PPCL Logo Process Plant Computing Limited

Tel: +44 1753 893090 |

LinkedIn Twitter

Capturing and Replicating Optimal Process Operation

 

The key to repeating best operation is understanding your process behavior. This is necessary to achieve higher throughput, lower costs, least waste and minimum downtime. While some of this is known in the process design, much of what is taken as true about your process is likely wrong or contradictory. The resolution to these conflicts and a true process picture are already present implicitly in your plant historian, built from years of experience. This webinar will show you visual tools which engineers can use to extract and exploit the wealth of information available.

An operating envelope is the relationship between variables that gives the best achievement of one or more objectives such as yield, throughput, quality, KPI attainment or cost. This can be very complex when multiple optima exist and, until now, very difficult to see with more than a handful of variables. In the past, targets would be found individually, discounting the inherent relationship between the variables. Doing this is tedious and it is difficult to move them in accordance with new or modified goals through KPI targets and would often conflict with other limits such as safety alarms, integrity operating limit and operator alarm limits.

Feed-to-product operating envelopes make it easy not only to reproduce best operation by identifying key variables and operating ranges to ensure consistent operation or provide MPC constraints, but also to generate new information by comparing operation under different external conditions to question why it is where it is. This can be done graphically without advanced statistical or analytic techniques.

In this webinar we demonstrate a straightforward visual method of quickly finding consistent limits for operating windows which capture desired performance for many variables at once. This can be repeated for different objectives so that conflicts between different target operations can be seen before they are implemented. You can even see how that newly proposed KPI and its limits will perform and whether the operating window is sufficient to achieve the required targets. 



Geometric Process Control – what is it and why do I need it in my plant?



Geometric Process Control (GPC) is a technology developed here at PPCL. It lets you see hundreds of process variables and tens of thousands of observations from process and other plant data in a single parallel coordinates graph with supplemental time-trends, distribution plots, and Pareto plots with every individual point traceable through all the views. It comes with an extremely powerful graphical Boolean query capability that lets you extract feed-to-product operating windows and operating envelopes to help you explain why yesterday was such a great day for your KPIs and what went wrong last week when the fractionator column flooded five times.

Why is this important? Well, it lets you benefit from a few hundred thousand years of evolution and use your amazing human powers of pattern recognition and interpretation to determine cause-and-effect relationships – as well as correlations - in multi-variable problems where previously you had fewer than 10 variables to work with. You also would have needed some heavy-duty mathematics, which requires first an expensive statistician and then a hypothesis to test and which tends to introduce new complications when trying to explain your results to others.

Having determined your best operating envelope from process history, GPC creates a real-time model complete with graphical operator interface for process operator guidance and an OPC Client to link to your DCS or historian to keep your process operating in the best operating envelope you have chosen far into the future.

People ask is GPC Machine Learning (ML)? Probably, with the caveat that definitions of ML vary. Is it Artificial Intelligence (AI)? We don’t think so, but it might feel like it to process operators and certainly might be if we added neural net algorithms. GPC is based on a novel method of capturing human intelligence. It can be very powerful for identifying training data for AI and for understanding how the black-box models produced by AI have arrived at their results.

In this webinar we show you how GPC unifies process control, quality control, KPI achievement, operating limits and operator alarms into one readily understandable framework. A huge step forward for chemical engineering and you won’t need more than basic maths!



Using Visualization to Analyze Big Data from Process Plants



Process plants generate thousands of variables of continuous time-series data. This data has enormous potential which is largely untapped by the conventional analysis tools available to most process and control engineers. Fashionable “Big Data” approaches are challenged by process plant data and have limited application for busy engineers since many of the assumptions and simplifications destroy the richness of process data. Geometric Process Control (GPC) – a technology unique to us here at PPCL – avoids these pitfalls and provides engineers with graphical tools to work with datasets spanning their entire plant and create low-cost, equation-free predictive models to develop new process understanding quickly and easily.

This webinar demonstrates our unique approach to analysis on a process with a medium size dataset spanning the process from feed to product, with 750 variables over a year at 10 minute intervals. We’ll discuss how to approach big datasets and explore them visually, using operating envelopes and finding interactions between variables. Covering the entire process including incoming analyses through processing conditions to final quality variables, KPIs and performance variables, GPC enables engineers to explore their data fully and make discoveries that they couldn’t before. We’ll also cover process stewardship, using your discoveries to achieve quality targets and operational excellence long into the future.


 


PPCL © Copyright 2011-2019. All Rights Reserved. | Terms of use | Privacy Policy | P.O. Box 43, Gerrards Cross, Buckinghamshire, SL9 8UX. UK.