Machine Learning for condition monitoring of production equipment

点击翻译按钮获取中文摘要

My project will help manufacturers increase the lifetime of their production equipment and keep it in use longer - when compared with traditional 'inspect and replace' maintenance schedules.By continually monitoring the condition of equipment and giving realtime intelligent insights into ongoing performance, manufacturers will be able to make informed decisions on when to replace components on their equipment, if there is a need to change production schedules or to reduce operating limits to increase the lifetime of the equipment being monitored.Underpinning our solution is a proprietry Artificial Intelligence and Machine Learning software system that will use data obtained from sensors (attached to production equipment) to spot changes in behaviour, and highlight potential points of failure before they occur. The sensors (such as vibration, voltage, heat, airflow, airborne particulates etc) will capture data that will be represented on a digital dashboard and show the current operating performance along with any boundaries that have been set (such as operational or legislative limits) and highlight with labels what is assumed to be going wrong - such as a bearing failure, filter blockage or shaft misalignment.This solution in itself is not unique, but what is innovative is that it can operate without an internet connection and is fully contained within the manufacturing facility. Other systems of this nature rely on cloud computing for the AI analysis of the sensory data, where data is sent across the internet to a remote computer and the analysis returned the same way. However, in some installations, such as those concerned with defence or national security, no data will be allowed to leave the facility and will therefore incapacitate a cloud based solution.Our approach, by contrast, handles all of the Artificial Intelligence processing from within a supplied server that is placed within the premises and has no need to send data elsewhere for processing. Making it free from transmission lag, external security breaches and can be operated in locations where there is no internet connection at all.