Mitsubishi Electric AI Laboratory has unveiled its MELSOFT MaiLab AI solution, which apparently allows for optimised operations on manufacturing lines while maximising equipment monitoring, visibility, and diagnostics.
According to Mitsubishi, shortages in resources as the budget needed to hire skilled data analysts and AI specialists, and the capability and time to process large volumes of data with which to develop accurate predictive models, are restricting development in manufacturing. Improvements in data analytics could provide actionable insights into the performance and status of machines and processes, it says, but these smart processes have not yet been reached in factories.
In response, MELSOFT MaiLab is intended to become a virtual alternative to an AI data scientist. Said to require little training to function, it apparently utilises machine learning to automate the gathering of data across a variety of systems, as well as predictive model creation, analysis, and the mining of large volumes of data. It then bases its recommendations on such intelligence, meaning that its operators do not need to be experts in the field.
Therefore, the solution is intended to remove any barriers between manufacturers and a transition into Industry 4.0 technology, which is thought to improve the production process with a quick return on investment. It is reportedly easy to install, as it can be accessed from a browser-based environment, can run on any industrial PC, and does not require any additional software to function.
Its user interface is also designed to be intuitive, with such features as web-based visualisations and step-by-step guidance intended to facilitate easy use for operators; it also supports them through all the phases of a data analysis project and helps them to understand the implications of the resultant data, according to the company. End goals can be selected by users, from which datasets are processed and analysis models created using Mitsubishi Electric’s Maisart AI.
Mitsubishi also states that the MELSOFT MaiLab can be developed to support various different application scenarios and can be tailored to be compatible with individual setups. It can also be used in off-line mode to feed existing empirical data and develop or refine predictive models, customisable using open Python scripts. These can subsequently be used for real-time diagnostics, which provide operators with insights on the status and performance of the line and how it can be optimised.
Additional information and functions to address the requirements of other departments and SMEs are also available through the system, says Mitsubishi. It is claimed that the information processed by the system is used to increase the accuracy of its algorithms, enhance its outputs, and help companies increase their productivity of time. Flexible licensing schemes are also reported to be available to address specific companies’ needs, including yearly subscription models or one-off payment options.
A survey conducted by the Brazilian National Confederation of Industry (CNI) revealed a 21% increase in digitalisation in the Brazilian manufacturing industry between 2016 and 2021. However, the small range of technology employed on factory floors and confusion about where to begin with the transition into digital solutions were named as roadblocks by packaging journalist Liliam Benzi.
Dr. Tim Foreman from Omron Industrial Automation previously spoke to Packaging Europe and named three key issues in the implementation of AI – the ways it can improve processes and production, and if cloud or edge computing should be implemented.