Piston-cylinder intelligent polishing equipment integrated with online roughness detection

Abstract

This invention relates to the field of grinding technology and discloses an intelligent grinding device for piston cylinders integrating online roughness detection. The device includes: a multi-source sensor signal fusion and truth arbitration module, a constant force grinding adaptive control module, a vibration active suppression and multi-axis collaborative module, and a grinding wheel state soft measurement and compensation decision module. A redundant sensing architecture composed of virtual observers and physical sensors is used for cross-verification and fault tolerance, outputting a reliable process dataset. A dual-closed-loop force control architecture is used to achieve adaptive planning and precise tracking of the target normal force. Adaptive notch filtering and cross-coupled contour compensation are used to suppress chatter and ensure surface accuracy. Compensation commands are sent through a multi-feature fusion passivation index hierarchical force control loop and a vibration suppression loop, forming a cross-module collaborative closed loop. This invention achieves system-level intelligent grinding from signal reliability assurance, dynamic process optimization, active vibration suppression to on-demand grinding wheel warning.

Description

Technical Field

[0001] This invention relates to the field of grinding and machining technology, specifically to an intelligent grinding device for piston cylinders that integrates online roughness detection. Background Technology

[0002] As a core component of power machinery such as engines and compressors, the piston cylinder block's inner wall surface quality directly affects sealing performance, friction and wear characteristics, and the overall service life of the machine. Grinding, a key finishing process in cylinder block manufacturing, requires achieving the designed surface roughness while ensuring dimensional accuracy, and also imposes strict requirements on the stability and consistency of the processing. With the continuous improvement in the efficiency and reliability requirements of power systems in the automotive and general machinery industries, cylinder block grinding technology is evolving from traditional manual or semi-automatic operations towards a fully automated process.

[0003] However, existing devices still have some problems in use, specifically as follows: sensor signals are directly accepted without effective verification and fault tolerance mechanisms; process parameters rely on factory settings or manual experience for tuning, lacking real-time adaptive adjustment capabilities based on the physical state of the grinding process; during grinding, the fluctuation of grinding force and the modal coupling of the equipment structure can easily cause regenerative chatter, and existing equipment mostly adopts passive methods such as reducing the rotational speed or feed rate to avoid chatter, sacrificing processing efficiency with limited effect, lacking means to actively suppress vibration energy at the control loop level; the timing of grinding wheel replacement depends entirely on the operator's experience or is forced to be replaced according to a fixed number of processed parts, lacking online assessment and graded early warning methods based on the actual deterioration state of the grinding wheel. Summary of the Invention

[0004] This invention provides an intelligent grinding device for piston cylinders that integrates online roughness detection. It has the advantages of sensor signal redundancy verification and fault tolerance, adaptive dynamic optimization of process parameters, active vibration suppression and multi-axis collaborative control, and online soft measurement and graded compensation of grinding wheel status, thus solving the problems mentioned in the background art.

[0005] This invention provides the following technical solution: an intelligent grinding device for piston cylinders integrating online roughness detection, comprising a support frame, a control component mounted on the left side of the support frame, a first moving mechanism fixedly mounted on the top of the support frame, a second moving mechanism mounted above the first moving mechanism, a third moving mechanism mounted in front of the second moving mechanism, a grinding mechanism mounted inside the third moving mechanism, a detection mechanism for online roughness detection mounted in front of the third moving mechanism, and a feeding mechanism mounted on the right side of the support frame. The feeding mechanism includes a fixed frame, a translation component, and a translation gripping component. The fixed frame is fixedly mounted on the right side of the support frame, the translation component is mounted on the right side of the fixed frame, and the translation gripping component is mounted in front of the translation component.

[0006] In a preferred embodiment, the control element (2) includes: The multi-source sensor signal fusion and truth arbitration module is used to form a redundant sensing architecture with virtual observers and physical sensors, cross-verify the grinding normal force and surface roughness, and replace the sensor with a virtual signal when the sensor fails, outputting a reliable process dataset. The constant force grinding adaptive control module is used to construct a dual closed-loop control architecture with reliable normal force as feedback, and outer loop target force planning and inner loop force tracking. The outer loop target force planning adaptively corrects the target normal force according to the real-time roughness deviation and grinding wheel wear trend. The vibration active suppression and multi-axis coordination module is used to perform three-axis linkage trajectory planning and contour error compensation control, and uses the vibration main frequency as the tracking target. A notch filter with adaptive center frequency is connected in series at the output of the position loop to suppress regenerative flutter. The grinding wheel condition soft measurement and compensation decision module is used to extract three characteristic quantities: unit normal force removal efficiency, root mean square of high frequency component of force signal, and contact frequency band vibration energy. It integrates and calculates the comprehensive passivation index, and sends compensation commands to the constant force grinding adaptive control module and the vibration active suppression module according to the passivation index.

[0007] The present invention has the following beneficial effects: 1. This invention constructs a virtual normal force observer using a motor current-torque mapping model and generates a theoretical roughness expectation value using a roughness-process parameter regression model. This value forms a redundant sensing pair with the measured values ​​of physical sensors. Furthermore, it introduces the root mean square value of vibration as a dynamic deviation threshold for cross-arbitration. In the event of sensor failure, the virtual signal is temporarily used to maintain the uninterrupted control loop. This solves the problem of existing equipment directly accepting sensor signals, lacking verification and fault tolerance mechanisms, and being prone to misjudgment due to sensor drift or contamination. This ensures the reliability of the input data of the control system.

