System and method for measuring size of continuous rolling steel material in real time based on active phased array

Abstract

The invention discloses a continuous rolling steel material size real-time measurement system and method based on an active phased array radar. The system utilizes the characteristics of strong penetrability of the millimeter wave radar and agile AESA wave beam, overcomes the influence of severe environment, and realizes the non-contact tracking measurement of the high-speed moving steel material. Through an innovative grading signal processing framework, precision requirements of different rolling stages are intelligently matched, finally, high-reliability and full-process real-time size feedback is provided for an automatic control system of the rolling mill, the size precision and production stability of continuous rolling products are remarkably improved, and basic technical support is provided for intelligentization and unmanned operation of a steel continuous rolling production line.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of metallurgical automation online detection and radar signal processing, and in particular relates to a real-time measurement system and method for the dimensions of continuously rolled steel based on an active phased array. Background Technology

[0002] During the rolling process, the rolls in each pass of a continuous rolling production line experience continuous wear, leading to unstable steel dimensions. This, in turn, causes fluctuations in rolling speed, resulting in poor product flowability, increased weight deviation, and even production and quality accidents. Only when the steel dimensions in each pass are stable can speed stability (stable stacking relationship) be maintained, accidents avoided, continuous production achieved, and high-precision products with stable flowability and small weight deviation fluctuations be produced. This, in turn, reduces production costs and enhances the company's competitiveness. Therefore, real-time and accurate measurement of steel is crucial.

[0003] The main technical problems currently existing include:

[0004] The steel material in the continuous rolling production line moves at a high speed (0-40 m / s, and the finished product speed can reach up to 40 m / s). Traditional contact measurement (such as width gauges and laser scanners) is easily affected by high temperature, water mist and iron oxide scale, resulting in low accuracy and easy wear. Manual measurement also poses significant safety hazards.

[0005] Non-contact optical measurements (such as CCD vision and laser triangulation) have poor stability in smoke and steam environments and insufficient sampling rate for high-speed moving targets.

[0006] Existing radar ranging technologies are mostly single-point or static scanning, which cannot meet the requirements of multi-channel, high dynamic, and bidirectional synchronous size measurement.

[0007] Different stages of roughing (±1mm), intermediate rolling (±0.5mm), and finishing rolling (±0.1mm) require differentiated precision control, which is difficult for existing systems to adapt to. Summary of the Invention

[0008] In view of this, the purpose of the present invention is to provide a real-time measurement system and method for the dimensions of continuously rolled steel based on an active phased array, so as to realize high-speed, high-precision, and anti-interference real-time monitoring of the dimensions of steel throughout the entire continuous rolling production line, and meet the progressive accuracy requirements from roughing to finishing rolling.

[0009] In a first aspect, embodiments of the present invention provide a real-time measurement system for the dimensions of continuously rolled steel based on an active phased array, comprising:

[0010] The dual orthogonal active phased array radar unit consists of two AESA radar modules deployed above and to the side of the rolling line; the upper radar module scans vertically downwards to measure the width W of the steel, while the side radar module scans horizontally to measure the height H of the steel.

[0011] The dynamic beam control system is connected to the dual orthogonal active phased array radar unit. By adjusting the phase difference of each radiating unit in real time, it realizes electronic scanning and dynamic focusing of the beam within a range of ±30°.

[0012] The FMCW-Doppler composite signal processing module uses a linear frequency modulated continuous wave waveform to compensate for the influence of the steel material's movement speed through Doppler frequency shift.

[0013] The graded signal processing unit has three parallel processing channels built-in for roughing, intermediate rolling, and finishing rolling. It automatically switches the algorithm mode according to the pass of the steel material, that is, it calls the beamforming algorithm, MIMO virtual array algorithm and super-resolution algorithm for processing respectively.

[0014] Preferably, the AESA radar module operates in the frequency band of 77GHz-81GHz, with a wavelength range of 3.7-3.9mm, a beam scanning range of ≥±30° for a single module, and a scanning switching time of ≤100μs.

