Data analysis for energy management
Fortress Specialty Cellulose, 2016 – 2017
The Kraft plant in in Thurso produces specialized cellulose for, among other things, the textile manufacturing industry. The plant’s steam system consists of four pressure levels, a 24 MW cogeneration system and four boilers producing approximately 250 t/h of high-pressure steam. This process and its systems are very complex and require advanced data analysis techniques for energy management.
3E Eng. provided an expert in on-site data analysis and energy efficiency supported by three other experts to carry out the following mandates over a two-year period:
- Analysis of multivariate data from energy systems to identify causes of inefficiency
- Programming machine learning algorithms and energy management tools (control maps, tracking causes of inefficiency, Pareto diagram)
- Programming key performance indicators in Pi Process Book and Pi Asset Framework among other things
- Modelling of energy systems (on CADSIM Plus and Excel)
- Identification and implementation of energy-saving measures
- Setting up an energy management system ISO 50001
- Training an energy manager
All of the studies have identified more than 1M $/year opportunities to increase the profitability of the cogeneration system without investment.
About data analysis and machine learning for energy management
Data analytics is a powerful tool to improve energy management in industry. In a context where processes are complex and data to monitor is numerous, it can quickly become difficult to properly track all important indicators. Data analysis identifies important parameters and ensures that these parameters are normal. This allows you to react quickly in the event of a drift or avoid reacting unnecessarily when you are in normal. Advanced analytical tools can also identify and correct inefficiencies in existing systems with very little investment.