As anyone who has been involved in a control system migration will tell you, it’s never an easy process. Whether it’s an upgrade, expansion, stepwise migration or rip-and-replace, the bigger and more complex the project, the more fraught with tension and risk. To help you get through the project with your sanity intact, Automation World readers share their recommendations and suggest pitfalls to avoid:
1. Determine strategy. Your migration strategy will depend on which type of automation you’re dealing with: scripts, workflow tools, policy-based orchestration, configuration or control systems. The different activities that can be automated (provisioning, maintenance, proactive incident response, production execution, etc.) and the different degrees of automation (automating just a few actions, partial workflows or end-to-end) will determine your strategy in terms of resources, time scale, production stops, etc.
2. Virtualize first. Automation upgrades or migrations need to be scheduled properly in terms of system commission date to extend the warranty or for a vendor’s obsolete notice date. The best practice is to conduct a virtualization of the new automation system. The future of automation will need virtualized infrastructure and platforms to deal with the IT spectrum, cyber security and better management capabilities. Virtualization has many benefits in terms of technology, investment, maintenance and lifecycle cost.
3. Take it one step at a time. Avoid changing the entire system or manufacturer if you are upgrading. Upgrading to the newer modules or systems of the same vendor provides a bit more reliability, since the basic architecture remains the same.
4. Don’t experiment. While innovation is important, there is a counter-argument for doing what you know will work. If rip-and-replace is possible (and that means you have to stop the plant for several days, weeks, or months depending on the circumstances) and you know that it works, that is the best choice. But if you can’t afford a shutdown, then go for a step-by-step migration. Make sure you work with an experienced vendor and proven technology.
5. Consider three critical migration issues. When doing a migration there are three points to think about: how to update software and whether you have the right conversion tools; what you need to do to avoid system failure or risk for the migration step; what is the expected lifecycle of the new system.
6. Make no assumptions. Try to foresee every small step in a migration implementation. Don’t assume anything. Every implementation is done to achieve some objective of the operation. The needs could range from some reporting or alarm functions to an action initiated due to alarm. Always visit the site to understand the requirements and the nuances completely.
7. Changing suppliers adds some complexity. The difficulty of a process migration usually increases when you change DCS suppliers, since different brands often don’t have similar functions. Factor that into your timeline and risk assessment when weighing whether to switch vendors.
8. Start with data needs. First you need to understand what data the user will require and how quickly the data is needed. That should be the starting point in developing your migration strategy. The second priority is to determine the impact on the safety and productivity of the plant.
9. Focus on controllers.The best strategy is to first upgrade the controllers, then replace the I/O chassis piece-by-piece going forward. Some I/O changes could be driven by other projects, such as a motor control center(MCC) replacement.
10. Do your homework. Do some up-front analysis to avoid creating problems for yourself by not choosing the right migration path. For example, migrating from one generation of processor to another one may not be a wise choice. Reviewing the instruction sets and information available about conversions and manufacturer recommendations will give you insight into the difficulty of the conversion. If you do your homework, you might choose a different processor to make the conversion easier.
11. Technology education. It is important to educate everyone on the new technology. Remember, it is easy to use "old" thinking instead of changing practices to take advantage of the benefits of the new technology.
12. Decentralize. The architecture has to be critically reviewed and transformed, keeping in view the improved performance of the local controllers. Your mantra should be to decentralize the controls as far as possible.
13. Aging equipment. Depending on the technology you have installed, when your equipment is more than 10 years old you will need to implement a rip-and-replace. If you are just making some modifications you can upgrade or make an expansion only. Most of the problems that arise during a migration are with the field equipment you have installed and control room facilities.
Liked this article? Download the entire playbook here.
