Contributed by: Abhinash Saikia
This is Abhinash. I completed my M.Tech in Manufacturing and B.Tech in Mechanical from KIIT University, Bhubaneswar, in 2015. I am working as an Asst. Manager, Mechanical Engineering with a leading pioneer in the Indian Textile Industry since July 2015. I look after engineering-related activities like breakdown and preventive maintenance management, inventory management of spares. Along with my regular activities, I am also a part of the Project Management team related to Energy Conservation, and Sustainability.
Utilities and energy consumption form a major chunk of the total manufacturing budget in any plant (around 30-35%). Therefore, it is important to understand the trend of energy consumption- be it equipment, process, or system-level and study the various variables affecting it to predict the expected energy consumption in the future and plan energy conservation initiatives accordingly. This is where I had applied the concept of Linear Regression to predict the future energy consumption of different machines with the help of past data.
The Industry plant is ISO 50001 (Energy Management System). Every year, the plant takes different targets to be achieved in energy conservation to solidify its certification.
To understand better, Linear Regression has been applied on all SEU (Significant Energy Users) equipment (Compressors, Boilers, Textile washing machines, weaving looms, etc.) to predict the energy consumption of the machines. Once the baseline is set, different initiatives are taken to achieve the targets, i.e., reducing energy consumption from the expected numbers.
Previously, there was no particular direction guiding the energy management system. However, the introduction of the linear regression concept has helped streamline the system to a great extent.
As mentioned, previously there was no proper energy monitoring system for the machinery. The introduction of the linear regression concept has helped streamline the system to a great extent. As a result, we are able to analyze factors affecting the pattern of energy consumption of every piece of equipment at a detailed level.
I will be giving an example of Drying and Heat Setting where a machine called Stenter is used. There are six stenter machines running simultaneously in the finishing department of the plant whose main purpose is drying the fabric at 180 to 200 degrees C.
It consumes a lot of electricity and coal as hot oil circulates the machine to facilitate the drying of the fabric. Therefore, it was decided that since it is a high energy-consuming system, these machines should be studied and analyzed to look for the scope of energy reduction.
Linear regression was applied using MS Excel.
Several variables were taken for trial and testing to understand the electricity consumption pattern of the stenter machines. It was found that the running hours and electricity consumption of Harish stenter machines is an important variable.
Basically, there are six stenters running simultaneously, out of which one of them is of Harish stenter make. Harish stenter is an old technology stenter machine which consumes a higher amount of electricity. In FY 2018-19, harish stenter contributed to around 23.5% of the total energy consumption from stenters.
Based on the analysis, it was decided to take a decision on the harish stenter machine to save energy.
It was decided that the production load on Harish stenter will be decreased during the upcoming year FY 2019-20. More load should be given to the other five stenter machines as they were equipped with modern technology and consume lesser energy comparatively. Harish stenter should be run as little as possible.
There was an immediate impact on the electricity consumption of the stenter process.
In FY 2019-20, harish stenter contributed to around 12.7% of the total energy consumption from stenters. This has helped reduce the energy consumption in the stenter process by 3.62% annually, which saved Rs 4.5 Lakhs/year. This was achieved without making any investment.
This exercise has helped me understand the importance of linear regression in predictive modeling. It showed that with the correct analysis, lots of benefits could be obtained.
The example I have given is only of one system. There are more than 80+ systems.
(compressors, boilers, textile machinery, department-level energy consumption, etc.) where regression analysis is done under the energy consumption pattern. Based on the observations, different projects are carried out to achieve energy conservation goals.
Here are a few graphs, plots, and charts to help you understand more about my work.
If you wish to learn more such concepts and also apply your learning at work, you can take up the PGP Data Science and Business Analytics Course by Great Learning. The course offers mentorship sessions and online classes to help you learn from the comfort of your home. Great Learning also offers career assistance.
For more such success stories, watch this space.
