Interested in linking to "Process Performance Improvement - Part 2"?
You may use the Headline, Deck, Byline and URL of this article on your Web site. To link to this article, select and copy the HTML code below and paste it on your own Web site.
By Greg McMillan and Stan Weiner
Greg McMillan and Stan Weiner bring their wits and more than 66 years of process control experience to bear on your questions, comments, and problems. Write to them at email@example.com.
Stan: We continue our interview with Mike Brown, who has focused on the bottom-line performance of process control tools for basic and advanced control systems in his role as Vice President, North America Solutions at Matrikon Inc.
Greg: Where are the greatest opportunities today to better leverage tools and technology in process control to help deliver true performance improvement?
Mike: The biggest weakness and scale is still at the regulatory level. Base-level control has really not improved much given all the changes and improvements on the systems side. We need to be more flexible about how we support process control requirements, and the tools also need this flexibility. Today, advanced process control (APC) tools in industries like air separation are often designed, implemented and commissioned completely with remote services and fully supported by a centralized model of non-APC experts. Without the flexibility of the tools to support this model, these advanced control benefits in the air separation business would not have been possible.
Stan: What are some specific examples of ways to get non-control people to use control technology to deliver process improvements?
Mike: Bring 24X7 operators into the tuning workflow. Recent developments in tuning tools, such as closed-loop testing and identification, tuning of closed loop cascades and tuning of interactive loops now make this possible. Tuning tools are also shifting away from complex math and process control talk towards simpler understanding for process engineers. Working with process response curves that show open-loop response time, closed-loop response time, stability and confidence for both old and proposed tuning can really change the paradigm.
Another example is to combine some of the simple PID metrics of a monitoring system, such as percent saturation of a control loop, and use some powerful visualization techniques, such as tree maps, to analyze this information across a plant. Giving the process engineer the ability to quickly visualize all the flow controllers in a plant that have high output saturation is a great way to understand hydraulic limits and the need for bypass management. Further, looking at the same information for all temperature controllers provides a good overview of heating and cooling constraints. This is information that can easily be presented to the process engineer on a plant-wide basis in less than five seconds and provides good process knowledge. How long would it take for most process engineers to pull this data?
Again, simplification and robustness of these tools are essential to make this model work. Technologies usually evolve by adding features to solve that last 5% of complexity that exists. In many cases, the tools become too complicated for the non-expert to apply, and the expert is too busy to get to the problem. Technology development needs to focus more on easily solving 80% of the cases with a simple and robust approach that can be implemented by non-control specialists, such as process engineers, operators and instrumentation technicians. Imagine the case where 80% of all regulator loops could be optimally maintained by the shift operator, or 80% of the advanced control applications could be implemented by process engineers who really understand the process economics. This will require sophisticated technologies, but the complexity needs to be directed at the robustness of the tool and hidden from the user.
Greg: Some of the changes that you are proposing are fundamental to how an organization does things. Why do you think they will recognize the need for these changes?
Mike: They really do not have a choice. The fact is that their organizations are changing, and plants do not have the same level of staffing or dedicated expertise that they carried twenty years ago. Management knows this, and they only see it getting worse as more expertise retires. Companies that twenty years ago had a process control team of six people now have a team of two people who are just able to keep the control system infrastructure running and updated to current releases. Twenty years ago, advanced control projects were often executed with a team of two consultants supported by two engineers from the operating company. Today, a single third-party consultant is expected to independently deliver the application. In many cases, the knowledge never gets effectively transferred to the operating company. If operating companies are going to re-establish ownership of their process control improvement initiatives, they need simpler technology that allows greater flexibility of the resources that can be assigned to these tasks.
Stan: What will be required for companies to support these efforts in the long term?
Mike: They need to see the economic results associated with these technologies and understand what is actually causing these economic effects.
There was a very good publication a few years ago by a U.S. chemical company trying to understand the real source of economics behind the recent application of a model-predictive controller (MPC) in an ethylene plant. The application demonstrated significant economic benefits mostly due to higher production rates, but there was a sense that over time, the benefits were not being sustained to the levels demonstrated post-commissioning. The immediate reaction was to bring back the consultant to review the design of the MPC controller, but one smart process engineer decided instead to take a Six Sigma approach to the analysis and to start documenting the magnitude and root cause of the benefits reduction each time the MPC controller was not able to push to maximum charge rates. After three months of analysis, a simple Pareto chart clearly showed that over 75% of the MPC controller benefits degradation was due to analyzer downtime, operational issues, operator training and instrumentation. There were no issues with the MPC controller design and its associated models. However, since the economics associated with these functions were now clearly linked to process control improvements and process benefits, there was now support for improvements in areas such as analyzer maintenance and operator training.
Taking the analysis one step further, the study extracted simple operator change data on MPC variables, such as constraint limits and manipulated variable limits. This data demonstrated that over time, as operators made more changes to the controller, the benefits dropped. Lack of operator understanding and intervention more often reduced the MPC controller's ability to truly optimize the process and maximize the production rates. Operations managers were now able to monitor the number of operator changes to MPC variable limits on a weekly basis and establish a threshold at which they would re-train the operator on the proper management of the controller. This is an example of operations people performing simple monitoring of control information in order to improve process control performance with a clear understanding of the economic benefits.
Greg: See if Mike's "Top Ten List" can help you get better traction within your organization for the adoption of process control technologies.
10) Explain that in a typical plant, 70% of manufacturing processes or production units are under some form of process control.
9) Tell them that most process control technologies involve using a joystick.
8) Tell them that process control improvements give instant gratification: Designing a plant may take years to see the results, but process control improvements are visible within days.
7) Tell them that on Fridays, control engineers get extra BBQ at the plant cafeteria.
6) Show them that advanced control technologies are the best way to understand real plant limits and bottlenecks.
5) Tell them that plant operators will be nicer to them.
4) The best control engineers are also very good process engineers … so they can do two jobs!
3) Tell them that control engineers get 50% of all their savings as a Christmas bonus.
2) Demonstrate how most plants can achieve up to a 5% production increase, 10% better energy efficiencies, or 15% better yield improvements with better process control.
1) Point them to the Control survey that shows how process control engineers have better sex lives.