; This represents a mapping function, obtained through training with a Long Short-Term Memory (LSTM) network. Based on historical machining data, this function can fit the variation pattern of tool wear under different cutting parameters and machining times. The purpose of this formula is to provide a theoretical basis for adaptive parameter optimization. By inputting the current spindle speed, feed rate, depth of cut, and machining time, it can accurately predict the current tool wear, thereby adjusting cutting parameters, extending tool life, and avoiding machining deviations caused by excessive tool wear.

[0047] The execution layer includes a high-precision servo drive module and an intelligent tool adjustment mechanism. The high-precision servo drive module is connected to the spindle and feed axis of the CNC machine and is used to receive parameter commands output by the intelligent decision layer and drive the spindle and feed axis to run according to the optimized parameters. The intelligent tool adjustment mechanism is installed at the tool holder and can fine-tune the position of the tool according to the compensation commands generated by the intelligent decision layer to realize online correction of machining deviations.

[0048] The feedback layer links visual inspection with digital twin simulation. The visual inspection module uses an industrial camera to capture real-time images of the workpiece's machining surface, extracts the workpiece's dimensional information, and compares it with preset standard dimensions to obtain machining deviation data. The digital twin simulation module constructs a digital twin model of the CNC machine and the machining process, inputs the machining deviation data obtained from visual inspection into the digital twin model, simulates the causes of machining deviations, provides a reference for the intelligent decision-making layer to generate compensation instructions, and simultaneously feeds the machining deviation data back to the intelligent decision-making layer in real time, forming a closed-loop control of "acquisition-analysis-decision-execution-feedback".

[0049] The local database is electrically connected to the intelligent decision-making layer to store various data during the processing, including processing parameters, sensor data, deviation data, tool wear data, etc. This data will serve as a reference for subsequent processing of similar workpieces. By analyzing and learning from historical data, the parameters of the fusion algorithm are continuously optimized to improve the system's self-learning and generalization capabilities, enabling the system to adapt to the processing needs of workpieces of different materials and specifications.

[0050] Specific implementation of CNC precision machining control method based on intelligent sensing:

[0051] The intelligent sensing-based CNC precision machining control method in this embodiment is applied to the intelligent sensing-based CNC precision machining control system described in the above embodiment. It strictly corresponds to the steps described in claims 6 to 10, and specifically includes the following five steps, which are executed sequentially to achieve precise control of the entire machining process.

[0052] The first step is sensor data acquisition and preprocessing. After starting the CNC machine, the piezoelectric force sensor, high-precision temperature sensor, acoustic emission sensor, and laser displacement sensor in the intelligent sensor acquisition layer start synchronously, respectively acquiring data on cutting force, spindle vibration, machining temperature, tool wear, and workpiece dimensional deviation during the machining process. During acquisition, the acquisition frequency of each sensor remains consistent to ensure data synchronization. The acquired raw data contains a large amount of electromagnetic interference and mechanical noise signals. These interference signals can affect the accuracy of the data; therefore, a wavelet threshold noise reduction algorithm is needed to preprocess the raw data.

[0053] The core formula of the wavelet thresholding noise reduction algorithm is as follows: ,in, These represent the wavelet coefficients after the original sensor data has undergone wavelet transform; This represents the wavelet coefficients after thresholding. Indicates the wavelet threshold; Indicates the scale of the wavelet transform; Indicates the position of the wavelet coefficients; This function represents the sign of the input value. It outputs 1 when the input is positive, -1 when the input is negative, and 0 when the input is 0. The formula's purpose is to filter the coefficients after the wavelet transform, removing those less than a threshold. The coefficients (mainly those corresponding to the noise signal) are retained if they are greater than the threshold. The coefficients (mainly the coefficients corresponding to the effective sensing signals) are then obtained through wavelet inverse transform to obtain the noise-reduced effective data, thereby eliminating electromagnetic interference and mechanical noise, retaining minor fault and deviation characteristics, and providing reliable data support for subsequent feature recognition and condition assessment.

[0054] The second step is feature recognition and status assessment. The preprocessed effective data is transmitted to the edge computing layer. The edge computing layer uses feature extraction algorithms to extract core features such as tool wear, spindle operating status, and workpiece dimensional deviation. Specifically, tool wear feature extraction is based on the variation patterns of acoustic emission signals and cutting force signals; spindle operating status feature extraction is based on the frequency distribution of spindle vibration signals; and workpiece dimensional deviation feature extraction is based on the difference between data collected by the laser displacement sensor and the standard dimensions.