[0008] 2. This invention constructs a dual closed-loop control architecture of outer-loop target force planning and inner-loop force tracking. The outer loop adaptively corrects the target normal force based on real-time roughness deviation and grinding wheel wear trend, while the inner loop achieves precise force tracking through a composite control law of proportional-integral-derivative plus feedforward. This solves the problem of existing equipment having fixed process parameters and being unable to dynamically adjust according to the attenuation of grinding wheel cutting ability, thus keeping material removal efficiency and surface quality stable throughout the entire life cycle of the grinding wheel.

[0009] 3. This invention achieves smooth trajectory planning for the cylinder block profile by using a parameterized B-spline path and a seven-segment S-curve acceleration / deceleration. It compensates for contour errors caused by differences in multi-axis dynamic response through a cross-coupled controller, and actively cuts off the regenerative excitation path of chatter energy by connecting a notch filter with adaptive center frequency in series at the output of the position loop with the vibration main frequency as the tracking target. This solves the problem that existing equipment lacks active vibration suppression methods and passively reduces speed to avoid chatter and sacrifice efficiency. It effectively suppresses the generation of surface grooves while ensuring processing efficiency.

[0010] 4. This invention calculates a comprehensive passivation index by integrating three characteristic quantities: unit normal force removal efficiency, root mean square of high-frequency force signal components, and contact frequency band vibration energy. It then implements a four-level graded compensation and early warning strategy from the sharpness period to the end of the lifespan, transforming the degradation trend into a quantitative closed-loop correction of the force control loop gain, vibration suppression threshold, and force safety boundary in real time. This solves the problem that existing equipment relies on manual experience for grinding wheel replacement and is difficult to balance consumable utilization and surface quality. It achieves intelligent maintenance decision-making with on-demand early warning and multi-module collaborative evolution. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of the overall structure of the present invention; Figure 2 This is a partial structural diagram of the entire invention; Figure 3 This is a schematic diagram of the feeding mechanism of the present invention; Figure 4 This is a schematic diagram of the module connection structure of the present invention; In the figure: 1. Support frame; 2. Control component; 3. First moving mechanism; 4. Second moving mechanism; 5. Third moving mechanism; 6. Grinding mechanism; 7. Detection mechanism; 8. Feeding mechanism; 80. Fixing frame; 81. Translation component; 82. Translation gripping component. Detailed Implementation

[0012] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The intelligent grinding equipment for piston cylinders with integrated online roughness detection involved in the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] Please see Figures 1-4 The intelligent grinding equipment for piston cylinders with integrated online roughness detection shown includes a support frame 1, a control component 2 installed on the left side of the support frame 1, the support frame 1 provides stable support for the whole machine, and the control component 2 realizes intelligent collaborative control to improve the level of automation; It also includes a first moving mechanism 3, which is fixedly installed on the top of the support frame 1. A second moving mechanism 4 is installed above the first moving mechanism 3, and a third moving mechanism 5 is installed in front of the second moving mechanism 4. The first moving mechanism 3, the second moving mechanism 4, and the third moving mechanism 5 are all guided and translated by guide rails to ensure smooth and accurate multi-axis movement and effectively improve the positioning accuracy of grinding and inspection. The grinding mechanism 6 is installed inside the third moving mechanism 5. The grinding mechanism 6 is driven by a synchronous belt motor, which ensures smooth power transmission and good grinding consistency. The detection mechanism 7 for online roughness detection is installed in front of the third moving mechanism 5. It can detect the surface roughness of the piston cylinder in real time, provide timely feedback on processing quality, and reduce defective products. It also includes a feeding mechanism 8, which is installed on the right side of the support frame 1. The feeding mechanism 8 includes a fixed frame 80, a translation component 81, and a translation gripping component 82. The fixed frame 80 is fixedly installed on the right side of the support frame 1 to provide stable support. The translation component 81 is installed on the right side of the fixed frame 80 and is guided by a guide rail for translation, ensuring accurate feeding stroke. The translation gripping component 82 is installed on the front side of the translation component 81. The translation gripping component 82 is guided by the guide rail and clamped by the gripper, which ensures reliable gripping and high material changing efficiency.

[0014] During operation, the translation component 81 and the translation gripping component 82 of the feeding mechanism 8 work together. The translation gripping component 82 is guided by the guide rail and firmly clamps the piston cylinder to be processed through the grippers. The translation component 81 moves smoothly along the guide rail, accurately transporting the workpiece to the clamping position and reliably clamping it. Subsequently, the first moving mechanism 3, the second moving mechanism 4, and the third moving mechanism 5 work together in an orderly manner under the guidance of their respective guide rails to adjust the spatial position of the grinding mechanism 6 and the detection mechanism 7 along the X, Y, and Z directions, so that the grinding tool and the detection probe are precisely positioned above the surface of the cylinder to be processed. The grinding mechanism 6 then operates at a constant speed under the drive of the synchronous belt motor, and performs uniform and stable grinding on the cylinder surface. At the same time, the detection mechanism 7 installed in front of the third moving mechanism 5 moves with it to perform online roughness detection on the grinding area, collect surface quality data in real time and quickly feed it back to the control component 2. The control unit 2 intelligently analyzes the test results and dynamically adjusts parameters such as grinding pressure, feed speed, and grinding time to form a closed-loop control until the surface roughness index fully meets the preset requirements. After processing, the grinding mechanism 6 and the testing mechanism 7 retract, the three-way moving mechanism resets, and the translation gripping component 82 moves again to remove the finished cylinder body, completing a full-process intelligent grinding operation.

