Solving for Engineering Solutions

By Kock Chee Kiong, ARM, ARiMI-CERM.

Data science and analysis are more developed in recent years. Accuracy of the algorithm depending on the number of questions, type of questions and engineering words used. Providing mixed integer answers toward problem solving.

The traditional engineering and maintenance philosophy relies on preventive and predictive based monitoring. Measurements of temperature and vibration on machines are used to trend the performance of the equipment. Abnormalities such as noise, vibration and temperature are the result used to determine the performance of machinery and equipment, and the need to replace. Result is obtained by trending past performance of the equipment, highlighting the abnormal operating condition when compared with past results.

Ways to do things better and faster.

Correlate machine and equipment perform with designed specification and actual operating condition.

Getting real time information by providing input parameters on drivers such as voltage, current, revolutions vs pressure, flow rate and temperature.

Innovate solutions (solutions to manage issues or solving problems)

Solutions should be customized and the approach to tackle the problem would vary from finding root cause, mitigate severity of the problem, creating temporary measures and towards preventing future issues from arising. Deriving these solutions and the approach towards problem solving very much depend on approaches taken to tackle the issue identified and experience level of personnel.  Data analysis helps to move a level beyond trending of equipment that enable predicting the operation and equipment functionality.

Using information and database, the issues can be attended by personnel in real time and thus making better engineering decision.

Think and Analyze (Heuristic learning for machines)

In teaching machines to do the guess work would shed light of the unknowns/uncertainty and enable human user to make better decision or allowing autonomous remedy actions to be taken by machines.

Methods in enabling machine learning

    • Data mining follows by data extraction and correlation.
    • Extracting insights, key information from data and problem statement.  Routine works are just checked and verified by the system.
    • Trending and learning after data analysis and detection on anomalies. Set decision making after issues detected. Human intervention is required to teach machines to performing non- repetitive work.

Leave a Reply

Your email address will not be published. Required fields are marked *