Inexperienced planners often underestimate the cost of safety instrumented system or SIS integration and updates. Anything that touches an SIS will cost more because of the specialized design devices and instrumentation required. Usually, any major SIS work will be handled as its own project, which is a much better approach. If we must go back and move or add things to provide additional safety interlocks, it gets expensive, really quickly.
In addition, many facilities have systems with a grandfathered status. Understanding how your migration will impact your facility’s grandfathered status or grandfathered systems is critical to your safety planning. Grandfathered status can be lost the moment a system is modified – that means you might need an SIS. In fact, it’s very important to understand that SIS control is much more expensive than DCS contro
Batch processing systems tend to be significantly more expensive to upgrade than continuous processes, and oftentimes companies underestimate the impact. While it may seem counterintuitive, batch processes are harder to handle than continuous, and migration costs reflect this reality. If you estimate a project based on I/O counts rather than the effort it will take to implement sequential control correctly, your estimate will be inaccurate. Unfortunately, this isn’t widely understood, and planners often apply the wrong cost model.
Another common mistake you should avoid is the trap of – it should be just like Unit 1 but different – when defining requirements. This can lead you down a path that will cause a great deal of problems. You really need to reverse engineer the functionality in the first unit, make sure you understand it and what it will take to put that in for the second unit. It is easy to skip the estimation step in this case or to make assumptions. If you don’t have the functionality documented, you really can’t develop a reliable estimate for that. Remember, good upfront definition saves you time from costly changes later.
The other pitfalls many face is regarding historian information, reporting and passing information to the manufacturing execution system or MES. That is often specialized work. If the skill-level requirements are not understood well, the estimate will not be accurate.
Too often, people don’t take adequate time to identify and plan all the interface requirements. Don’t forget vendor-supplied skids, your laboratory system, etc.
And if you’re implementing a phased migration, do you have a plan for how your new and your legacy systems are going to communicate to each other during each phase
Other things to consider include:
– Do you have cable tray space for the wiring associated with the new process you’re including in the migration?
– After years of gradual physical changes in your process area, is there still a viable path into your rack rooms or any enclosures?
– What about environmental conditions in the location of any new hardware – do you need any investment? What about the power for those areas as well?
– A construction plan is more than just: Do we have the space for new panels? Maybe we’re going to reuse existing wiring, and we need to evaluate usability of that.
So, going hand in hand with constructability is cutover planning. Don’t leave cutover planning to the end. You need to initialize it during planning and budgeting. We’ll cover this more in the next course but think about whether you can take advantage of major process outages or turnarounds. Is the project schedule required to make the outage work feasible with the capital allotment you’re likely to get? After all, expedited project schedules cost more. Maybe you need to plan to take advantage of a series of mini outages, but you can’t plan mini outages without knowing what process elements must migrate with the other elements.
Even more damaging problems are associated with a lack of vision into the future, and these are much harder to fix. Far too many companies do not adequately explore the capabilities a new system can bring. They look at the system presently in place and decide all they want or all they currently can do is implement the same thing on an existing platform. Replicating the old system is an enormous lost opportunity for improvement (and justification!), and it preserves all the problems and shortcomings of the legacy system. Companies rationalize this thinking by deciding it isn’t a good idea to add new things because it will annoy operators and require training. Rather, you should be open to exploring ways in which the facility can work better and take advantage of newer technologies available with new control system platforms. One of the worst mistakes we see is when a company insists that the new system must use existing HMI graphics. Not only is this a poor idea in general, it often costs more to force a current platform to look like an old one than to create new, optimized graphical interfaces.