[0015] The control unit (2) uses a multi-core heterogeneous embedded processor as its core hardware platform, integrating an EtherCAT real-time industrial bus master module, a multi-channel synchronous analog signal acquisition unit, and a large-capacity process database storage unit. It features a 16-bit precision force sensor signal conditioning interface, a high-speed differential input interface for a laser displacement sensor, an accelerometer IEPE constant current source power supply interface, and a multi-channel servo motor encoder feedback interface. It can efficiently adapt to heterogeneous peripherals such as roughness probes, force sensors, triaxial accelerometers, laser displacement sensors, spindle frequency converters, and servo drives for each axis, meeting the requirements for high-frequency synchronous acquisition of multiple physical quantities and strong real-time control command issuance during the grinding process. The software system running on this hardware platform adopts a strict modular design, including: a multi-source sensor signal fusion and truth arbitration module, a constant force grinding adaptive control module, a vibration active suppression and multi-axis coordination module, and a grinding wheel state soft measurement and compensation decision module. The modules interact with each other in real time through shared memory and high-speed data bus, forming a complete intelligent closed loop from "reliable identification of original signals, closed-loop control of core processes, directional suppression of vibration energy to compensation for tool deterioration trend", so as to achieve consistent surface quality of cylinder grinding and adaptive evolution of process parameters.

[0016] The multi-source sensor signal fusion and truth arbitration module achieves online diagnosis of sensor status and fault-tolerant operation under abnormal conditions through multi-dimensional signal cross-verification, ensuring the absolute reliability of the data input to the control system. The specific implementation includes the following steps: S101: Redundant Observation and Cross-validation of Grinding Normal Force A high-rigidity strain gauge force sensor is installed on the spindle housing of the grinding mechanism (6) to directly measure the grinding normal force. The sampling frequency is 1kHz.

[0017] Basis and method of acquisition Basis for measurement: The sensor is based on the resistance strain effect. The deformation of an elastic element under force causes a change in the resistance of a strain gauge attached to it. This change in resistance is converted into a measurable differential voltage signal via a Wheatstone bridge. The voltage amplitude is linearly related to the applied normal force. The upper limit of the sensor's measurement range is no less than 1.5 times the maximum permissible normal force, and the overall accuracy is no less than ±0.5%FS.

[0018] Acquisition method: The sensor is fixed to the spindle housing of the grinding mechanism (6) by flange mounting, and the force direction marking is strictly consistent with the axis direction of the grinding tool. The differential voltage signal output by the sensor is transmitted to the analog signal acquisition unit of the control component (2) via a shielded cable. After differential amplification and low-pass filtering by the instrumentation amplifier, it is converted into an analog-to-digital signal by the ADC module at a sampling frequency of 1kHz. Then, based on the sensitivity coefficient and zero-point offset obtained from the installation calibration, it is converted into the measured value of the normal force. .

[0019] Meanwhile, the module internally uses real-time current feedback from the servo motor driver. With spindle encoder speed feedback Based on this, the virtual tangential force is calculated in real time using a calibrated motor output torque-force mapping model:

[0020] in, The motor torque constant (N·m / A) is verified from the parameters on the motor nameplate. The no-load current (A) is measured at different speeds before the equipment leaves the factory, and a one-dimensional lookup table indexed by speed is established. When online, it is based on real-time speed feedback. Interpolate to obtain the no-load current value corresponding to the current speed; The efficiency of the transmission mechanism is dimensionless and was determined to be 0.93 by bench calibration test. Let be the radius (m) of the grinding tool.

[0021] The above virtual tangential force This represents the tangential grinding force at the contact point between the grinding wheel and the workpiece, reflecting the real-time load state of the motor drive system. To enable redundant comparison between this virtual observation signal and the measured normal force value from the physical force sensor, the module introduces a normal force-tangential force ratio coefficient calibrated through grinding tests under the same working conditions. ( Its value is determined by averaging the force measured under at least three sets of force benchmark tests with different grinding parameters (and stored in the process database), converting the virtual tangential force into an estimated virtual normal force value:

[0022] Therefore, the virtual tangential force Compared with the measured value of the solid force sensor This constitutes a redundant force sensing pair.

[0023] S102: Comparison of process model and actual measurement of surface roughness The stylus-type roughness detection mechanism (7) installed on the front side of the third moving mechanism (5) outputs real-time roughness values ​​at a sampling frequency of 150Hz. The module synchronously calculates the theoretical expected roughness value based on an empirically calibrated roughness-process parameter regression model.

