Computational intelligence in mass soldering of electronics - A survey

Highlights? An insight to the application of computational intelligence to mass soldering is given. ? Applications of computational intelligence (CI) have expanded rapidly in the industry. ? The methods used mostly in mass soldering are neural networks and fuzzy sets. ? Multilayer perceptrons have been the mostly used type of artificial neural network. ? The main application field has been the inspection of soldered joints. Mass soldering of electronic components is one of the key processes in electronics production, because it affects directly the functionality of the final product. Mass soldering, like other processes of manufacturing electronics, is constantly facing new challenges arising from the evolving production environment, increasing product variety and complexity, miniaturization of components and products, new environmental regulations, and increasing time-based competition. In the last two decades, advancements in information technology and data acquisition systems have promoted the use of manufacturing-related data in process improvement, which has also promoted the use of new computational methods and made them more applicable to industrial problems. Especially computational intelligence and its applications have expanded among the different fields of industry. The benefits of these so called intelligent methods include an ability to learn from experience, to self-organize and to adapt in response to dynamically changing conditions, and a considerable potential in solving real world problems. This survey provides an insight to the application of computational intelligence to mass soldering of electronics. The survey includes a summary on the main application fields of these methods in the past, what methods have been typically used, and what the most probable application fields will be in the future.