[0055] After extracting the core features, an improved long short-term memory network is used to predict the remaining tool life. To avoid excessive prediction errors, a prediction accuracy evaluation index is introduced, and its calculation formula is as follows: ,in, This indicates the accuracy of the remaining tool life prediction. This indicates the number of samples where the predicted result matches the actual result. This represents the total number of samples involved in the prediction. This formula is used to quantitatively evaluate the prediction performance of the improved Long Short-Term Memory (LSTM) network. The higher the value, the more accurate the prediction result. This formula can be used to adjust network parameters in real time, improve the prediction accuracy, provide a basis for subsequent parameter optimization, and avoid machining failures and deviations caused by excessive tool wear.

[0056] The third step is adaptive parameter optimization. The intelligent decision layer receives core features and tool life prediction results transmitted from the edge computing layer, and combines them with the workpiece material database and historical machining data stored in the local database. It then dynamically adjusts machining parameters such as feed rate, spindle speed, and depth of cut using a deep Q-network algorithm. During parameter optimization, the optimization objectives are to achieve the highest machining accuracy, minimum tool wear, and highest machining efficiency. A parameter optimization objective function is established, and its calculation formula is as follows: ,in, This indicates the optimization of the objective function value; Indicates the workpiece dimensional deviation; Indicates the amount of tool wear; Indicates processing time; , , These represent the weighting coefficients for workpiece dimensional deviation, tool wear, and machining time, respectively. These coefficients are adjusted according to actual machining requirements, and their sum is 1. The purpose of this formula is to quantitatively evaluate the advantages and disadvantages of different combinations of machining parameters by minimizing... By finding the optimal combination of machining parameters, the problem of tolerance fluctuations caused by manual experience parameters can be solved, thereby improving machining accuracy and efficiency.

[0057] The fourth step is real-time execution and deviation correction. The intelligent decision-making layer transmits the optimized machining parameter instructions to the execution layer. After receiving the instructions, the high-precision servo drive module in the execution layer drives the spindle and feed axis of the CNC machine to operate according to the optimized parameters, while the intelligent tool adjustment mechanism is simultaneously in standby mode. During the machining process, the laser displacement sensor detects the dimensional deviation of the workpiece in real time and feeds the deviation data back to the feedback layer. The feedback layer analyzes the cause of the deviation through visual inspection and digital twin simulation, and then transmits the deviation data back to the intelligent decision-making layer.

[0058] The intelligent decision-making layer generates corresponding compensation instructions based on the deviation data. The calculation basis for the compensation instructions is as follows: ,in, Indicates the distance at which the cutting tool needs fine-tuning; This represents the compensation coefficient, which is adjusted according to the processing material and processing accuracy requirements, and its value ranges from 0.8 to 1.2. This indicates the workpiece dimensional deviation. The formula's function is to accurately calculate the tool's fine-tuning distance based on the workpiece dimensional deviation, ensuring the accuracy of the compensation command. After receiving the compensation command, the intelligent tool adjustment mechanism performs online fine-tuning of the tool's position, achieving seamless integration of data acquisition, analysis, decision-making, and correction. This eliminates machining deviations in real time and improves machining accuracy.

[0059] The fifth step is the accumulation of data throughout the entire lifecycle. During the machining process, the local database stores various data in real time, including machining parameters, sensor data, deviation data, tool wear data, and prediction results. This data serves as the foundation for the system's self-learning. After machining is completed, the system organizes and analyzes the machining data, optimizes the parameters of the fusion algorithm and the weighting coefficients of the machining parameters, and compares the current machining data with historical machining data of similar workpieces to improve the workpiece material database and machining parameter database. Through continuous data analysis and accumulation, the system achieves self-learning iteration, enabling it to adapt to the machining needs of workpieces of different materials and specifications, and improving the system's generalization ability and stability.

[0060] Implementation Results Explanation:

[0061] Through the implementation of the above embodiments, the intelligent sensing-based CNC precision machining control system and method of the present invention can effectively solve the technical pain points of existing CNC machining, significantly improve machining accuracy, extend tool life, and improve machining efficiency. By combining an anti-electromagnetic interference packaging structure with a wavelet threshold noise reduction algorithm, the accuracy of sensor data is effectively improved; by using a fusion algorithm of long short-term memory networks and deep Q-networks, adaptive optimization of machining parameters is achieved, eliminating reliance on human experience; and through full-process closed-loop control, real-time correction of machining deviations is achieved, ensuring the stability of machining accuracy. The implementation process of this embodiment is clear and highly operable, providing reliable technical support for the machining of high-precision parts and possessing extremely high industrial application value.

