Surrogating Online Analyzers With A Real-Time Arithmetic Prediction
Online analyzers can be an effective instrument in providing real-time information for refineries to monitor their current performance in terms of unit operations, blending, emission, etc. While being very useful in providing valuable information – they can also be expensive, time-consuming to maintain, and inaccurate. These days, a lot of technical solutions are proposed to refining operators to install online analyzers in solving a business challenge, which usually involves a cumbersome, time-consuming process including initiation, feasibility study, approval, installation, testing, and implementation.
When working with refining operators around the world, we often see the misconception that an online analyzer is a must-have solution to enable mature process control – which is not necessarily the case. With the advancement in data analytics and modeling, we have successfully supported our clients to install Online Analyzer Surrogates in the past couple of years – a Real-Time Arithmetic Prediction that does the work just as well as an online analyzer.
To enable this opportunity, there are several prerequisites that the refinery needs to achieve before installing this solution:
- Commitment to continually improving their modeling practice in their modeling software such as linear programming and blend optimizers
- Availability of a real-time data analytics platform that can retrieve key information such as lab data, unit temperature parameters, crude information, etc.
- An enthusiastic team willing to learn new knowledge and take ownership of the data analytic solution to ensure long term sustainability
A typical solution can be installed within 2-3 months without capital expenditures or any modification to existing refinery infrastructures. Our experience indicates that such a solution can often achieve a measurement accuracy as good as an online analyzer.
If you are looking for a cost-effective solution that can be installed within a short period of time to monitor your refinery’s performance on a real-time basis – this approach may just be the right answer for you.
The author of this blog Kai Y. Wan is an Engagement Manager at Trindent Consulting.