[0015] Preferably, the super-resolution algorithm called by the graded signal processing unit in the finishing rolling stage is the MUSIC algorithm or the ESPRIT algorithm.

[0016] Preferably, it also includes a data interface with the rolling mill control system, used to feed back the real-time calculated steel height, width dimensions and changing trends to the rolling mill, forming a closed-loop control system.

[0017] Preferably, the dynamic beam control system employs a predictive tracking algorithm to predict the beam pointing angle at the next moment based on the steel material position and velocity data from the previous N scanning cycles.

[0018] Preferably, the system further includes a real-time temperature monitoring module, which uses an infrared thermometer to collect the surface temperature of the steel in real time, providing input parameters for rolling processes such as controlled cooling and controlled rolling.

[0019] Secondly, a method for real-time measurement of steel dimensions includes the following steps:

[0020] S1. Dual radar modules transmit FMCW millimeter-wave signals to the moving steel material and receive the echoes.

[0021] S2. Mix, filter, and convert the echo signal to an AD converter to obtain the intermediate frequency signal;

[0022] S3. Perform two-dimensional FFT processing on the intermediate frequency signal to obtain preliminary range-velocity information of the target;

[0023] S4. The dynamic beam control system controls the radar beam to perform tracking and scanning based on the predicted value of the steel material speed.

[0024] S5. Estimate the angle of arrival of the multi-frame echo data acquired by the tracking scan and generate a three-dimensional point cloud on the surface of the steel material.

[0025] S6. Based on the steel material's pass, select the corresponding graded processing algorithm to optimize the point cloud, fitting the upper surface and side edges respectively.

[0026] S7 outputs high-precision height and width values.

[0027] Preferably, the specific steps of the hierarchical processing algorithm include:

[0028] In the roughing section, a beamforming algorithm is used to control the accuracy to ±1mm.

[0029] MIMO virtual array technology is used in the intermediate rolling section to control the accuracy to ±0.5mm;

[0030] In the finishing rolling section, a super-resolution spectral analysis algorithm is used to control the accuracy to ±0.1mm.

[0031] The embodiments of the present invention bring the following beneficial effects:

[0032] 1. Strong non-contact anti-interference capability: Millimeter waves can effectively penetrate water vapor / dust, adapting to harsh rolling environments;

[0033] 2. High speed and good compatibility: Microsecond-level beam control can match high-speed moving targets;

[0034] 3. Accuracy Adaptive: Hierarchical algorithm resources are allocated on demand to optimize computational efficiency;

[0035] 4. Bidirectional synchronous measurement: The orthogonal layout eliminates the height-width coupling error of traditional systems;

[0036] 5. High safety: Non-contact measurement avoids the risk of personal injury;

[0037] 6. Good real-time performance: The electronic scanning rate is >10kHz, which meets the real-time sampling requirements of high-speed steel materials.

[0038] Other features and advantages of the invention will be set forth in the following description, and some features will be obvious from the description or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description, claims, and drawings.

[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0040] Figure 1 : Schematic diagram of the system deployment on the continuous rolling production line.

[0041] Figure 2 Schematic diagram of a dual orthogonal AESA radar unit.

[0042] Figure 3 FMCW-Doppler composite signal waveform and processing principle diagram.

[0043] Figure 4 Schematic diagram of dynamic beam tracking of steel material movement.

[0044] Figure 5 Comparison of switching logic and effects of hierarchical signal processing algorithms.

[0045] Figure 6 System hardware and software architecture diagram.

[0046] Figure 7 : A schematic diagram of the closed-loop control formed by the rolling mill control system.

[0047] Figure 8 Schematic diagram and algorithm flowchart of point cloud fitting and size calculation for steel surface;

[0048] Figure 9 Flowchart of point cloud fitting algorithm. Detailed Implementation

[0049] 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.