[0024] in, This represents the current synthesis feed rate (m / min). is the current normal force (N), and n is the spindle speed (r / min); The roughness coefficient is related to the workpiece material and the grinding wheel grit size. The four parameters, which are the influence indices of each process variable, were determined by nonlinear least squares fitting after conducting a 5-factor, 3-level orthogonal grinding test on the same batch of cylinder block materials. The results are then stored in the workpiece material process database of the control components.

[0025] and This constitutes a roughness redundancy sensing pair.

[0026] S103: Vibration signal acquisition and feature extraction A triaxial accelerometer is installed on the spindle housing of the grinding mechanism (6) to synchronously acquire vibration acceleration signals in the X, Y, and Z directions at a sampling frequency of 2kHz. After being powered by an IEPE constant current source and undergoing signal conditioning, the signal enters the analog acquisition channel. The module performs the following processing on the raw acceleration signal: The three-axis signals are each subjected to a low-pass filter with a cutoff frequency of 500Hz to remove high-frequency noise, resulting in the filtered signals. .

[0027] Calculate the root mean square value of the combined triaxial acceleration as a comprehensive indicator of the current vibration intensity:

[0028] Where N is the number of sampling points in the sliding window, the window width is 100ms (i.e., 200 sampling points), and the window is updated every 10ms.

[0029] Synthetic acceleration signal Perform a short-time Fourier transform (STFT) to search for the frequency component with the largest amplitude within the 0-500Hz frequency band and extract the dominant vibration frequency. (Hz) and its corresponding amplitude (m / s) 2 ).

[0030] This forms the vibration characteristic vector. ,in Reflects the total energy level of vibration. Indicates the frequency of the currently excited structurally weak modes. It reflects the intensity of the excitation of this mode.

[0031] S104: Multi-signal joint arbitration and adaptive fault tolerance The module uses deviation statistics within a sliding time window (100ms wide) as a basis to perform dynamic threshold arbitration for each redundant sensing pair. For the normal force channel, an adaptive deviation threshold is defined:

[0032] in, The historical average (N) of the force signal deviation over the last 5 seconds. The corresponding historical standard deviation (N); The root mean square value of the current triaxial acceleration (m / s²) 2 ), The vibration reference value during normal and stable grinding (m / s) 2 (The threshold is determined by data collected during the equipment commissioning phase). When vibration intensifies, the threshold is adaptively widened to avoid misjudgments triggered by fluctuations in operating conditions.

[0033] The arbitration logic is as follows: If only the force sensor is deviating Exceeding roughness deviation It did not exceed its corresponding threshold. (If the construction method is similar), then it is determined that the force sensor has malfunctioned or drifted, and the module is automatically isolated. With virtual power It replaces the calculation of subsequent control loops and outputs a warning signal encoded as "FORCE_SENSOR_FAULT" through the communication interface, while the other functions of the equipment operate normally without interruption; If only roughness deviation Exceeding If the force deviation is normal, then the roughness detection mechanism (7) is determined to be abnormal, based on the model roughness. Replace it and issue a "ROUGHNESS_PROBE_FAULT" warning; If both force and roughness deviation exceed the limit, or acceleration If a sudden increase of more than 50% occurs without any change in process instructions, it is determined to be a "serious process abnormality of the equipment body" (such as grinding wheel breakage or workpiece loosening), generating an interlock signal coded "CRITICAL_PROCESS_ANOMALY", triggering rapid retraction and spindle emergency braking.

[0034] Credibility value after arbitration Confidential roughness value and vibration characteristics (in The dominant frequency of vibration, The main frequency amplitude is packaged into a trusted process dataset and synchronously distributed to subsequent modules via a high-speed bus.

[0035] The constant force grinding adaptive control module: uses the arbitrated reliable normal force Using the physical process of workpiece material removal as the control basis, a dual closed-loop force control architecture is constructed to achieve precise control of the amount of material removed from the cylinder surface and over-wear protection. The specific implementation includes the following steps: S201: Outer Ring Target Force Planning and Adaptive Correction The outer ring is responsible for dynamically planning the target normal force for the current pass based on the target roughness requirements and the real-time grinding status. Initial target force Based on a pre-set process database, the parameters are determined by the cylinder block material, grinding wheel specifications, and target surface roughness. The table provides the information. The module performs the following adaptive corrections within each force control cycle (2ms cycle):

[0036] The first correction term is the surface quality closed-loop term: The force-roughness gain coefficient (N / μm) is determined by a stepped force grinding test on the same material; when Higher than Increase the target force appropriately to increase the removal amount and improve the roughness.

[0037] The second correction item is the grinding wheel wear compensation item: The volume of material removed per unit time (mm) 3 / s), by Online calculation, The material removal coefficient; The baseline removal rate of the new grinding wheel under the same parameters; when Significantly lower due to grinding wheel passivation At the same time, the target force is gradually increased to compensate for the decrease in cutting capability. The wear compensation gain (N) was calibrated through a full-life-cycle wear test of the grinding wheel.

[0038] It needs to be limited, and the upper limit is the maximum allowable normal force that will not cause the workpiece to burn or elastically deform.

