The EngineerLive website has an article, System migration: make sure you improve control, not just replicate, which shares many of the thoughts we’ve expressed in recent modernization posts.
I turned once again to Emerson principal modernization consultant, John Dolenc, for his thoughts and additions to this article. I’ll include them in their entirety:
You can summarize the author’s comments as:
Process control systems will get old. The hardware components will eventually fail. Older configurations tools limit your ability to optimize the process. Integration to MES/ERP systems may be limited by older systems. Engineering expertise on these older systems has retired or moved on to new state-of-the-art systems. Spare parts are becoming scarce and expensive.
These are correct statements. So the bottom line is that at some time, older legacy systems will need to be replaced. The questions become: When will these system components fail? What are the options to replace the system? How much will it cost? How can the cost to replace the system be justified?
The author also had some excellent comments on the planning and implementation of a system migration project. He infers that migration projects are not easily and quickly implemented. The planning process is very important, and during that planning process it is import to identify the business and process issues that drive the plant. The migration plan should then be designed to meet the business and operation goals and not to simply replace-in-kind what exists.
Emerson Process Management provides automation feasibility studies to answer these questions. An automation study is always tailored to the specific needs of each client. Typically, the deliverables of an automation study include identification of benefits from automation and estimates of the financial value. An automation modernization / migration plan is developed to address business and operational issues; many times with migration options. A fairly detailed design is completed including bill of materials, scope of work for engineering and installation, cost estimates and schedules. The final report provides the information the client needs to make the decision to move to the front end engineering or implementation phase of the project.
The author also mentions that a holistic approach should be taken for the migration planning process. Although we cannot be sure what the author means by a “holistic approach,” we do have our own interpretation.
First; one must expand the scope of a “process control system” to really be a “process automation system” including instrumentation, control valves and other associated applications such as advanced control, safety systems, and asset management. Then one must understand that the solution to a control problem is not necessarily found in the process control system. Is the measurement device the correct choice for the application and process conditions? Is it located in the proper position for optimal control? Is the control valve the correct type and sized properly? Is the positioner up to the task of providing rock steady control? In other words, a control system can not correct all process control issues alone. All components that make up the control loop need to be considered.
Experiences from one of the early system migration projects we completed provide some examples. We were asked to migrate an existing control system in a batch chemical unit. The existing system was an integrated system consisting of PLC I/O equipment, PC-based control, and a custom, PC-based operator interface. As with most integrated systems we encounter, this system had the typical symptoms: Limited flexibility, the system integrator had to make all control modifications and provide maintenance, high cost for integrator services, and limited spare part availability just a few years after being installed. The only migration option was complete replacement.
During the early design phase of the project we discovered that many of the existing instrumentation was in bad repair, weigh cells were used to add the raw materials, and the right-the-first-time factor was in the mid-80s range. We convinced the client that replacement of the bad instrumentation was in order along with the inclusion of mass flow meters and control valves for the raw material feeds to the reactors. The client required that we reverse engineer the system configuration, because they had invested quite a bit into system modifications and they did not want to lose the engineering investment. So in essence, the control logic was close to a replace-in-kind situation.
The right-the-first-time factor was in the low nineties after the first few weeks of operation upon completion of the migration. The only issue preventing this factor being in the high nineties was certain control logic that was added to the legacy system to overcome its control deficiencies. These logic modifications were soon removed.
The lessons learned here are two fold. First, the new control system performance was impeded by retaining the existing control logic in kind versus retaining the desired control functions and allowing the control system to be configured as designed. Second, the addition of better instrumentation and control valves substantially improved the performance of the process, despite the control logic not being optimal.
John, thanks for adding your perspectives to this article!