[0050] Example 1: A real-time measurement system for the dimensions of continuously rolled steel based on an active phased array, such as... Figure 1 and Figure 6 As shown, it includes:

[0051] 1. A dual orthogonal active phased array radar unit, consisting of two AESA radar modules deployed above and to the side of the rolling line, is used to form an orthogonal measurement coordinate system;

[0052] The two AESA arrays are arranged perpendicularly to each other, and each measurement point includes a height H and a width W measurement array. The AESA radar module operates in the frequency band of 77GHz-81GHz, with a wavelength range of 3.7-3.9mm. The beam scanning range of a single module is ≥±30°, and the scan switching time is ≤100μs.

[0053] Specifically, it includes:

[0054] Top radar module: installation height 5-10m, beam elevation angle -30° to -60°;

[0055] Side radar module: horizontal range 5-10m, beam azimuth ±30°.

[0056] Furthermore, this unit features real-time closed-loop control, meaning that the measured data is fed back to control the mill parameters in real time, with a system delay of ≤10ms to meet the closed-loop control requirements.

[0057] 2. A dynamic beam control system, connected to a dual orthogonal active phased array radar unit, is used to adjust the phase of each radiating element in real time to achieve electronic beam scanning. The electronic beam scanning rate of the dynamic beam control system is ≤100μs / beam, and the scanning rate is >10kHz. The dynamic beam control system predicts the trajectory of the steel material using a Kalman filter algorithm, with a beam switching time ≤100μs, ensuring a beam hysteresis distance <4mm under a steel material velocity of 40m / s.

[0058] 3. FMCW-Doppler composite signal processing module, such as Figure 3 As shown, this is used to combine speed compensation and distance resolution to eliminate measurement errors caused by the high-speed movement of steel.

[0059] 4. The graded signal processing unit is configured to call beamforming algorithms, MIMO virtual array algorithms, and super-resolution algorithms respectively for processing based on the roughing, intermediate, and finishing stages of the steel material (e.g., ...). Figure 5 (As shown); the processing mode is automatically switched according to the rolling pass, and the corresponding precision of the steel material size data is output.

[0060] Among them, the roughing section processing module adopts a beamforming algorithm to achieve a measurement accuracy of ±1mm;

[0061] The intermediate rolling section processing module adopts MIMO virtual array technology to achieve a measurement accuracy of ±0.5mm;

[0062] The finishing rolling section processing module uses a super-resolution spectrum analysis algorithm to achieve a measurement accuracy of ±0.1mm.

[0063] Furthermore, the super-resolution algorithm used in the finishing rolling stage is either the MUSIC algorithm or the ESPRIT algorithm.

[0064] In this embodiment of the invention, a thermometer can also be installed near the finishing mill to correct for thermal expansion and contraction of the finished product after it returns to room temperature (generally based on a 20°C reference temperature) after it is stored in the warehouse.

[0065] In this embodiment of the invention, a data interface with the rolling mill control system (such as...) is also included. Figure 7As shown in the figure, it is used to feed back the real-time calculated height, width and trend of the steel material to the rolling mill, forming a closed-loop control system.

[0066] In this embodiment of the invention, the measurement system supports multi-target identification and tracking. For example, in the multi-wire cutting process of bar stock, there may be 2-5 steel bars in the finishing rolling stage. The core capability of active phased array radar is precisely "scanning and tracking" multiple targets simultaneously. Its key lies in "simultaneous multi-beam" technology, that is, the radar beam can "jump" in different directions within microseconds to illuminate and track multiple different targets respectively.

[0067] Furthermore, the performance test data for Example 1 includes:

[0068] Test conditions: steel velocity 25 m / s, ambient temperature 150-300℃

[0069] Measurement results:

[0070] Standard deviation of height measurement: 0.08mm

[0071] Standard deviation of degree measurement: 0.12mm

[0072] Data update delay: <5ms

[0073] Continuous working time: >1000 hours

[0074] Example 2: A method for real-time measurement of steel dimensions, comprising the following steps:

[0075] S1. Dual radar modules transmit FMCW millimeter-wave signals to the moving steel material and receive the echoes.

