Bidirectional cooperation and adaptive control method and system driven by wall thickness follow-up measurement data

Abstract

The invention relates to the technical field of ultrasonic measurement equipment, in particular to a bidirectional cooperation and adaptive control method and system driven by wall thickness follow-up measurement data, and the method comprises the steps: evaluating the confidence coefficient weight of each eddy current sensor according to pre-sampling data; an eddy current sensor is adopted to measure a workpiece with the thickness to be measured to obtain eddy current vector data, and confidence coefficients are adopted to perform fusion to obtain an adjustment vector of the mechanical arm; and controlling the mechanical arm based on the adjustment vector, and then measuring the thickness by adopting a water immersion ultrasonic probe. In order to solve the problem that in the actual machining process of the eddy current sensors, due to the influence of a complex machining environment, a plurality of eddy current sensors are arranged on a plurality of measuring positions respectively and grouped, and in the process of calculating the eddy current vectors, the reliability of the eddy current sensors is analyzed in advance, so that the reliability of the eddy current sensors is improved. And the corresponding fusion weight is given, and the fused eddy current vector is obtained based on the fusion weight, so that the accuracy of eddy current measurement is improved.

Description

Technical Field

[0001] This invention relates to the field of ultrasonic measurement equipment technology, specifically to a bidirectional collaborative and adaptive control method and system driven by wall thickness follow-up measurement data. Background Technology

[0002] Mirror milling is a machining method designed for large, thin-walled aircraft skins. Its main difference from traditional milling lies in the application of a support and measuring device on the back of the skin. During machining, the milling cutter and the support and measuring device are aligned with the skin from both sides, and milling is performed synchronously. Compared to chemical etching milling used in traditional skin machining, it achieves better product uniformity. Typically, to achieve higher machining accuracy, mirror milling systems often integrate various sensors, such as water immersion ultrasonic probes. An ultrasonic probe is a device based on the principle of ultrasound. It emits ultrasonic waves to the workpiece and receives the reflected echoes, calculating the workpiece thickness by analyzing the time-of-flight of the ultrasonic waves. Water immersion ultrasonic probes, based on traditional industrial ultrasonic probes, use water as a coupling agent to achieve higher detection sensitivity and resolution, and are often used in precision machining processes.

[0003] In the existing technology, there are thickness measurement systems based on water immersion ultrasonic probes.

[0004] For example, patent document CN202223121051.9 discloses a water immersion C-scan ultrasonic probe adjustment device, relating to the field of detection. It includes a fixed plate with multiple adjustment mechanisms on the fixed plate. Each adjustment mechanism includes an eccentricity adjustment component for adjusting the probe's eccentricity, a distance adjustment component for adjusting the probe's distance from the water layer, and a first angle adjustment component and a second angle adjustment component for adjusting the probe's rotation angle. The probe is mounted on the second angle adjustment component. The rotation axis of the first angle adjustment component is perpendicular to the rotation axis of the second angle adjustment component. By adjusting the position of the ultrasonic probe through four-axis linkage, the position of the ultrasonic probe becomes more accurate, improving the accuracy of ultrasonic probe detection. Using multiple adjustment mechanisms to adjust different ultrasonic detection angles allows the ultrasonic probe to more accurately detect transverse waves, longitudinal waves, and the thickness of pipes or rods.

[0005] For example, patent document CN201310174143.X discloses a thickness measurement device and method for ultrasonic testing of multilayer absorbing coatings, belonging to the field of ultrasonic nondestructive testing and evaluation technology. This device comprises a portable digital ultrasonic flaw detector with a bandwidth of 0-35MHz, a delay block probe or a partially water-immersed ultrasonic delay line probe, a coating sound velocity calibration sample, and a computer integrating a thickness measurement algorithm. The device calculates the coating thickness by selecting Δt or fn based on the ultrasonic echo characteristics. Through iterative windowing analysis combining the autocorrelation method and the sound pressure reflection coefficient power spectrum method, an accurate fn is selected to achieve coating thickness measurement. This thickness measurement device and method overcome the limitations of existing ultrasonic thickness measurement technologies, such as high requirements for the frequency band of the flaw detector and probe, the need for manual intervention in data interception, and applicability only to single-layer coatings. The equipment used is small in size and lightweight, suitable for on-site thickness measurement of various substrates and multilayer coatings, offering significant economic and social benefits.

