During the processing of castings, various production defects can occur during the molding, casting and solidification of the material. These are generally referred to as casting defects. If they are not detected, weak points can occur in components that lead to defects in the finished product. Significant quality losses or even safety defects are the consequences. In most cases, these production errors have to be rectified through extensive reworking. Committee can also be the result of casting errors. In all cases, high costs are incurred that can be avoided by automated quality inspection with AI.
Automatic casting defect detection
in production with AI.SEE™
The casting defect - a diverse defect pattern in castings
As a form of production defect in the casting of melts, the term casting defect occurs in connection with various defect patterns:
Detection of casting defects in production through AI
These gas inclusions are particularly fatal if they occur inside the casting and only come to light after milling or turning off the surface or even as a defect at the end customer.
However, as it is not possible to completely avoid casting defects in most castings, ways must be found to integrate reliable control into the production process.
Many production facilities still rely on the use of trained employees for this step of quality assurance. Even though they are usually highly qualified and extensively trained, human visual quality assurance is error-prone, expensive and often too slow to ensure an optimal and efficient production flow. In addition, it is very difficult to ensure reliable, consistent and objective results due to fluctuations in the concentration and personal condition of employees.
On the other hand, there are rule-based computer vision systems. These often cannot be scaled efficiently due to the large number of surface structures and the variety of defect patterns. In addition, very short cycle times also pose a major challenge for automated optical systems for surface inspection. Suitable evaluation software based on state-of-the-art deep learning technologies (AI) is therefore crucial in order to efficiently obtain usable results and integrate quality assurance into the ongoing production process.
trusted by industry leaders
Automatic detection of casting defects leads to an increase in production quality
Automatic casting defect detection therefore brings significant efficiency gains and leads to a measurable increase in quality in production.
Automatic casting defect detection therefore brings significant efficiency gains and leads to a measurable increase in quality in production. In order to avoid fluctuations in the quality of the end product, every casting produced must be analyzed in the production chain and critical casting defects must be detected extremely reliably and quickly. In addition, clear identification and comprehensive documentation of these defects are essential. The documentation of the application serves as the basis for measures to feed existing blowholes or bubbles and prevent the development of new casting defects. However, for a fully automated software-based solution to improve the existing production process, it must be possible to implement the application in the existing system. elunic is therefore developing modern solutions for Industry 4.0 that integrate the detection of casting defects as an automated step in production in a cost-efficient and fluctuation-free manner. In this way, a high expenditure of time in employee training is to be avoided and fluctuations in human condition reduced. Efficient processes therefore enable a significant increase in the quality of the end product.
AI.SEE™ - the efficient solution for casting defect detection
With the automated quality control AI.SEE™, elunic offers an application specially programmed for the respective production that can analyse each part fully automatically and check for all types of outward casting defects. With its simple and extremely intuitive handling for every quality assurance employee, AI.SEE™ is the optimum solution for the visual quality inspection of all metal or other cast components, reliably recognising and documenting defects. Extensive feeding with unambiguous training data makes it possible to create an adaptive learning model based on various defect patterns and reject criteria, which is constantly evolving and is thus able to reliably recognise even very small or changing casting defects. New defect patterns in ongoing production or even entire product series can be integrated into the system in just a few steps. Simple and intuitive operation of the application ensures that every quality assurance employee can feed the application with training data. The open system also opens up possibilities for other departments within the company. These are also able to develop models or pipelines and integrate them into the system. Extensive data collection and analysis of the cloud-based software enables comprehensive tracing of fault-prone components and their revision. In addition to measures to optimise future production batches, the implementation of real-time monitoring can also offer numerous opportunities to reduce the number of rejects caused by casting defects in the current batch.
Are you interested? Then find out now about the numerous possibilities for optimum detection of casting defects with AI.SEE™!
Learn more about the use case.