[0039] S202: Inner Loop Force Tracking and Force-Position Coordination Control The inner loop, using a control frequency of 2kHz, adjusts the Z-axis position command of the third moving mechanism (5) through a composite control law of PID + feedforward, so that the actual normal force... Precisely follow the target force The control law is:

[0040] in, The Z-axis position fine adjustment amount (mm). Force deviation (N). , , , The PID gain is determined through system identification and gradient optimization tuning.

[0041] feedforward term Used to compensate for changes in cylinder block surface height caused by cumulative material removal. To reduce the tracking lag of the force ring, among which This represents the grinding contact area.

[0042] Among them, the grinding contact area The basis and method for obtaining the information are as follows: Basis for obtaining information: Grinding contact area The basis for obtaining this value is the principle of grinding contact geometry. In grinding, the contact area between the grinding wheel and the workpiece is not the entire surface of the grinding wheel, but rather the product of the local contact arc length and the grinding width, which is determined by the grinding depth and the grinding wheel diameter. ,in For the contact arc length, This refers to the grinding width.

[0043] Method of obtaining: Contact arc length According to the grinding contact arc length formula Calculation, where The grinding depth (mm) is obtained in real time from the actual feed position of the Z-axis of the third moving mechanism (5); The grinding wheel diameter (mm) is read from the process database of the control unit (2) based on the currently installed grinding wheel specifications. Grinding width The grinding contact area A is determined based on the grinding wheel width and cylinder profile geometry parameters, and also derived from the workpiece size information and grinding wheel specifications in the process database. These parameters are obtained in real-time and input into the formula to calculate the grinding contact area A under the current working conditions online.

[0044] When force deviation When the value remains positive and the Z-axis position has reached the hard limit (indicating severe wear of the grinding wheel, which cannot reach the target force even if it makes contact), the module automatically triggers a grinding wheel dressing or replacement prompt and switches the force ring to position holding mode to prevent damage to the mechanism.

[0045] The vibration active suppression and multi-axis coordination module is responsible for the three-axis linkage trajectory planning and high-precision tracking control of the first moving mechanism (X-axis), the second moving mechanism (Y-axis), and the third moving mechanism (Z-axis), and integrates vibration active suppression function to limit the regenerative chatter energy caused by grinding force fluctuations and structural modal coupling during grinding to a safe level. The specific implementation includes the following steps: S301: Adaptive trajectory interpolation oriented towards cylinder block profile For characteristic surfaces such as the cylindrical surface and transition arcs of the cylinder block, a parametric B-spline path description is used. Within each interpolation cycle (250 μs), the module calculates the surface curvature radius based on the characteristic surface curvature radius. With programmed feed rate Adaptive speed adjustment is performed to prevent excessive centripetal acceleration from causing contour errors.

[0046] in, The maximum permissible normal acceleration of the machine tool (m / s²) 2 ).

[0047] A continuous velocity profile is generated by using a seven-segment S-curve acceleration and deceleration to ensure that the acceleration is continuously limited and to avoid the inherent modes of the impact excitation structure.

[0048] S302: Real-time compensation for cross-coupled contour errors To suppress cylinder block surface contour errors caused by differences in multi-axis dynamic response, an online contour error estimator is constructed. For a cylindrical surface grinding path, the instantaneous contour error is... The calculation is approximately based on the tangential tracking error. The module introduces a cross-coupled controller in the velocity loop, through coupling gain... The estimated contour error is compensated for by the speed commands on the X and Y axes:

[0049] in, This is the path tangent angle at the current position. This mechanism increases the decay rate of contour errors to 2-3 times that of traditional uncoupled control, keeping the target within ±12μm.

[0050] S303: Adaptive Notch Filter for Flutter Frequency Regenerative flutter energy during the grinding process is typically concentrated near the weakest modal frequencies of the structure. The module's vibration frequency is determined by a triaxial accelerometer. To track the target in real time, a second-order notch filter with an adaptive center frequency is connected in series at the output of the position loop of the third moving mechanism (Z-axis). Its z-domain transfer function is:

[0051] in, The control period is 250 μs. This is the notch width factor (0.995 is chosen to balance selective attenuation and phase loss).

[0052] When detected The notch filter automatically activates when the flutter warning threshold is exceeded; when When the temperature drops below the safety line and remains there for more than 30 cycles, the notch filter smoothly exits. This solution cuts off the regenerative excitation path of flutter energy at the control algorithm level, which is more economical and efficient than simply reducing the speed or feed.

[0053] The grinding wheel condition soft measurement and compensation decision module utilizes existing multi-source signals such as force, vibration, and power from the control system to achieve online soft measurement of the grinding wheel's wear degree and sharpness through physical feature fusion. Based on the deterioration trend, it outputs an active compensation strategy or a replacement warning. The specific implementation includes the following steps: S401: Extraction of Grinding Wheel Wear Trend Features Three measurable features strongly correlated with the grinding wheel condition are extracted to construct a wear observation vector: Unit normal force removal efficiency (mm) 3 / (s·N)), this value decreases monotonically with the passivation of the grinding wheel; Root mean square of high frequency component of force signal The high-frequency fluctuation component of the force signal is obtained by performing a 100Hz high-pass filter on the force sensor signal. and calculate This value increases as abrasive fragmentation intensifies; The integral energy of the vibration power spectral density in the grinding wheel-workpiece contact frequency band (4-8kHz) : S402: Calculation of the passivation index for multi-feature fusion The three features mentioned above are normalized and weighted, and then combined to calculate the comprehensive passivation index:

[0054] in, The initial removal efficiency of the new grinding wheel. This represents the removal efficiency at the end of the grinding wheel's life. This is a reference value for the high-frequency force component at the end of the grinding wheel's life. This serves as a reference for the contact frequency band vibration energy under normal conditions. This is a reference value for the contact frequency band vibration energy at the end of the grinding wheel's life. , , The weighting coefficients are determined by linear discriminant analysis of the full life bench test of the grinding wheel, and satisfy the condition that the sum is 1.