[0006] However, in actual implementation, the inventors found that when this type of technical solution is applied to a mirror milling system, due to the relatively complex geometric surface of the workpiece itself, the emission path and echo path of the ultrasonic probe during the thickness measurement process may have a significant offset relative to the geometric normal, which will lead to inaccurate ultrasonic thickness measurement results. Summary of the Invention

[0007] To address the aforementioned problems in the existing technology, a bidirectional collaborative and adaptive control method driven by wall thickness servo measurement data is provided.

[0008] On the other hand, a system is also provided for implementing the bidirectional collaborative and adaptive control method driven by the wall thickness follow-up measurement data.

[0009] The specific technical solution is as follows:

[0010] A bidirectional collaborative and adaptive control method driven by wall thickness servo measurement data, applicable to ultrasonic thickness measurement systems;

[0011] The ultrasonic thickness measurement system includes a water immersion ultrasonic probe located at the front end of the robotic arm and multiple eddy current sensors surrounding the water immersion ultrasonic probe.

[0012] The wall thickness tracking measurement method includes:

[0013] Step S1: Use the eddy current sensor to perform pre-sampling to obtain pre-sampled data;

[0014] Step S2: Evaluate the confidence weight of each of the eddy current sensors based on the pre-sampled data;

[0015] The eddy current sensor is pre-assigned multiple eddy current sensor groups, and each eddy current sensor group includes at least two eddy current sensors.

[0016] Step S3: Use the eddy current sensor to measure the thickness of the workpiece to be measured to obtain eddy current vector data;

[0017] Step S4: The adjustment vector of the robotic arm is obtained by fusing the eddy current vector data, and the confidence weight is used for fusion during the fusion process;

[0018] The adjustment vector is used to control the water immersion ultrasonic probe to maintain a constant distance and normal orientation relative to the workpiece to be measured;

[0019] Step S5: Control the robotic arm based on the adjustment vector, and then use the water immersion ultrasonic probe to measure the thickness.

[0020] On the other hand, step S1 includes:

[0021] Step S11: Clamp a standard test block on the worktable and move the robotic arm to the calibrated position;

[0022] Step S12: Use the eddy current sensor to perform pre-sampling sequentially to obtain pre-sampled data;

[0023] The pre-sampling data includes the distance corresponding to the standard test block measured by the eddy current sensor, and the ambient noise when the eddy current sensor does not emit eddy currents.

[0024] On the other hand, the method for generating the confidence weights includes:

[0025] w i =α·w p,i +β·w c,i +γ·w e,i ,α+β+γ=1;

[0026] In the formula, w i The confidence weight of the i-th eddy current sensor;

[0027] w p,i The accuracy weight of the i-th eddy current sensor;

[0028] w c,i The consistency weight is the i-th eddy current sensor.

[0029] w e,i The environmental adaptation weights for the i-th eddy current sensor;

[0030] α, β, and γ are weighting coefficients used to adjust the proportion of each part's weight in the overall confidence level, and their sum is 1.

[0031] On the other hand, step S2 includes:

[0032] Step S21: Within each eddy current sensor group, calculate the mean and standard deviation of the distance based on the pre-sampled data, and calculate the measurement distance of each eddy current sensor based on the mean and standard deviation of the distance to generate a consistency weight for each eddy current sensor;

[0033] Step S22: Filter the environmental noise collected by the eddy current sensor to obtain the residual noise intensity, and generate the environmental adaptation weight according to the correspondence between the residual noise intensity and the preset threshold range.

[0034] On the other hand, step S3 includes:

[0035] Step S31: In each of the eddy current sensor groups, the eddy current sensors are screened according to the confidence weight to remove the eddy current sensors to be calibrated;

[0036] Step S32: Use the remaining eddy current sensor to measure the workpiece to be measured to obtain eddy current vector data.

[0037] On the other hand, step S4 includes:

[0038] Step S41: In each of the eddy current sensor groups, the eddy current vector data are fused according to the confidence weight to obtain a fused vector;

[0039] Step S42: In the machine tool coordinate system, calculate the eddy current normal vector between the multiple fused vectors;

[0040] Step S43: Calculate the adjustment vector based on the eddy current normal vector and the theoretical normal vector of the workpiece to be measured.

[0041] On the other hand, step S5 includes:

[0042] Step S51: Control the robotic arm based on the adjustment vector so that the water immersion ultrasonic probe has a predetermined spacing and normal;

[0043] Step S52: Use the water immersion ultrasonic probe to measure the thickness of the workpiece, and filter the original signal to obtain a filtered signal;

[0044] Step S53: Calibrate the filtered signal to obtain the measurement result.

