CN-122017030-A - Multi-product synchronous ultrasonic bubble removal and scanning analysis method

CN122017030ACN 122017030 ACN122017030 ACN 122017030ACN-122017030-A

Abstract

The invention discloses a multi-product synchronous ultrasonic bubble removal and scanning analysis method, which relates to the field of ultrasonic bubble removal, wherein a product is positioned to an ultrasonic processing tank matrix type station through a six-axis mechanical arm and a hydraulic lifting platform, a three-dimensional standing wave field directional migration broken bubble is formed through a phase conjugation self-adaptive focusing technology, detection precision is improved by integrating ultrasonic waves and phased array scanning and combining a frequency domain synthetic aperture focusing algorithm, a bubble boundary is extracted through a local entropy self-adaptive threshold and a region growing algorithm, a bubble motion track is tracked through a Kalman filtering and particle filtering fusion algorithm, a dual-attention mechanism depth residual error network intelligent classification bubble is constructed, and a closed loop system is formed by optimizing processing parameters in real time based on a model prediction control algorithm. The invention dynamically matches ultrasonic parameters to improve processing adaptability, the multi-mode scanning technology enhances micro-bubble recognition capability, closed loop feedback control ensures optimal processing effect, and provides an efficient and intelligent quality control solution for industrial production.

Inventors

ZHANG SHUGUANG
ZHANG XU
FAN YUHANG
SHAN GUIBIN

Assignees

宿迁市河海大学研究院

Dates

Publication Date: 20260512
Application Date: 20260128

Claims (10)

1. The multi-product synchronous ultrasonic bubble removal and scanning analysis method is characterized by comprising the following steps of: the multi-station positioning step is to position the product to an ultrasonic treatment tank matrix type station by utilizing a six-axis mechanical arm, wherein the station is provided with an independent hydraulic lifting table and a vacuum adsorption clamp, and a grating ruler secondary calibration position is adopted; The ultrasonic bubble removal step is to dynamically match ultrasonic frequency based on the material density of the product, construct a three-dimensional standing wave field by adopting a phase conjugate focusing technology, adjust sound field parameters according to the size of bubbles, and directionally migrate and break bubbles; integrating ultrasonic C scanning and phased array B scanning, adopting a time reversal mirror algorithm to synchronously image multiple focuses, and detecting sensitivity by utilizing an ultrasonic back scattering integral and frequency domain synthetic aperture focusing technology and a self-adaptive matched filtering algorithm; A dynamic threshold segmentation step, namely, adopting a self-adaptive threshold adjustment algorithm based on local entropy to initially segment an image, removing noise by combining morphological opening operation, extracting bubble boundaries by a region growing algorithm, and removing a pseudo-bubble region by utilizing topology analysis; The bubble motion track tracking step is to monitor the bubble removal process in real time through a high-frequency pulse sequence, calculate bubble motion vectors by means of an optical flow method, introduce a Kalman filtering prediction model, combine particle filtering correction tracks, and track high-speed motion bubbles by adopting a multi-scale pyramid optical flow algorithm; The intelligent classification decision step comprises the steps of constructing a depth residual error network, fusing ultrasonic image time domain, frequency domain and space domain information by an input layer, adopting a migration learning strategy, fine tuning in an ImageNet pre-training model, and balancing samples through a focus loss function; and a closed-loop feedback control step, namely based on a bubble dynamics model, adjusting the ultrasonic frequency, the power and the acting time in real time by using a model predictive control algorithm, and optimizing the bubble removal efficiency and the product damage risk.
2. The multi-product synchronous ultrasonic bubble removal and scanning analysis method according to claim 1, wherein the ultrasonic bubble removal step optimizes excitation parameters of a phased array transducer by adopting a quantum genetic algorithm, optimizes uniformity of energy density of an acoustic field by constructing a quantum bit coding population, and controls bubbles in a non-contact manner by utilizing an acoustic tweezer effect.
3. The multi-product synchronous ultrasonic bubble removal and scanning analysis method according to claim 1 is characterized in that an ultrasonic elastography auxiliary module is developed in a multi-mode scanning step, tissue displacement field distribution is measured by applying low-frequency modulation stress waves, an elastic modulus image is reconstructed by combining an inversion algorithm, elastic information and ultrasonic gray level image characteristics are fused, bubbles hidden in a high-density area are detected, and a stress concentration area in a material caused by the bubbles is identified.
4. The multi-product synchronous ultrasonic bubble removal and scanning analysis method according to claim 1, wherein the dynamic threshold segmentation step is conducted by preprocessing an image by a super-pixel segmentation technology, optimizing a segmentation boundary by combining a graph segmentation algorithm, and constructing an energy function by calculating the gray difference and texture similarity of adjacent super-pixels.
5. The multi-product synchronous ultrasonic bubble removal and scanning analysis method according to claim 1, wherein the bubble motion trajectory tracking step adopts an instance segmentation network to separate targets for the overlapped bubble scene, the overlapped bubbles are segmented by training a synthetic bubble image dataset, and the separated bubble motion trajectories are matched by a trajectory correlation algorithm.
6. The multi-product synchronous ultrasonic bubble removal and scanning analysis method according to claim 1, wherein the intelligent classification decision step designs an integrated transfer learning and integrated learning hybrid model, integrates a pre-training EFFICIENTNET-B7 model and a random initialization ResNet-34 model, and fuses model prediction results through a stacked generalization strategy.
7. The multi-product synchronous ultrasonic bubble removal and scanning analysis method according to claim 1 is characterized in that the closed loop feedback control step establishes a digital twin body real-time simulation system, utilizes finite element analysis to establish a multi-physical field coupling model, combines ultrasonic echo data to calibrate model parameters, predicts different parameter combination treatment effects through the digital twin body, and optimizes parameters and energy consumption.
8. The multi-product synchronous ultrasonic bubble removal and scanning analysis method according to claim 1, wherein the multi-station positioning step adopts a multi-sensor fusion positioning technology, combines laser radar, visual identification and inertial navigation data, and optimizes the positioning time of a high-speed motion product and controls positioning errors by expanding Kalman filtering estimation states.
9. The multi-product synchronous ultrasonic bubble removal and scanning analysis method according to claim 1, wherein a quality traceability system is constructed, a alliance chain architecture is adopted to hash and uplink product material information, processing parameters and detection results, and a zero knowledge proof technology is utilized to protect data privacy.
10. The multi-product simultaneous ultrasonic bubble removal and scanning analysis method of claim 1, further comprising: and the man-machine collaborative optimization module is used for generating a parameter optimization scheme by adopting a reinforcement learning algorithm, and providing optimization requirements through natural language interaction, and automatically converting manual experience into model parameters.

