Machine learning in manufacturing: advantages, challenges, and applications

Abstract

The nature of manufacturing systems faces ever more complex, dynamic and at times even chaotic behaviors. In order to being able to satisfy the demand for high-quality products in an efficient manner, it is essential to utilize all means available. One area, which saw fast pace developments in terms of not only promising results but also usability, is machine learning. Promising an answer to many of the old and new challenges of manufacturing, machine learning is widely discussed by researchers and practitioners alike. However, the field is very broad and even confusing which presents a challenge and a barrier hindering wide application. Here, this paper contributes in presenting an overview of available machine learning techniques and structuring this rather complicated area. A special focus is laid on the potential benefit, and examples of successful applications in a manufacturing environment.

Predictive Manufacturing Systems in Industry 4.0: Trends, Benefits and Challenges

“…The development of the algorithms that are able to learn from and to handle different situations is still a great issue which can be considered as a general challenge not only for the PMS, but for entire Industry 4.0 [3,13]. Also, the collected set of data can contain an irrelevant and redundant information which can have an impact on the performance of the learning capabilities of PMS, and today's generation of PMS can only operate with continuous and nominal data values [27]. Another major concern is data security in PMS since the big data are stored in virtual cloud platforms [15].…”

“…[1] and Wuest et al [27] pointed out, the PMS represents a very complex system for managing that requires development of appropriate planning and explanatory algorithms and development of a single set of common standards which are still the big issues for the scientists. The planning and explanatory models will provide a basis for managing complex manufacturing systems, and standardisation will support a collaboration and technical description of standards.…”

“…The new knowledge supports the process of decision-making or it is used to improve the system directly. The end goal of machine learning techniques is detection of the patterns or regularities that describe the relationships and structure between the elements [27].…”

“…The reinforcement learning refers to the kind of learning where less feedback is given, because knowledgeable expert has to evaluate which of the chosen action gives the best results [27,28]. The supervised learning is machine learning where the correct response is provided by the knowledgeable expert, while the unsupervised learning is intelligent learning where evaluation of the action is not provided, because there is no knowledgeable expert [26,27]. According to Wuest et al [27], the supervised learning is the most commonly used, because it is the best fit for most manufacturing systems due to the fact that the majority of applications can provide labelled data.…”

Machine Learning in Chemical Industry

“…The chemical manufacturing industry today is facing an increasing volume of data which compromise a variety of different formats, semantics, and quality. ML techniques have been successfully utilized in various process optimization, monitoring, and control applications in manufacturing [5]. ML can help manufacturers find solutions faster.…”