[0045] On the other hand, step S52 includes:

[0046] Step S521: Perform median filtering on the original signal to obtain the first intermediate signal;

[0047] Step S522: Perform adaptive mean filtering on the first intermediate signal to obtain the second intermediate signal;

[0048] Step S523: Perform Kalman filtering on the second intermediate signal to obtain the third intermediate signal.

[0049] On the other hand, step S53 includes:

[0050] Step S531: Perform temperature compensation correction on the filtered signal in combination with material properties to obtain the compensated sound velocity, and calculate the intermediate thickness measurement result based on the compensated sound velocity;

[0051] Step S532: Perform nonlinear error correction on the intermediate thickness measurement results to obtain the measurement results.

[0052] A bidirectional cooperative and adaptive control system driven by wall thickness servo measurement data is provided for implementing the above-mentioned bidirectional cooperative and adaptive control method driven by wall thickness servo measurement data.

[0053] The above technical solution has the following advantages or beneficial effects:

[0054] To address the issue that existing water immersion ultrasonic probes are prone to inaccurate measurements when measuring workpieces with complex geometric surfaces due to normal deviation and changes in coupling agent thickness, this solution adds multiple eddy current sensors arranged in a ring around the water immersion ultrasonic probe. These sensors measure the distance around the detection center point, obtaining multiple eddy current vectors. The normal vector is calculated based on the installation position of the eddy current sensors, thus adjusting the normal deviation of the current water immersion ultrasonic probe relative to the detection plane and ensuring the normal orientation of the probe is maintained.

[0055] Furthermore, to address the issue that eddy current sensors may fail due to complex processing environments during actual manufacturing, this solution sets up multiple eddy current sensors at various measurement locations and groups them together. During the calculation of the eddy current vector, the reliability of the eddy current sensors is analyzed in advance, and corresponding fusion weights are assigned. The fused eddy current vector is obtained based on the fusion weights, thereby improving the accuracy of eddy current measurement. Attached Figure Description

[0056] Embodiments of the invention will be described more fully with reference to the accompanying drawings. However, the drawings are for illustration and explanation only and do not constitute a limitation on the scope of the invention.

[0057] Figure 1This is an overall schematic diagram of an embodiment of the present invention;

[0058] Figure 2 This is a schematic diagram of step S1 in an embodiment of the present invention;

[0059] Figure 3 This is a schematic diagram of step S2 in an embodiment of the present invention;

[0060] Figure 4 This is a schematic diagram of step S3 in an embodiment of the present invention;

[0061] Figure 5 This is a schematic diagram of step S4 in an embodiment of the present invention;

[0062] Figure 6 This is a schematic diagram of step S5 in an embodiment of the present invention;

[0063] Figure 7 This is a schematic diagram of step S52 in an embodiment of the present invention;

[0064] Figure 8 This is a schematic diagram of step S53 in an embodiment of the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0067] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0068] This invention includes:

[0069] A bidirectional collaborative and adaptive control method driven by wall thickness servo measurement data, applicable to ultrasonic thickness measurement systems;

[0070] The ultrasonic thickness measurement system includes a water immersion ultrasonic probe located at the front end of the robotic arm and multiple eddy current sensors surrounding the water immersion ultrasonic probe.

[0071] like Figure 1 As shown, the wall thickness tracking measurement method includes:

[0072] Step S1: Use an eddy current sensor to perform pre-sampling to obtain pre-sampled data;

[0073] Step S2: Evaluate the confidence weight of each eddy current sensor based on the pre-sampled data;

[0074] The eddy current sensor is pre-assigned with multiple eddy current sensor groups, and each eddy current sensor group includes at least two eddy current sensors.

[0075] Step S3: Use an eddy current sensor to measure the thickness of the workpiece to be measured to obtain eddy current vector data;

[0076] Step S4: The adjustment vector of the robotic arm is obtained by fusing the eddy current vector data. Confidence weights are used in the fusion process.

[0077] The adjustment vector is used to control the water immersion ultrasonic probe to maintain a constant distance and normal orientation relative to the thick workpiece being measured;

[0078] Step S5: Control the robotic arm based on the adjustment vector, and then use a water immersion ultrasonic probe to measure the thickness.

[0079] Specifically, addressing the issue that existing water immersion ultrasonic probes are prone to inaccurate measurements when measuring workpieces with complex geometric surfaces due to normal deviation and changes in coupling agent thickness, this solution adds multiple eddy current sensors arranged in a ring around the water immersion ultrasonic probe. By measuring the distance around the detection center point using these eddy current sensors, multiple eddy current vectors are obtained. Combined with the installation position of the eddy current sensors, the normal vector is calculated, thereby adjusting the normal deviation of the current water immersion ultrasonic probe relative to the detection plane, ensuring that the normal of the water immersion ultrasonic probe is maintained.