Description

Multi-product synchronous ultrasonic bubble removal and scanning analysis method Technical Field The invention relates to the technical field of ultrasonic bubble removal, in particular to a multi-product synchronous ultrasonic bubble removal and scanning analysis method. Background In modern industrial production, the problem of bubbles inside the product seriously affects the quality and performance of the product. Taking products such as optical lenses, precision castings, polymer composite materials and the like as examples, residual bubbles in the interior can lead to the decrease of the material strength, the deterioration of the optical performance and even the initiation of functional failure. The traditional ultrasonic bubble removal technology generally adopts ultrasonic waves with single frequency and fixed power to act on products, and is difficult to adapt to the characteristic differences of products with different materials and different structures. For example, for higher density metal castings and lower density engineering plastics, uniform ultrasonic parameters are not only ineffective in removing bubbles, but also may result in product damage due to energy concentration. Meanwhile, a single bubble removing process lacks a real-time monitoring mechanism, so that whether bubbles are completely removed cannot be judged, and the conditions of insufficient treatment or excessive treatment are easy to occur. There are significant limitations in the prior art in terms of bubble detection analysis. The conventional ultrasonic scanning technology has low resolution, is difficult to identify tiny bubbles (with the diameter smaller than 0.1 mm), and has a blind area for detecting the internal bubbles of a product with a complex structure. For example, when detecting a multilayer composite material, an ultrasonic echo signal is easily interfered by an interlayer interface, which leads to erroneous judgment or omission of bubbles. In addition, the traditional detection method mostly adopts single-mode imaging, only relies on an ultrasonic gray level image for analysis, lacks comprehensive judgment capability on physical characteristics (such as elastic modulus and internal pressure) of bubbles, and cannot accurately evaluate potential influence of the bubbles on product performance. Along with the development of industrial production to automation and intellectualization, the demand for parallel processing of multiple products is increasingly urgent. However, most of the existing bubble removal and detection devices are designed in a single station, so that the high-efficiency processing requirements of batch products are difficult to meet. Even if multi-station equipment exists, cooperative control mechanisms are lacked among stations, and processing parameters cannot be dynamically adjusted according to product characteristics. Meanwhile, the data processing and analysis links are relatively lagged, the detection results are mainly manually interpreted, an automatic classification and decision system is lacked, an effective quality evaluation report is difficult to generate rapidly, and the production efficiency and quality control level are severely restricted. Disclosure of Invention The invention provides a multi-product synchronous ultrasonic bubble removal and scanning analysis method, which aims to solve the problems in the prior art. In order to achieve the purpose, the invention adopts the following technical scheme that the multi-product synchronous ultrasonic bubble removal and scanning analysis method comprises the following steps: the multi-station positioning step is to position the product to an ultrasonic treatment tank matrix type station by utilizing a six-axis mechanical arm, wherein the station is provided with an independent hydraulic lifting table and a vacuum adsorption clamp, and a grating ruler secondary calibration position is adopted; The ultrasonic bubble removal step is to dynamically match ultrasonic frequency based on the material density of the product, construct a three-dimensional standing wave field by adopting a phase conjugate focusing technology, adjust sound field parameters according to the size of bubbles, and directionally migrate and break bubbles; integrating ultrasonic C scanning and phased array B scanning, adopting a time reversal mirror algorithm to synchronously image multiple focuses, and detecting sensitivity by utilizing an ultrasonic back scattering integral and frequency domain synthetic aperture focusing technology and a self-adaptive matched filtering algorithm; A dynamic threshold segmentation step, namely, adopting a self-adaptive threshold adjustment algorithm based on local entropy to initially segment an image, removing noise by combining morphological opening operation, extracting bubble boundaries by a region growing algorithm, and removing a pseudo-bubble region by utilizing topology analysis; The bubble motion track tracking step is to