[0055] Of the three terms above: the first term represents the percentage decrease in unit normal force removal efficiency relative to the initial value. When the grinding wheel is sharp, the numerator approaches 0, and when the lifespan ends, the numerator equals the denominator, with a value that monotonically increases within the interval [0,1]. The second term represents the ratio of the root mean square of the high-frequency component of the force signal to the limit value at the end of the lifespan, with a value within the interval [0,1]. The third term represents the ratio of the increment of contact band vibration energy relative to the normal reference value to the maximum increment at the end of the lifespan. When the grinding wheel is sharp, the numerator approaches 0, and when the lifespan ends, the numerator equals the denominator, with a value within the interval [0,1]. Therefore, after weighted summation, The value range is [0,1], where 0 represents complete sharpness and 1 represents complete passivation, thus achieving a dimensionless representation of the passivation degree.

[0056] S403: Hierarchical Compensation Decision-Making and Multi-Module Collaborative Closed Loop according to The module executes a progressively hierarchical compensation strategy based on the real-time value, and synchronously distributes the compensation commands to the constant force grinding adaptive control module and the vibration active suppression module, forming a cross-module collaborative closed-loop response: Level 1 Response (Sharp Period): When When the value is less than 0.3, the grinding wheel is considered to be in a sharp state, requiring no additional intervention. The wear compensation gain in the constant force grinding adaptive control module... Maintaining the baseline value, the flutter warning threshold of the active vibration suppression module remains at the default setting. The module continues to record the current... , and As a dynamic benchmark for subsequent comparisons.

[0057] Secondary response (normal wear period): when 0.3≤ When the value is less than 0.6, the grinding wheel is determined to have entered the normal wear stage, and its cutting ability begins to decrease at a predictable rate. The module automatically performs the following compensation actions: Send a compensation command to the constant force grinding adaptive control module to increase the wear compensation gain. according to Linear adjustment gradually enhances the wear compensation term in the outer ring target force planning, stabilizing material removal efficiency; Send parameter adjustment instructions to the vibration active suppression module to appropriately lower the activation threshold of the adaptive notch filter by 10% in order to cope with the increase in vibration trend caused by grinding wheel passivation and intervene in flutter boundary management in advance.

[0058] Level 3 Response (Severe Wear-Out Period): When When the value is ≥0.6, the grinding wheel is determined to be severely worn, and parameter compensation alone is insufficient to maintain process quality; the module executes the following power limit protection and warning actions: Send a force upper limit constraint command to the constant force grinding adaptive control module. Reduce to 85% of the rated value to prevent thermal damage to the workpiece surface or elastic deformation of the thin wall of the cylinder caused by excessive increase in normal force. The human-machine interface of the control unit (2) outputs a "grinding wheel replacement warning" signal to prompt the operator to replace the grinding wheel in a planned manner after the current processing batch is completed; The current machining cylinder number and the overall passivation index are used. Record it in the process traceability database and mark it as "grinding wheel end section processed part" so that subsequent quality inspection can focus on its surface integrity.

[0059] Level 4 Response (End-of-Life Forced Shutdown): When When the value is ≥0.85, the grinding wheel is determined to have reached the end of its service life, and continuing processing poses an unacceptable risk of workpiece scrapping or equipment damage; the module executes the following mandatory protection actions: Immediately interrupt the closed-loop output of the constant force grinding adaptive control module and force the Z-axis command of the third moving mechanism (5) to switch to a rapid retraction sequence, so that the grinding tool can leave the workpiece surface within a preset safe time. Lock the spindle start signal to prevent the equipment from re-entering the automatic grinding cycle before the grinding wheel replacement confirmation signal is reset. The human-machine interface outputs a continuously flashing "forced grinding wheel replacement" alarm, and simultaneously records all process data at the moment of shutdown. Final value, cumulative number of parts processed by this grinding wheel, and cumulative total material removed) are used for statistical analysis of the entire life cycle of the grinding wheel.

[0060] Through the aforementioned four-level progressive hierarchical response mechanism, this module not only completes online measurement and early warning of the grinding wheel's condition, but also utilizes the passivation index, which characterizes the deterioration trend. This is transformed in real time into quantitative closed-loop correction of the constant force control loop gain, vibration suppression threshold, and force safety boundary, enabling the control strategies of the first three modules to sense changes in the grinding wheel state and adjust collaboratively. As a result, the four modules no longer operate independently, but instead achieve a system-level intelligent closed loop from "detection-control-suppression-compensation" to "perception-fusion-linkage-evolution" through the core time-varying factor of grinding wheel deterioration, which runs through the entire grinding process.