[0080] Furthermore, to address the issue that eddy current sensors may fail due to complex processing environments during actual manufacturing, this solution sets up multiple eddy current sensors at various measurement locations and groups them together. During the calculation of the eddy current vector, the reliability of the eddy current sensors is analyzed in advance, and corresponding fusion weights are assigned. The fused eddy current vector is obtained based on the fusion weights, thereby improving the accuracy of eddy current measurement.

[0081] Specifically, the above technical solution is mainly implemented as a software embodiment in the ultrasonic thickness measurement system. The ultrasonic thickness measurement system mainly includes a robotic arm, with a water immersion ultrasonic probe and an eddy current sensor arranged around the probe mounted at its front end. The eddy current sensor is used to generate eddy currents during the ultrasonic probe's measurement process to measure the eddy current distance between the ultrasonic probe and the thin-walled component.

[0082] Multiple eddy current sensors are set in the circumferential direction of the ultrasonic probe, which can generate corresponding eddy currents in the direction the ultrasonic probe is pointing and collect echo signals. Based on the intensity of the echo signal and combined with a pre-calibrated reflection intensity-distance comparison function, the eddy current spacing between the ultrasonic probe and the thin-walled part can be easily calculated and used as the machine tool control parameter of the mirror milling system to realize closed-loop control of the milling cutter cutting process.

[0083] In a typical arrangement, there are four eddy current sensors, evenly distributed at the 12, 3, 6, and 9 o'clock positions on the ultrasonic probe.

[0084] Furthermore, to achieve the aforementioned spacing control process, the relationship between eddy current intensity and spacing needs to be calibrated before processing. The calibration process includes:

[0085] A metal calibration block is pre-introduced. The normal of the eddy current sensor is then aligned with the surface of the metal calibration block, maintaining a certain gap, which serves as the zero-point gap. Eddy current values ​​are then read and associated with the zero-point position for storage. Based on this, the eddy current sensor is moved along the normal in specific steps to increase the gap, and eddy current values ​​are recorded and stored. By performing linear fitting on the collected distances and eddy current values, the reflection intensity-distance correlation function can be easily obtained.

[0086] During the measurement process of the ultrasonic probe, the ultrasonic probe is controlled to point towards the back normal of the thin-walled part by the eddy current normal holding process.

[0087] The eddy current normal-maintaining process includes:

[0088] Obtain the distance between multiple eddy current sensors and the back surface of the thin-walled component;

[0089] A machine tool coordinate system is constructed based on the eddy current distribution generated by the eddy current probe, and the back face distance is mapped in the machine tool coordinate system respectively;

[0090] The distribution center of the eddy currents coincides with the origin of the machine tool coordinate system;

[0091] The eddy current normal vector is calculated in the machine tool coordinate system based on the back face distance obtained by mapping.

[0092] The orientation of the ultrasonic probe is adjusted based on the eddy current normal vector so that the eddy current normal vector coincides with the back normal vector of the thin-walled component.

[0093] Specifically, to achieve better measurement results, in this embodiment, the ultrasonic probe is kept aligned with the back normal of the thin-walled component by utilizing the feedback signal from the eddy current sensor. In this embodiment, the ultrasonic probe and eddy current sensor are mounted on a ball joint, allowing their orientation to be freely changed.

[0094] After obtaining the real-time thickness using eddy currents, the distances of multiple eddy current sensors relative to the back surface of the thin-walled part can be obtained according to this process. Then, three of the eddy current sensors are selected, and a machine tool coordinate system is established based on the distribution surface of the eddy currents they generate. This machine tool coordinate system has the distribution centers of the three eddy currents as its origin, and the measurement distances of the three eddy current sensors are mapped to the machine tool coordinate system.

[0095] Then, calculate the eddy current normal vector between the three eddy currents, which is: n = n1 × n2.

[0096] In the formula, n1 represents the vector of the first eddy current and the second eddy current, with the direction from the second eddy current to the first eddy current; n2 represents the vector of the first eddy current and the third eddy current, with the direction from the third eddy current to the first eddy current; n represents the eddy current normal vector of the first eddy current, with the direction being the cross product of vectors n1 and n2. The calculation process follows the right-hand rule.