[0076] S2. Mix, filter, and convert the echo signal to an AD converter to obtain the intermediate frequency signal;

[0077] S3. Perform two-dimensional FFT processing on the intermediate frequency signal to obtain preliminary range-velocity information of the target;

[0078] S4. The dynamic beam control system controls the radar beam to perform tracking and scanning based on the predicted steel material velocity (e.g., ...). Figure 4 (as shown)

[0079] S5. Estimate the angle of arrival of the multi-frame echo data acquired by the tracking scan and generate a three-dimensional point cloud on the surface of the steel material.

[0080] S6. Based on the steel material's pass, select the corresponding graded processing algorithm to optimize the point cloud, fitting the upper surface and side edges respectively.

[0081] S7. Outputs high-precision height and width values. Further, the specific steps for calculating steel dimensions include:

[0082] The width value is output by fitting the upper surface of the steel material with the top-view array point cloud data;

[0083] The height value is output by fitting the side edge of the steel material using side-view array point cloud data;

[0084] Decoupled calculations for height and width measurements are achieved through orthogonal layout.

[0085] Furthermore, the specific steps of the hierarchical processing algorithm include:

[0086] In the roughing section, a beamforming algorithm is used to control the accuracy to ±1mm.

[0087] MIMO virtual array technology is used in the intermediate rolling section to control the accuracy to ±0.5mm;

[0088] In the finishing rolling section, a super-resolution spectral analysis algorithm is used to control the accuracy to ±0.1mm.

[0089] In this embodiment of the invention, the prediction algorithm uses motion model fitting, wherein the multi-algorithm fusion technique is described as follows:

[0090] The least squares method is used in the roughing stage, which has a fast calculation speed and meets the accuracy requirement of ±0.3mm;

[0091] The RANSAC algorithm is used in the intermediate rolling stage, which has strong anti-abnormal point interference capability and meets the accuracy requirement of ±0.2mm;

[0092] The finishing rolling stage uses a combination of RANSAC and least squares methods. First, RANSAC is used to remove outliers, and then least squares is used for accurate fitting to meet the accuracy requirement of ±0.1mm.

[0093] Furthermore, the present invention employs a multi-layered, adaptive strategy, specifically including:

[0094] Edge detection based on point cloud gradients: First, the spatial gradient (such as changes in normal vector, density, and intensity) of each point or local region in the point cloud is calculated to initially identify candidate edge points. The gradient calculation is optimized for the spatial distribution characteristics of the point cloud to ensure robustness under complex surfaces and noise interference.

[0095] Edge extraction through multi-feature fusion: To further improve the completeness and accuracy of edge recognition, multiple features are fused for collaborative judgment. These features include, but are not limited to, spatial geometric gradient, local curvature, point density distribution, and reflection intensity information. Through feature-level or decision-level fusion algorithms (such as weighted fusion or machine learning-based discriminators), the probability of each point being a true edge point is comprehensively evaluated, effectively suppressing noise and connecting broken edges.

[0096] Edge detection with adaptive thresholds: An adaptive threshold mechanism is introduced to dynamically determine the edge detection threshold based on the statistical characteristics of local point cloud regions (such as the distribution and density of gradient magnitudes). This method overcomes the shortcomings of fixed thresholds, which have poor adaptability to different scenes and distance regions, and achieves a balance between suppressing false alarms in uniform regions and enhancing sensitivity in feature-rich regions.

[0097] This invention integrates the following compensation mechanisms:

[0098] Velocity-induced Doppler compensation: When using lidar with a coherent detection mechanism (such as FMCW lidar), the radial velocity between the target and the radar induces a Doppler frequency shift, causing deviations in the range or position calculations of the point cloud. This invention obtains the platform's own motion attitude information (such as from IMU, GNSS) and preliminary point cloud data, estimates or directly reads the radial velocity component, and performs real-time compensation and correction of the point cloud's coordinate position based on the Doppler effect model, thereby eliminating or significantly reducing geometric deformation and edge blurring caused by motion.