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

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

Claims

1. A piston cylinder intelligent grinding device integrating online roughness detection, comprising a support frame (1), a control component (2) mounted on the left side of the support frame (1), and a first moving mechanism (3), the first moving mechanism (3) being fixedly mounted on the top of the support frame (1), a second moving mechanism (4) mounted above the first moving mechanism (3), a third moving mechanism (5) mounted in front of the second moving mechanism (4), and a grinding mechanism (6) mounted inside the third moving mechanism (5), for online roughness detection. The detection mechanism (7) is installed on the front side of the third moving mechanism (5), and also includes a feeding mechanism (8). The feeding mechanism (8) is installed on the right side of the support frame (1). The feeding mechanism (8) includes a fixed frame (80), a translation component (81), and a translation gripping component (82). The fixed frame (80) is fixedly installed on the right side of the support frame (1). The translation component (81) is installed on the right side of the fixed frame (80). The translation gripping component (82) is installed on the front side of the translation component (81). The control element (2) includes: The multi-source sensor signal fusion and truth arbitration module is used to form a redundant sensing architecture with virtual observers and physical sensors, cross-verify the grinding normal force and surface roughness, and replace the virtual signal with a reliable process dataset when the sensor fails. The constant force grinding adaptive control module is used to construct a dual closed-loop control architecture with reliable normal force as feedback, and outer loop target force planning and inner loop force tracking. The outer loop target force planning adaptively corrects the target normal force according to the real-time roughness deviation and grinding wheel wear trend. The vibration active suppression and multi-axis coordination module is used to perform three-axis linkage trajectory planning and contour error compensation control, and uses the vibration main frequency as the tracking target. A notch filter with adaptive center frequency is connected in series at the output of the position loop to suppress regenerative flutter. The grinding wheel condition soft measurement and compensation decision module is used to extract three characteristic quantities: unit normal force removal efficiency, root mean square of high frequency component of force signal, and contact frequency band vibration energy. It integrates and calculates the comprehensive passivation index, and sends compensation commands to the constant force grinding adaptive control module and the vibration active suppression module according to the passivation index.

2. The intelligent piston cylinder grinding equipment integrating online roughness detection according to claim 1, characterized in that, The multi-source sensor signal fusion and truth arbitration module includes: The actual value of the grinding normal force is measured by a strain gauge force sensor installed on the spindle housing of the grinding mechanism. At the same time, based on the real-time current feedback of the servo motor driver and the speed feedback of the spindle encoder, the virtual tangential force is calculated in real time through the mapping relationship between the motor output torque and the grinding tangential force. Then, the virtual normal force is estimated by converting it into a normal force-tangential force ratio coefficient, thus forming a redundant force sensing pair. The surface roughness is measured by a stylus-type roughness detection mechanism installed on the front side of the third moving mechanism. At the same time, a theoretical roughness expectation value is generated based on a pre-calibrated roughness and process parameter regression model, forming a roughness redundant sensing pair. The vibration acceleration signals in three directions are simultaneously collected by a triaxial accelerometer installed on the spindle housing of the grinding mechanism. The root mean square value of the composite vibration acceleration is calculated, and the vibration main frequency and its corresponding amplitude are extracted by short-time Fourier transform to form a vibration feature vector. The deviations between the measured values ​​and the corresponding theoretical expected values ​​are calculated for the redundant force sensing pairs and the roughness redundant sensing pairs, respectively. The dynamic deviation threshold, which is introduced as an adaptive factor, is used for arbitration: when the deviation of a single sensor exceeds its corresponding dynamic deviation threshold while the deviations of other sensors do not exceed the limit, the single sensor is determined to be faulty, its measured value is isolated and replaced with the corresponding theoretical expected value, and a sensor fault warning signal is output; when the deviations of multiple sensors exceed the limit simultaneously or the vibration acceleration composite root mean square value increases abruptly without any change in process instructions, the equipment body process is determined to be abnormal, and an interlock signal is generated to trigger rapid retraction and emergency braking of the spindle; the arbitrated reliable normal force, reliable surface roughness, and vibration feature vector are packaged into a reliable process dataset.

3. The intelligent piston cylinder grinding equipment integrating online roughness detection according to claim 1, characterized in that, The constant force grinding adaptive control module includes: Based on the initial target force corresponding to the cylinder material, grinding wheel specifications and target roughness in the pre-set process database, a surface quality closed-loop correction term driven by real-time roughness deviation and a wear compensation term driven by grinding wheel wear trend are superimposed to generate a target normal force after amplitude limiting. The upper limit of the amplitude limiting is the maximum allowable normal force that prevents the workpiece from burning or elastically deforming. Using the force deviation between the target normal force and the reliable normal force as input, the Z-axis position command of the third moving mechanism is adjusted by a composite control law of proportional-integral-derivative plus feedforward, so that the actual normal force follows the target normal force. The feedforward term is used to compensate for the change in cylinder surface height caused by the cumulative removal of material. When the force deviation remains positive and the Z-axis position of the third moving mechanism has reached the hard limit, a grinding wheel replacement prompt is automatically triggered and the force ring is switched to position holding mode.