[0097] Based on the above process, the eddy current normal vector can be obtained. The back side of the thin-walled part has a predetermined theoretical normal vector (0, 0, 1) in the machine tool coordinate system. For these two vectors, the vector angle between them can be easily calculated, which corresponds to the deviation angle of the ultrasonic probe. Then, the spherical joint is rotated based on the deviation angle to make the two vectors coincide, thus completing the normal control process of the ultrasonic probe.

[0098] Based on the aforementioned normal control process, and considering the potential for cumulative error and electromagnetic noise in the environment during actual processing, this solution employs multiple eddy current sensors arranged in groups at predetermined locations. For example, using the arrangement of four eddy current sensors as an example, the individual sensors are replaced with four groups, each containing six hexagonal sensors arranged in a hexagonal pattern. The center of this hexagon corresponds to the original sensor mounting position. In the normal vector calculation, the reliability of each eddy current sensor in a single group is first calculated using the methods described above. Then, multiple measurement results from the eddy current sensor groups are fused to form a fused vector. This fused vector, mapped to the center of the hexagon, is equivalent to the eddy current vector acquired by a single sensor. The normal vector holding and thickness measurement processes described above are then performed, achieving good accuracy.

[0099] Its specific implementation methods include:

[0100] In one embodiment, such as Figure 2 As shown, step S1 includes:

[0101] Step S11: Clamp the standard test block on the worktable and move the robotic arm to the calibrated position;

[0102] Step S12: Use an eddy current sensor to perform pre-sampling sequentially to obtain pre-sampled data;

[0103] The pre-sampling data includes the distance corresponding to the standard test block measured by the eddy current sensor, and the ambient noise when the eddy current sensor is not emitting eddy currents.

[0104] Specifically, in order to effectively calculate the confidence weight of the eddy current sensor, in this embodiment, a standard test block is first clamped on the workbench, and then the robotic arm is controlled to move to the calibration position.

[0105] The standard test block is typically a rectangular block made of the same material as the workpiece used in actual machining, and has a certain thickness. The calibration position, in actual implementation, should be a position perpendicular to the standard test block, and the distance between the water immersion ultrasonic probe and the standard test block should be controlled to maintain a fixed distance.

[0106] Subsequently, pre-sampling is performed sequentially using eddy current sensors to obtain pre-sampled data. This sampling process is carried out in two steps: sequentially collecting ambient noise for each eddy current sensor, and measuring the distance corresponding to the standard test block using a single eddy current sensor. This distance should be repeated multiple times to form a time series.

[0107] In one embodiment, the method for generating confidence weights includes:

[0108] w i =α·w p，i +β·w c，i +γ·w e，i α+β+γ=1;

[0109] In the formula, w i For the confidence weight of the i-th eddy current sensor;

[0110] w p,i Let be the accuracy weight of the i-th eddy current sensor;

[0111] w c，i Let be the consistency weight of the i-th eddy current sensor;

[0112] w e,i Let be the environmental adaptation weight for the i-th eddy current sensor;

[0113] α, β, and γ are weighting coefficients used to adjust the proportion of each part's weight in the overall confidence level, and their sum is 1.

[0114] Specifically, in order to evaluate the reliability of the measurement results output by the eddy current sensor, this scheme introduces three weights to form the confidence weight.

[0115] The accuracy weight is the accuracy of a single eddy current sensor under ideal conditions, which is obtained through experimental calibration in advance, and is assigned a corresponding basic weight.

[0116] The consistency weight is calculated based on the difference between the output of other sensors and the output of the i-th sensor in a single eddy current sensor group. The smaller the difference, the higher the consistency.

[0117] The environmental adaptation weight is based on the environmental noise measured by the eddy current sensor during the sampling process. The residual noise intensity is obtained after filtering the signal, and the weight is assigned according to the correspondence between the residual noise intensity and the preset threshold range.

[0118] By calculating the weights of the three parts mentioned above, the confidence weights can be obtained for vector fusion in a better way.

[0119] In one embodiment, such as Figure 3 As shown, step S2 includes:

[0120] Step S21: Within each eddy current sensor group, calculate the mean and standard deviation of the distance based on the pre-sampled data, and calculate the consistency weight of each eddy current sensor based on the mean and standard deviation of the distance for the measurement distance of each eddy current sensor.

[0121] Step S22: Filter the environmental noise collected by the eddy current sensor to obtain the residual noise intensity, and generate environmental adaptation weights based on the correspondence between the residual noise intensity and the preset threshold range.

[0122] Specifically, in order to achieve a better calculation effect on the confidence weight of the eddy current sensor, in this embodiment, firstly, in each eddy current sensor group, the mean distance and standard deviation of the distance within the group are calculated based on the distance sequence measured by a single eddy current sensor in the pre-sampled data. This calculation process obtains the corresponding results by forming a matrix of multiple sampling points in multiple distance sequences.

