This tip seems obvious, yet some of the biggest project failures I have witnessed were a result of the team not understanding the process, and programming what they thought was correct. The folly of this approach was usually not discovered until late in the project (possibly during start-up). At that point, recovery of the schedule was impossible, and the effort required to correct the problem resulted in massive budget overruns.
Concept: Designing and/or improving a control system is practically impossible if the team does not understand the process. The first step of ANY control project should be a study of the P&IDs and a conversation with the process gurus of each affected area.
Details: Before beginning any major project, pull out the P&IDs, batch sheets (or other process documentation), and any operator instructions that can be found and read through them. Then take the time to talk through the process with the plant engineers and operations staff. Be sure to inquire about the normal process flow and any non-routine cleanout, emergency, or abnormal situations that may occur. If the process documentation is sparse or not updated, take the time to interview plant personnel and document the process and the functionality of the current control system. This thorough understanding of the process is absolutely critical to project success. Armed with this knowledge you can not only design the system correctly, but offer ideas for improvements that the plant might not have considered.
The project team’s understanding is as crucial as the project leader’s. If you are the project leader, distribute copies of the P&IDs to the entire team, and walk them through the process. Many project teams find it useful to map out the process on a large wipe board in the project team area during the discussions. This “process roadmap” can be left up for the entire project so team members can easily refer to it and discuss control options and ideas during the system design.
Watch Out: Always ask about “abnormal” scenarios. Clean outs, shutdown/startup sequences, and other non-routine events may require more programming effort than the original project itself. They can also be the most difficult to program due to various interactions and undocumented operations.
Exceptions: There really are none. Even if the job is a control system retrofit with no significant software changes, take the time to learn the process.
Insight: In many systems, you will find programming errors that have been there for years. When asked the operators will often say, “Yeah, it has always done that, and we always wondered why.” If the team encounters something that just looks wrong, ASK! Either the team does not understand the process nearly as well as they think they door the original program was in error. Either way you are well ahead by raising the question.
Rule of Thumb: Process understanding is a crucial first step in any automation project. A failure to achieve that understanding can place the entire project in jeopardy.
Look for another tip next Friday.