[0099] This invention organically combines real-time edge detection with an error compensation mechanism to form a complete point cloud front-end processing solution. Its advantages are:

[0100] High real-time performance: The algorithm is optimized for point cloud flow data to meet online processing requirements. High accuracy and robustness: Multi-feature fusion and adaptive thresholding significantly improve the accuracy and environmental adaptability of edge extraction. High reliability: The Doppler compensation mechanism reduces motion artifacts from the data source, ensuring the geometric fidelity of edge features in dynamic scenes.

[0101] It should be noted that, in this document, 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.

[0102] 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 real-time measurement system for the dimensions of continuously rolled steel based on an active phased array, characterized in that, include: The dual orthogonal active phased array radar unit consists of two AESA radar modules deployed above and to the side of the rolling line; the upper radar module scans vertically downwards to measure the width W of the steel, while the side radar module scans horizontally to measure the height H of the steel. The dynamic beam control system is connected to the dual orthogonal active phased array radar unit. By adjusting the phase difference of each radiating unit in real time, it realizes electronic scanning and dynamic focusing of the beam within a range of ±30°. The FMCW-Doppler composite signal processing module uses a linear frequency modulated continuous wave waveform to compensate for the influence of the steel material's movement speed through Doppler frequency shift. The graded signal processing unit has three parallel processing channels built-in for roughing, intermediate rolling, and finishing rolling. It automatically switches the algorithm mode according to the pass of the steel material, that is, it calls the beamforming algorithm, MIMO virtual array algorithm and super-resolution algorithm for processing respectively.

2. The system according to claim 1, characterized in that, The AESA radar module operates in the 77GHz-81GHz frequency band, with a wavelength range of 3.7-3.9mm. The beam scanning range of a single module is ≥±30°, and the scanning switching time is ≤100μs.

3. The system according to claim 1, characterized in that, The super-resolution algorithm called by the graded signal processing unit during the finishing rolling stage is either the MUSIC algorithm or the ESPRIT algorithm.

4. The system according to claim 1, characterized in that, It also includes a data interface with the rolling mill control system, which is used to feed back the real-time calculated steel height, width and trend to the rolling mill to form a closed-loop control system.

5. The system according to claim 1, characterized in that, The dynamic beam control system employs a predictive tracking algorithm to predict the beam pointing angle at the next moment based on the steel material position and velocity data from the previous N scanning cycles.

6. A method for real-time measurement of steel dimensions using the system described in any one of claims 1-5, characterized in that, Includes the following steps: S1. Dual radar modules transmit FMCW millimeter-wave signals to the moving steel material and receive the echoes. S2. Mix, filter, and convert the echo signal to an AD converter to obtain the intermediate frequency signal; S3. Perform two-dimensional FFT processing on the intermediate frequency signal to obtain preliminary range-velocity information of the target; S4. The dynamic beam control system controls the radar beam to perform tracking and scanning based on the predicted value of the steel material speed. S5. Estimate the angle of arrival of the multi-frame echo data acquired by the tracking scan and generate a three-dimensional point cloud on the surface of the steel material. S6. Based on the steel material's pass, select the corresponding graded processing algorithm to optimize the point cloud, fitting the upper surface and side edges respectively. S7 outputs high-precision height and width values.

7. The method according to claim 6, characterized in that, The specific steps of the hierarchical processing algorithm include: In the roughing section, a beamforming algorithm is used to control the accuracy to ±1mm. MIMO virtual array technology is used in the intermediate rolling section to control the accuracy to ±0.5mm; In the finishing rolling section, a super-resolution spectral analysis algorithm is used to control the accuracy to ±0.1mm.

Images