[0123] Subsequently, for multiple sampling points in a single distance sequence, the mean and standard deviation of the distance are calculated to determine the output difference of each sampling point relative to other eddy current sensors and assign a consistency weight.

[0124] Then, the environmental noise collected by each eddy current sensor is filtered, and the noise intensity is obtained by sampling based on the filtering result. The environmental adaptation weights of the eddy current sensors are generated by looking up the table based on the noise intensity.

[0125] In one embodiment, such as Figure 4 As shown, step S3 includes:

[0126] Step S31: In each eddy current sensor group, the eddy current sensors are screened according to the confidence weight to remove the eddy current sensors to be calibrated.

[0127] Step S32: Use the remaining eddy current sensor to measure the workpiece to be measured and obtain eddy current vector data.

[0128] Specifically, in order to achieve a better sampling process, in this embodiment, the eddy current sensors are first screened according to the confidence weight. Eddy current sensors with a confidence weight of less than 0.6 are usually marked as eddy current sensors to be calibrated. The output data of these eddy current sensors are isolated and the corresponding calibration process is triggered when appropriate.

[0129] Subsequently, the remaining eddy current sensor was used to measure the thickness of the workpiece to be measured to obtain eddy current vector data.

[0130] In one embodiment, such as Figure 5 As shown, step S4 includes:

[0131] Step S41: In each eddy current sensor group, the eddy current vector data are fused according to the confidence weight to obtain the fused vector;

[0132] Step S42: In the machine tool coordinate system, calculate the eddy current normal vector between multiple fused vectors;

[0133] Step S43: Calculate the adjustment vector based on the eddy current normal vector and the theoretical normal vector of the workpiece to be measured.

[0134] Specifically, in order to achieve better computational results, in this embodiment, for each eddy current sensor group, the eddy current vector data is first fused according to the confidence weight within the group to obtain a fused vector. This process can be obtained by weighted calculation combined with pose mapping, thereby forming the fused vector calculated in the group.

[0135] A typical calculation process involves six eddy current sensors arranged in a hexagonal pattern within an eddy current sensor group. The raw eddy current vector data collected by the eddy current sensors is first mapped to a mapping vector at the center of the hexagon. Then, multiple mapping vectors are weighted together to obtain a fused vector, which serves as the fused vector for the entire eddy current sensor group.

[0136] Based on the obtained fused vector, the eddy current normal vector between multiple fused vectors is calculated in the machine tool coordinate system. Taking a triangular arrangement of multiple eddy current sensor groups as an example, in this embodiment, the eddy current normal vector can be obtained by cross multiplying the eddy current normal vectors.

[0137] At this point, since the surface of the workpiece to be measured has a theoretical plane in the machine tool coordinate system, it is easy to determine the normal of the theoretical plane by combining the theoretical plane. Subtracting the normal of the theoretical plane from the eddy current normal vector makes it easy to obtain the vector angle that the adjustment vector should have.

[0138] Furthermore, by calculating the distance between the theoretical plane and the plane containing the eddy current normal vector, the plane spacing can be obtained. Subtracting this distance from the measured spacing of the immersion ultrasonic probe can determine the distance that needs to be adjusted. The distance that needs to be adjusted is then used as the vector length of the adjustment vector, thus obtaining the complete adjustment vector.

[0139] In one embodiment, such as Figure 6 As shown, step S5 includes:

[0140] Step S51: Control the robotic arm based on the adjustment vector so that the immersion ultrasonic probe has a predetermined spacing and normal;

[0141] Step S52: Use a water immersion ultrasonic probe to measure the thickness of the workpiece, and filter the original signal to obtain a filtered signal;

[0142] Step S53: Calibrate the filtered signal to obtain the measurement results.

[0143] Specifically, to further improve the accuracy of thickness measurement using the immersion ultrasonic probe, in this embodiment, the robotic arm is first controlled based on the adjustment vector before measurement to ensure that the immersion ultrasonic probe has a predetermined spacing and normal. Then, the immersion ultrasonic probe is used to measure the thickness of the workpiece, and the original signal is filtered to obtain a filtered signal, removing noise caused by factors such as water flow, back pressure, and motion. The filtered signal is then calibrated based on the characteristics of the actual workpiece being processed to obtain accurate measurement results.