4. The intelligent piston cylinder grinding equipment integrating online roughness detection according to claim 1, characterized in that, The vibration active suppression and multi-axis coordination module includes: The cylinder block profile is described by a parametric B-spline path. The speed is adaptively adjusted according to the surface curvature radius and the programmed feed rate in each interpolation cycle. A continuous speed profile is generated by using a seven-segment S-curve acceleration and deceleration. The estimated contour error is distributed and compensated to the speed commands of the first and second moving mechanisms by a cross-coupled controller to suppress cylinder surface contour error caused by differences in multi-axis dynamic response. Using the vibration dominant frequency as the real-time tracking target, a notch filter with adaptive center frequency is connected in series at the position loop output of the third moving mechanism. When the amplitude corresponding to the vibration dominant frequency exceeds the preset activation threshold, the notch filter is automatically activated. When the amplitude corresponding to the vibration dominant frequency falls back to below the activation threshold and remains below the preset number of cycles, the notch filter smoothly exits.

5. The intelligent piston cylinder grinding equipment integrating online roughness detection according to claim 1, characterized in that, The grinding wheel condition soft measurement and compensation decision module includes: The reliable normal force and the combined feed rate are obtained from the constant force grinding adaptive control module, and the unit normal force removal efficiency is calculated. The root mean square of the high frequency component of the force signal is calculated after high-pass filtering of the original force sensor signal. The power spectral density of the vibration acceleration signal is estimated, and the integral energy in the contact frequency band between the grinding wheel and the workpiece is calculated, thereby forming the wear observation vector. The unit normal force removal efficiency, the root mean square of the high-frequency component of the force signal and the integral energy in the contact frequency band are normalized and weighted to calculate the comprehensive passivation index. A graded compensation strategy is executed based on the real-time value of the comprehensive passivation index: when the comprehensive passivation index is in the sharp period, the actual initial operating parameters of the grinding wheel are recorded as a dynamic baseline; when entering the normal wear period, an instruction is sent to the constant force grinding adaptive control module to increase the wear compensation gain, and an instruction is sent to the vibration active suppression and multi-axis coordination module to decrease the activation threshold of the notch filter; when entering the severe wear period, an instruction is sent to the constant force grinding adaptive control module to decrease the maximum allowable normal force, and a grinding wheel replacement warning signal is output; when the end of the life is reached, the closed-loop output of the force loop is forcibly interrupted and the spindle start signal is locked, and a forced grinding wheel replacement alarm is output.

6. The intelligent piston cylinder grinding equipment with integrated online roughness detection according to claim 5, characterized in that, During the normal wear period, the process includes: sending a command to the constant force grinding adaptive control module to increase the wear compensation gain, and sending a command to the vibration active suppression and multi-axis coordination module to decrease the activation threshold of the notch filter. Specifically, this includes: linearly increasing the wear compensation gain according to the ratio of the comprehensive passivation index to the upper limit threshold of the normal wear period; and decreasing the activation threshold by a preset percentage based on the current operating value, wherein the decreased activation threshold is not lower than a preset percentage of its initial value.

7. The intelligent piston cylinder grinding equipment integrating online roughness detection according to claim 5, characterized in that, During the severe wear period, the process includes: sending a command to the constant force grinding adaptive control module to reduce the maximum allowable normal force and outputting a grinding wheel replacement warning signal. Specifically, this includes: reducing the maximum allowable normal force to a preset percentage of the original value based on the current operating value to prevent thermal damage to the workpiece surface or elastic deformation of the cylinder thin wall caused by excessive increase in normal force; outputting the grinding wheel replacement warning signal through the human-machine interface of the controller, and recording the number of the currently processed cylinder and the comprehensive passivation index into the process traceability database.

8. The intelligent piston cylinder grinding equipment with integrated online roughness detection according to claim 5, characterized in that, When the lifespan reaches its end, the following steps are taken: forcibly interrupting the closed-loop output of the force loop and locking the spindle start signal, and outputting a forced grinding wheel replacement alarm. Specifically, when the comprehensive passivation index reaches a preset lifespan end threshold, the closed-loop output of the constant force grinding adaptive control module is immediately interrupted, and the Z-axis command of the third moving mechanism is forcibly switched to a rapid retraction sequence, so that the grinding tool is removed from the workpiece surface within a preset safe time; the spindle start signal is locked to prevent the equipment from re-entering the automatic grinding cycle before the grinding wheel replacement confirmation signal is reset; the forced grinding wheel replacement alarm is output through the human-machine interface of the control unit, and all process data at the time of shutdown are recorded simultaneously.

9. The intelligent piston cylinder grinding equipment integrating online roughness detection according to claim 5, characterized in that, The comprehensive passivation index includes: the ratio of the decrease in the unit normal force removal efficiency relative to its reference value, the ratio of the root mean square of the high-frequency component of the force signal relative to its upper limit reference value at the end of its life, and the ratio of the integral energy in the contact frequency band relative to its normal working condition reference value, weighted and summed according to a preset weighting coefficient to obtain the comprehensive passivation index with a value between zero and one, where zero represents complete sharpness and one represents complete passivation. The weighting coefficient is determined by linear discriminant analysis of the grinding wheel's full life cycle test data.

Images