Testing Product Lines Research Articles

Cross-Domain Transfer Learning for Galvanized Steel Strips Defect Detection and Recognition

Abstract Defect detection is a crucial direction of deep learning, which is suitable for industrial inspection of product quality in strip steel. As the strip steel production line continuously outputs products, it is necessary to take corresponding measures for the type of defect, once a subtle quality problem is found on steel strips. We propose a new defect area detection and classification method for automation strip steel defect detection. In order to eliminate the way of insufficient data in industrial production line scenarios, we design a transfer learning scheme to support the training of defect region detection. Subsequently, in order to achieve a more accurate classification of defect categories, we designed a deep learning model that integrated the detection results of defect regions and defects feature extraction. After applying our method to the test set and production line, we can achieve extremely high accuracy, reaching 87.11%, while meeting the production speed of the production line compared with other methods. The accuracy and speed of the model realize automatic quality monitoring in the manufacturing process of strip steel.

Anomaly Candidate Extraction and Detection for automatic quality inspection of metal casting products using high-resolution images

This paper proposes an AI-based detection method called ACED (Anomaly Candidate Extraction and Detection) that can find tiny defects efficiently by inspecting the surface of cast manufactured products through high-resolution images. A major challenge is to minimize false detection to a level competent to skilled human inspectors. Also, high-resolution imaging is necessary to capture meaningful information about tiny flaws, but it may cause overfitting of AI models and increase computational burdens. We have designed ACED in a few steps to overcome the challenges: first, we create image partitions using factory-defined inspection area information and create regions of interest containing the most abnormal spots found by a convolutional autoencoder. Next, we filter out regions with minor flaws using the specification of significant flaws, where the information is available from a production line. Finally, we train an ensemble of convolutional neural networks to classify the defect candidates as normal or defective. We implemented and tested the proposed method ACED for two cast-manufactured automobile parts in a test production line of our collaborator’s factory. Our experimental results showed that ACED was particularly better in reducing false positive detection, outperforming other AI-based detection methods by ×10.5∼×43.0 in terms of accurate rates and ×5.3∼×18.2 in F1-scores, depending on the tested products. Also, ACED was competent to human inspectors, achieving ×1.9∼×4.0 performance gains in F1-scores. We also provide analyses of execution time, resource utilization, and inspection time reduction of factory workers, showing the potential of ACED to be used in actual production lines.

Software product line regression testing based on fuzzy clustering approach using distance method

Testing is a process that takes much time and effort in software companies. This becomes even more difficult and boring when it comes to testing a software product line (SPL). The SPL is a model in which multiple products from the same family are made simultaneously. Testing of all products is not possible. Hence a lot of testing methods have been given from time to time to test the product line, given by researchers based on contemporary conception. In the direction of testing product lines, this article has proposed a method, which used fuzzy C-means clustering with the Jaro-Winkler distance method. Variable features of the product form the basis for cluster development. The proposed method is compared with other distance methodologies. After comparison, it is concluded that the proposed method provides better results than other methods. This article has resorted to some product lines to compare with the proposed methods.

Rath installs test facilities and production line at Meissen site

Rath has invested in advanced test facilities and a second production line at its Hot Gas Filtration Competence Center in Meissen, Germany.

Performance Evaluation and Improvement of Chipset Assembly & Test Production Line Based on Variability

“Factory physics principles” provided a method to evaluate the performance of a simple production line, whose fundamental parameters are known or given. However, it is difficult to obtain the exact and reasonable parameters in actual manufacturing environment, especially for the complex chipset assembly & test production line (CATPL). Besides, research in this field tends to focus on evaluation and improvement of CATPL without considering performance interval and status with variability level. A developed internal benchmark method is proposed, which established three-parameter method based on the Little′s law. It integrates the variability factors, such as processing time, random failure time, and random repair time, to meet performance evaluation and improvement. A case study in a chipset assembly and test factory for the performance of CATPL is implemented. The results demonstrate the potential of the proposed method to meet performance evaluation and emphasise its relevance for practical applications.

Cycle time prediction and improvement of chipset assembly and test production line based on variability

Variability plays a very important role in the system performance in a chipset assembly and test production line with stochastic environment. So far only few variations have been considered in chipset assembly and test production line performance prediction and improvement. In this paper, a novel approach about cycle time prediction and improvement is proposed by combining with internal benchmarking and variability. Firstly, the composing of cycle time segment through queueing model based on Factory Physics is established, which is under the condition variability quantification and machine sudden large failure. Secondly, to obtain the optimal direction of the cycle time, an internal benchmarking is applied to analyze the influence of variability on cycle time among the workstations. Finally, a case study is performed to evaluate the effectiveness of the proposed method. Results show that the proposed method is an effective method to predict and improve the cycle time of the chipset assembly and test production line with variability.

An Engraving Character Recognition System Based on Machine Vision

In order to solve the problem of low efficiency, low quality and false detection of metal surface engraving character detection, we proposed a kind of engraving character recognition system based on machine vision. This system can identify and detect through the image acquisition of engraving area, and a series of mathematical and morphological treatments of binary image. It can extract the contour using optimized Laplace edge detection operator and combining with the convexity characteristic of engraved words. The system successfully carried out the experiment of engraving character recognition and it can detect the character recognition with high efficiency and high qualified rate. It has realized the automation of engraving character recognition test production line, and it has certain application value and promotion significance.

Dynamic Reentrant Scheduling Simulation for Assembly and Test Production Line in Semiconductor Industry

Production scheduling in semiconductor production line is a complex combinatorial optimization problem. It is featured by reentrant production, product variety, frequent machine breakdown and especially reentrant phenomenon and dynamic randomness, which conduces enormous complexity to production line scheduling and management. Based on the resource conflict resolution strategies, dynamic scheduling models for a semiconductor production system are proposed here, aiming at finding maximum machine utilization, optimal man-machine ratio (MMR) and maximum output. Then the validity of the model is illustrated by a simulation case.