[0144] In one embodiment, such as Figure 7 As shown, step S52 includes:

[0145] Step S521: Perform median filtering on the original signal to obtain the first intermediate signal;

[0146] Step S522: Perform adaptive mean filtering on the first intermediate signal to obtain the second intermediate signal;

[0147] Step S523: Perform Kalman filtering on the second intermediate signal to obtain the third intermediate signal.

[0148] Specifically, in order to achieve better filtering results, this embodiment employs a three-stage filtering strategy for the original signal.

[0149] The first step is median filtering, which eliminates pulse interference signals caused by vibrations during machining when measuring with a water immersion ultrasonic probe.

[0150] The filtering method is implemented as follows:

[0151] The original signal is acquired using a sliding window of length N. Within this sliding window, the output value is obtained by processing the signal using the following formula:

[0152] y(k)=median{x(kN / 2),...,x(k),...,x(k+N / 2)};

[0153] In the formula, y(k) is the window output value, x(k) is the sampled value of the kth signal sampling point, and k∈[(kN / 2), (k+N / 2)].

[0154] Meanwhile, during the filtering process, the signal change rate is calculated for the previous sliding window. The signal change rate in the previous window is compared with the signal change rate threshold to adaptively adjust the window length.

[0155] Generally, there are two thresholds for the signal rate of change. When the signal rate of change is greater than the first threshold, N is set to 3 to improve the fast response to high-frequency signals.

[0156] When the rate of change of the signal is less than the second rate of change threshold, N is set to 7 to enhance the filtering effect.

[0157] In other cases, N can be set to 5 to achieve a moderate filtering effect.

[0158] The second step is adaptive mean filtering, which can effectively suppress random noise caused by motion.

[0159] In this step, the signal-to-noise ratio of the first intermediate signal is first calculated. The noise power is obtained by fast Fourier transform and then low-pass filtering. The noise power is then obtained by power spectral density calculation of the noise signal.

[0160] Then, the weighted average number of times M is calculated, M = max(3, min(20, round(15 / SNR))), where SNR is the calculated signal-to-noise ratio.

[0161] Finally, adaptive mean filtering is performed based on the weighted average number of times.

[0162] The third step is Kalman filtering, which can correct the systematic errors of the water immersion ultrasonic system itself.

[0163] The above three-level correction strategy achieves a better filtering effect.

[0164] In one embodiment, such as Figure 8 As shown, step S53 includes:

[0165] Step S531: Perform temperature compensation correction on the filtered signal in combination with material properties to obtain the compensated sound velocity, and calculate the intermediate thickness measurement result based on the compensated sound velocity;

[0166] Step S532: Perform nonlinear error correction on the intermediate thickness measurement results to obtain the measurement results.

[0167] Specifically, to achieve more accurate measurement results, in this embodiment, the processing data of the workpiece is first searched, the material type is determined based on the processing data, and then a multi-material sound velocity database is searched, for example:

[0168] Material: 2024 aluminum alloy; v0 = 6260 m / s; α = -1.2 m / (s·℃)

[0169] Material: 7075 aluminum alloy: v o =6420m / s, α=-1.1m / (s·℃)

[0170] Material 2195 aluminum-lithium alloy: v0=6180m / s, α=-1.4m / (s·℃)

[0171] Then, the filtered signal is temperature-compensated and corrected using the retrieved data to obtain the first corrected signal, eliminating sound velocity drift caused by temperature changes during processing, including:

[0172] v(T)=v0×[1-α×(T-T0)]

[0173] In the formula, v(T) is the corrected sound velocity, v0 is the sound velocity at the standard room temperature, α is the material temperature coefficient, T0 is the reference temperature, and T is the actual processing temperature.

[0174] The intermediate thickness measurement results are obtained by calculating based on the corrected sound velocity.

[0175] Finally, nonlinear correction is performed on the intermediate thickness measurement results to eliminate systematic errors, including:

[0176] When the intermediate thickness measurement result is ≤2mm, the measurement result is obtained by multiplying by the correction coefficient k1 = 1.002;

[0177] When the intermediate thickness measurement result is between 2mm and 10mm, the measurement result is obtained by multiplying by the correction factor k2 = 1.000;

[0178] When the intermediate thickness measurement result is greater than 10mm, the measurement result is obtained by multiplying by the correction coefficient k3 = 0.998.

[0179] A bidirectional cooperative and adaptive control system driven by wall thickness servo measurement data is provided for implementing the aforementioned bidirectional cooperative and adaptive control system driven by wall thickness servo measurement data.

