The Fertilizer Industry Gains Actionable Timely Insights with Advanced Analytics –
Part 2 of 2
In Part 1 of this blog series, we gave a brief overview of the history of fertilizer and then went on to talk about the industry’s challenges and how advanced analytics could help. To catch up to speed and start this blog with a bit of background and context, have a look: The Fertilizer Industry Gains Actionable Timely Insights with Advanced Analytics – Part 1 of 2.
For Part 2 of 2, we take it a step further by looking at a generic process workflow using advanced analytics and then discuss a real use case with OCI Nitrogen. But first, we’ll tell you about a generic advanced analytics workflow.
What a Generic Process Workflow Looks Like Using Advanced Analytics
If you’re thinking about doing analytics in your plant because you want to make data-driven decisions to optimize operations, you don’t have to wait until you have a 100% complete infrastructure to start. You can start with the data your plant has already been collecting, the time-series data that is usually found in your plant’s historian. So, for a generic workflow, here are some general steps to approach a process data analytics problem using advanced analytics.
- You need to first know what process issue you want to analyze. What has happened?
- Once you have this information, you next need to know if this issue has happened before because if it has, you can derive insight from past occurrences. And you want this important information. TrendMiner’s filtering capabilities allow process experts to isolate relevant time periods for further data analysis. Filters can be easily set up manually or automatically using its dynamic search capabilities to quickly isolate data by startups, shutdowns, grade transitions, specific product campaigns for TrendMiner’s deeper statistical analysis and machine learning.
- If you find there have been similar occurrences, you can move one step further and ask the software to help you find the root cause. (That’s one of the main points of investigating process issues anyway.)
- Then, when you find the root cause, you can set some fingerprints of good operational zones to monitor the process. Additionally, the bad operational zones can be included in this fingerprint so when the larger issue occurs, process experts can understand what is happening. With these in place, you will receive early warnings for process deviations, for when your parameters are going out of range or for when there are problems coming about. This lets you know that you need to take corrective action. With these early warnings, you can notify the key stakeholders, for example with an email message, so that they know where to react and how to react.
- Also, apart from sending a notification to these stakeholders, you can capture all of this information within the tool itself as an event, so you can contextualize your data. Meaning, if you have time-series data, you can automatically or manually create an event which can include a comment describing the problem that happened in that instance, so that the next time you look at the data, you already have a bit more context on why you are seeing the kind of trends you are seeing. If you’ve manually created the event, you can also include communication or attachments. All of this information is to keep records straight and give process experts within different teams the information they need to work efficiently.
- Lastly, all the information from your analysis can be visualized in a “production cockpit”. This dashboard gives you a live overview of your process, showing you the live state, the live trends of the process. Essential to staying on top of your production.
This is a simplistic explanation of how you or any process expert can use advanced analytics to gain production insights themselves without even having a data science background. But you can get a better understanding of this by watching Empowering Engineers with Advanced Analytics. In this webinar, TrendMiner engineer Ruchika Tawani gives a short demonstration of this type of workflow using our software.
The example she covers deals with an unexpected temperature spike in a reactor. The product in question is extremely temperature sensitive, so an increase in temperature is definitely very bad for the product quality. The process experts did not want to let this issue go unnoticed, so they needed to find the root cause. They also wanted to prevent such occurrences in the future to avoid consequent issues and losses. Now for a real-life use case.
OCI Nitrogen Use Case:
Carbon Dioxide Washing Unit
Our customer OCI Nitrogen had a problem in the carbon dioxide washing unit in their ammonia process. After washing, their process experts often observed carbon dioxide peaks in the washing column. When carbon dioxide peaks, it causes a very high temperature in the downstream process which adversely affects the product quality. The team tried to analyze this problem using the data they had but weren’t able to find a root cause that could explain why this was happening. They also tried to solve this issue by increasing the steam consumption, but this did not solve the problem either.
Using TrendMiner, the team of process experts tackled this problem. They used the Search functionality to search through production data to compare periods where there had been carbon dioxide peaks with periods where operations were normal. Through the layer comparison, they discovered that many variables were influencing the process and thus causing the the carbon dioxide peaks. The team was able to solve this problem and decrease the number of carbon dioxide peaks by lowering the pressure before the methanation step.
OCI Nitrogen was able to achieve a more stable operation and increase production. They were also able to reduce the steam consumption which improved their carbon footprint as well. In terms of gained value, they were able to get a 5% increase in the revenue which represented approximately 2.4 million Euros per year.
Sitech Process Engineer for OCI Nitrogen
The data-driven discovery monitoring and predictive capabilities of TrendMiner were a game changer in our constant quest for operational excellence.
Advanced Analytics for Everyday Use by Process Experts
Growing the analytics maturity of your team is essential to keeping your fertilizer production competitive and sustainable to meet the world’s crop needs. And now there is an advanced analytics tool that is especially designed for your plant managers and process experts to let them do their everyday process analyses themselves.
With this approach, you can bring about quick contributions to improvements with little time investment and with significant economic benefits. Another added benefit, our software will enable you and your team to collaborate globally, connecting your people and your teams to further optimize your operations. As Gerd Decramer said, “TrendMiner is a game changer.”