[0062] This invention effectively improves the accuracy and reliability of sensor data acquisition, addressing the technical pain point of sensor data being easily interfered with in existing technologies. The intelligent sensor acquisition layer in the intelligent sensing-based CNC precision machining control system of this invention features an anti-electromagnetic interference (EMI) encapsulation structure. This structure employs a double-layer shielding design, effectively isolating power frequency interference and servo motor pulse width modulation signals in industrial scenarios. This prevents these interference signals from interfering with the minute sensor signals acquired by piezoelectric force sensors, high-precision temperature sensors, acoustic emission sensors, and laser displacement sensors. Simultaneously, in conjunction with the wavelet threshold noise reduction algorithm used in the edge computing layer, EMI and mechanical noise in the original sensor data are further eliminated, while retaining minor fault and deviation characteristics. This ensures the accuracy of various data acquired by the intelligent sensing acquisition layer, such as cutting force, spindle vibration, machining temperature, tool wear, and workpiece dimensional deviations. This provides reliable data support for subsequent feature recognition, state assessment, and parameter optimization, fundamentally improving the control precision of the entire intelligent sensing-based CNC precision machining control system.

[0063] This invention achieves adaptive optimization of machining parameters, eliminating reliance on manual experience and enhancing the intelligence level of the machining process. The intelligent decision layer of the intelligent sensing-based CNC precision machining control system of this invention features an adaptive decision engine. This engine employs a fusion algorithm of Long Short-Term Memory (LSTM) networks and Deep Q-Networks. Based on core features extracted from the edge computing layer, such as tool wear, spindle operating status, and workpiece dimensional deviations, combined with workpiece material data and historical machining data stored in a local database, it dynamically adjusts machining parameters such as feed rate, spindle speed, and depth of cut. This eliminates the need for manual parameter configuration, solving the problems of reliance on manual experience, inconsistencies in parameter configurations by different operators, and tolerance fluctuations in existing technologies. Furthermore, by establishing a parameter optimization objective function, it achieves synergistic optimization of machining accuracy, tool life, and machining efficiency, improving the intelligence and stability of CNC machining.

[0064] This invention constructs a closed-loop control system for the entire process, enabling real-time correction of machining deviations and significantly improving machining accuracy. The intelligent sensing-based precision machining control system for CNC machines employs a closed-loop architecture where an intelligent sensing acquisition layer, edge computing layer, intelligent decision-making layer, execution layer, and feedback layer are sequentially electrically connected. The feedback layer, through visual inspection and digital twin simulation linkage, can detect workpiece dimensional deviations in real time and feed them back to the intelligent decision-making layer. The intelligent decision-making layer generates compensation commands based on the deviation data, driving the intelligent tool adjustment mechanism in the execution layer to fine-tune the tool position online. This achieves seamless integration of "sensor acquisition - data processing - feature recognition - parameter optimization - execution correction - deviation feedback," effectively eliminating dimensional deviations generated during machining and raising machining accuracy to a higher level. It is suitable for high-precision parts machining needs in aerospace, medical devices, and precision instruments, reducing workpiece defect rates.

[0065] This invention enables accurate prediction of remaining tool life, extending tool life and reducing machining costs. The intelligent sensing-based precision machining control method for CNC machines utilizes an improved long short-term memory network, combined with data such as acoustic emission signals and cutting force signals collected by the intelligent sensing acquisition layer. This allows for accurate prediction of remaining tool life, early detection of abnormal tool wear, and prevention of machining failures, workpiece scrap, and unplanned downtime caused by excessive tool wear. Furthermore, by adaptively optimizing cutting parameters, it reduces tool wear rate, extends tool life, lowers tool replacement frequency and machining costs, and improves production efficiency.

[0066] Possessing self-learning and iterative capabilities, this invention enhances the system's generalization ability and adaptability. The intelligent sensing-based CNC precision machining control system of this invention is equipped with a local database. This local database can store various data in real time during the machining process, including machining parameters, sensor data, deviation data, tool wear data, and prediction results. Through the organization, analysis, and learning of historical data, it continuously optimizes the fusion algorithm parameters of the adaptive decision engine and the weight coefficients of machining parameters, and improves the workpiece material database and machining parameter database. This allows the intelligent sensing-based CNC precision machining control system to gradually adapt to the machining needs of workpieces of different materials and specifications, improving the system's generalization ability and stability. It can adapt to new machining scenarios without large-scale system modifications, reducing equipment upgrade costs.

[0067] This invention reduces data transmission latency and improves system response speed. The edge computing layer of the intelligent sensing-based CNC precision machining control system is equipped with a lightweight chip, enabling local real-time processing of raw data acquired by the intelligent sensing layer. This includes noise reduction, feature extraction, and local inference, eliminating the need to transmit raw data to the cloud. This avoids machining deviations caused by cloud transmission lag, keeps data processing and decision-making delays within a reasonable range, ensures real-time optimization of machining parameters and deviation correction, and further improves machining accuracy and efficiency.