[0180] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. A bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data, characterized in that, Suitable for ultrasonic thickness measurement systems; The ultrasonic thickness measurement system includes a water immersion ultrasonic probe located at the front end of the robotic arm and multiple eddy current sensors surrounding the water immersion ultrasonic probe. The wall thickness tracking measurement method includes: Step S1: Use the eddy current sensor to perform pre-sampling to obtain pre-sampled data; Step S2: Evaluate the confidence weight of each of the eddy current sensors based on the pre-sampled data; The eddy current sensor is pre-assigned multiple eddy current sensor groups, and each eddy current sensor group includes at least two eddy current sensors. Step S3: Use the eddy current sensor to measure the thickness of the workpiece to be measured to obtain eddy current vector data; Step S4: The adjustment vector of the robotic arm is obtained by fusing the eddy current vector data, and the confidence weight is used for fusion during the fusion process; The adjustment vector is used to control the water immersion ultrasonic probe to maintain a constant distance and normal orientation relative to the workpiece to be measured; Step S5: Control the robotic arm based on the adjustment vector, and then use the water immersion ultrasonic probe to measure the thickness.

2. The bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data according to claim 1, characterized in that, Step S1 includes: Step S11: Clamp a standard test block on the worktable and move the robotic arm to the calibrated position; Step S12: Use the eddy current sensor to perform pre-sampling sequentially to obtain pre-sampled data; The pre-sampling data includes the distance corresponding to the standard test block measured by the eddy current sensor, and the ambient noise when the eddy current sensor does not emit eddy currents.

3. The bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data according to claim 2, characterized in that, The method for generating the confidence weights includes: w i =α·w p，i +β·w c，i +γ·w e，i ,α+β+γ=1; In the formula, w i The confidence weight of the i-th eddy current sensor; w p，i The accuracy weight of the i-th eddy current sensor; w c，i The consistency weight is the i-th eddy current sensor. w e,i The environmental adaptation weights for the i-th eddy current sensor; α, β, and γ are weighting coefficients used to adjust the proportion of each part's weight in the overall confidence level, and their sum is 1.

4. The bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data according to claim 3, characterized in that, Step S2 includes: Step S21: Within each eddy current sensor group, calculate the mean and standard deviation of the distance based on the pre-sampled data, and calculate the measurement distance of each eddy current sensor based on the mean and standard deviation of the distance to generate a consistency weight for each eddy current sensor; Step S22: Filter the environmental noise collected by the eddy current sensor to obtain the residual noise intensity, and generate the environmental adaptation weight according to the correspondence between the residual noise intensity and the preset threshold range.

5. The bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data according to claim 1, characterized in that, Step S3 includes: Step S31: In each of the eddy current sensor groups, the eddy current sensors are screened according to the confidence weight to remove the eddy current sensors to be calibrated; Step S32: Use the remaining eddy current sensor to measure the workpiece to be measured to obtain eddy current vector data.

6. The bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data according to claim 1, characterized in that, Step S4 includes: Step S41: In each of the eddy current sensor groups, the eddy current vector data are fused according to the confidence weight to obtain a fused vector; Step S42: In the machine tool coordinate system, calculate the eddy current normal vector between the multiple fused vectors; Step S43: Calculate the adjustment vector based on the eddy current normal vector and the theoretical normal vector of the workpiece to be measured.

7. The bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data according to claim 1, characterized in that, Step S5 includes: Step S51: Control the robotic arm based on the adjustment vector so that the water immersion ultrasonic probe has a predetermined spacing and normal; Step S52: Use the water immersion ultrasonic probe to measure the thickness of the workpiece, and filter the original signal to obtain a filtered signal; Step S53: Calibrate the filtered signal to obtain the measurement result.

8. The bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data according to claim 7, characterized in that, Step S52 includes: Step S521: Perform median filtering on the original signal to obtain the first intermediate signal; Step S522: Perform adaptive mean filtering on the first intermediate signal to obtain the second intermediate signal; Step S523: Perform Kalman filtering on the second intermediate signal to obtain the third intermediate signal.

9. The bidirectional collaborative and adaptive control method driven by wall thickness follow-up measurement data according to claim 7, characterized in that, Step S53 includes: Step S531: Perform temperature compensation correction on the filtered signal in combination with material properties to obtain the compensated sound velocity, and calculate the intermediate thickness measurement result based on the compensated sound velocity; Step S532: Perform nonlinear error correction on the intermediate thickness measurement results to obtain the measurement results.

10. A bidirectional cooperative and adaptive control system driven by wall thickness follow-up measurement data, characterized in that, This is used to implement the bidirectional collaborative and adaptive control method driven by wall thickness servo measurement data as described in any one of claims 1-9.

Images