[0068] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A precision machining control system for CNC machines based on intelligent sensing, characterized in that: The system comprises an intelligent sensing and acquisition layer, an edge computing layer, an intelligent decision-making layer, an execution layer, and a feedback layer, which are electrically connected sequentially to form a closed-loop control architecture. The intelligent sensing and acquisition layer integrates a piezoelectric force sensor, a high-precision temperature sensor, an acoustic emission sensor, and a laser displacement sensor. It also features an anti-electromagnetic interference packaging structure to isolate interference signals in industrial environments. The edge computing layer carries a lightweight chip for real-time noise reduction, feature extraction, and local inference of the data acquired by the intelligent sensing and acquisition layer. The intelligent decision-making layer has an adaptive decision engine that uses a fusion algorithm to dynamically optimize cutting parameters based on the machining material and tool wear status. The execution layer includes a high-precision servo drive module and an intelligent tool adjustment mechanism to respond to parameter commands output by the intelligent decision-making layer. The feedback layer, through visual inspection and digital twin simulation linkage, detects machining deviations in real time and feeds them back to the intelligent decision-making layer.

2. The CNC precision machining control system based on intelligent sensing according to claim 1, characterized in that: The electromagnetic interference-resistant packaging structure is used to solve the problem of power frequency interference and interference of servo motor pulse width modulation signals on small signals in industrial scenarios. The sensors in the intelligent sensing and acquisition layer synchronously collect various data during the processing.

3. The CNC precision machining control system based on intelligent sensing according to claim 1, characterized in that: The lightweight chip in the edge computing layer can control data processing and decision-making delays within a preset range, avoiding processing deviations caused by cloud transmission lag, and enabling rapid local data processing.

4. The CNC precision machining control system based on intelligent sensing according to claim 1, characterized in that: The fusion algorithm of the intelligent decision layer is a fusion algorithm of long short-term memory network and deep Q network, which can overcome the deficiency of insufficient generalization ability of single algorithm and realize dynamic adaptive adjustment of processing parameters.

5. The CNC precision machining control system based on intelligent sensing according to claim 1, characterized in that: The system also includes a local database, which is electrically connected to the intelligent decision-making layer and is used to store processing parameters, sensor data, and deviation data, providing parameter support for the processing of similar workpieces in the future and enabling the system to learn and iterate.

6. A precision machining control method for CNC machines based on intelligent sensing, characterized in that: The method, applicable to the CNC precision machining control system based on intelligent sensing as described in any one of claims 1 to 5, comprises the following steps: First, sensor data acquisition and preprocessing: Cutting force, spindle vibration, machining temperature, tool wear, and workpiece dimensional deviation data are synchronously acquired by sensors in the intelligent sensing acquisition layer. Wavelet threshold noise reduction algorithm is used to process the acquired data, eliminating electromagnetic interference and mechanical noise while retaining minor fault and deviation features. Second, feature recognition and state assessment: Core features such as tool wear, spindle operating status, and workpiece dimensional deviation are extracted through an edge computing layer. An improved long short-term memory network is used to predict the remaining tool life. Third, adaptive parameter... The first step involves several steps: First, data optimization. By combining a workpiece material database with real-time machining status, a deep Q-network algorithm dynamically adjusts machining parameters such as feed rate, spindle speed, and depth of cut to address tolerance fluctuations caused by manual experience-based parameters. Second, real-time execution and deviation correction. The execution layer executes the optimized machining parameters, and a laser displacement sensor detects workpiece dimensional deviations in real time. This deviation data is fed back to the intelligent decision layer, which generates compensation commands to drive the intelligent tool adjustment mechanism for online correction. Third, full lifecycle data accumulation. Machining parameters, sensor data, and deviation data are stored in a local database to provide parameter support for subsequent machining of similar workpieces, enabling the system to self-learn and iterate.

7. The CNC precision machining control method based on intelligent sensing according to claim 6, characterized in that: The wavelet threshold denoising algorithm in the first step is an original design that can accurately remove various interference signals, ensuring the accuracy and effectiveness of the extracted processing data.

8. The CNC precision machining control method based on intelligent sensing according to claim 6, characterized in that: The feature recognition process in the second step is completed at the edge computing layer, eliminating the need to transmit the raw data to the cloud and shortening the response time for feature extraction and status assessment.

9. The CNC precision machining control method based on intelligent sensing according to claim 6, characterized in that: The online correction process in the fourth step achieves seamless integration of data collection, analysis, decision-making, and correction, ensuring that processing deviations can be eliminated in real time and improving processing accuracy.

10. The CNC precision machining control method based on intelligent sensing according to claim 6, characterized in that: The data accumulation process in the fifth step enables continuous optimization of processing parameters, allowing the system to adapt to the processing needs of workpieces of different materials and specifications, and improving the system's generalization ability.

Images