The following discussion is part of an occasional series, "Ask the Automation Pros," authored by Greg McMillan, industry consultant, author of numerous process control books, and 2010 ISA Life Achievement Award recipient. Program administrators will collect submitted questions and solicits responses from automation professionals. Past Q&A videos are available on the ISA YouTube channel. View the playlist here. You can read all posts from this series here.
Looking for additional career guidance, or to offer support to those new to automation? Sign up for the ISA Mentor Program.
See Part 1 here. See Part 3 here. See Part 4 here. See Part 5 here.
What are the biggest mistakes you have seen in automation system design, configuration, calibration, installation, checkout, commissioning, and maintenance? What were the consequences and the fixes and what can be done to prevent future occurrences?
As a maintenance and reliability professional that spends a significant amount of time reviewing scope development through detailed engineering, I think sometimes we focus too much on the bells and whistles associated with advanced control and diagnostics while missing fundamentals. We need to shift the “high value engineering” mindset where we just look at the installation costs and compare it against the total cost of ownership (TCO). I understand there are competing priorities as the project teams are constantly being indexed against cost and schedule, but how much is saved when the maintenance costs can be increased exponentially?
Consider the following example: A differential pressure transmitter with process diaphragm is being used to measure level in a tank with a 120,000lb storage capacity of an extremely caustic material. Installing a double block and bleed allows for in-situ maintenance, but direct mounting saves the project money. This $1,500 installation savings now requires a little more pre-work for maintenance/calibration:
Preventive maintenance (PM) to verify the calibration now requires:
How much did that $1,500 installation savings really save? While this seems to be an extreme example, it happens every day in facilities that may not yet have site standards promoting good engineering practices and reliance on minimum code requirements. We need to look past the installation and minimum code requirements to promote good engineering practices that deliver a lower TCO for the client.
Direct cost savings:
Indirect cost savings (and personnel safety):
In conclusion, I suggest making an increased effort to design for maintenance. It can be the best “high value” effort we offer to our customers by reducing revolving costs over the equipment or system life cycle while increasing efficiency and safety. Isn’t that the end goal for automation?
Here are a few:
After mechanical completion, enough time must be given to execute a point-to-point checkout. This is to make sure all instrumentation, sensors, and final control elements are properly wired from the field to the control system. As the startup date is getting close, each discipline gets into a fast-track mode that makes it easy for the people to fall in mistakes. Most common are loose bolts or loosening of wires at the terminal blocks, connections at the wrong channel, and/or wires at the wrong terminal of the input/output module.
A very good practice after checkout is to perform what is sometimes called “water batching.” This, in essence, refers to testing a given piece of the process which will be run for some time at a low rate or for some batches to determine if all performs as intended. This helps the team to identify early issues. I have seen how water batching can reveal design and construction mistakes:
The omission of factory acceptance testing of implemented control strategies based on functional requirements can add to commissioning delays and in some cases production loss and impact on safety integrity. Essential to the execution of the factory acceptance test is a detailed written procedure with expected behavior described. The control narrative and functional logic diagrams should be used as reference documents for the tests and form and, together with the procedure and recorded results, the record of compliance with the design intent.
In the implementation (configuration) of complex control strategies, even with the best intent, errors can occur in the selection of configurable options at the function block level that affect behavior. Issues with controller initialization, for example in ganged or split range control and timing issues in Boolean logic, if not discovered during testing can be a challenge to resolve during commissioning. Functional testing also provides the opportunity to verify the design of the operator interface, and with the participation of a control room operator, can provide the opportunity for refinement and early acceptance of the operator interface.
For complex loops, a simple process model can provide the feedback to the controllers to reduce the time required to carry out functional testing. Since in contemporary digital control systems the suite of functions provided for process control can be readily used to build simplified process models, feedback can be added with surprisingly little effort. Some DCS vendors may go further in providing virtual input channels which minimizes any temporary adjustment of the configuration to allow the use of simulated process measurements.
To sum up, a well-executed and documented factory acceptance test, in addition to the aforementioned benefits, will be appreciated by any who come under the pressure that typically accompanies the commissioning, on-line testing, and tuning of the control system.
One of the largest mistakes I’ve seen stems from a lack of awareness of the ASME B16.5 material rating tables and the subsequent impact on the piping/instrument interface flange. I have heard several stories about significant re-work being required when an instrument has been bought with a flange that isn’t compatible with its piping isolation valve. In the instance below, I’ve shown an example where a 600# piping flange has been specified against a 900# instrument.
Within B16.5 common metals used in the industry are arranged into formal groups with tables specifying working pressures of that particular group by pipe class (150#,300#,600#, etc.). Each table extends to 538°C (approx. 1,000°F) and from the various tables, it's evident that the working pressure of all groups derate as the temperature is increased. An important point to note is that the factor by which a group’s working pressure derates differs across the material groups and therein lies the potential for error.
A common material of construction for main piping is carbon steel (A105) which is a group 1.1 material. I’ve shown the PT derating curve for group 1.1 below, with a fictional P&T design of 65 bar(g) & 300°C. A piping engineer will provide a 600# flange as standard in this service, in-line with the pipe class.
The most common material for all things instrumentation is stainless steel (316SS), which is a group 2.2 material. The derating curve is shown below for group 2.2, at the same P&T design as above.
In essence, this design mistake is a communications issue, but given the ramifications to rectify the installation at site, it can be one of the most problematic. Regardless of the potential for re-work and re-design, the derating of materials at elevated temperatures is something that all instrument engineers should be aware of as it can present significant safety concerns at elevated temperatures. This problem can be overcome early in the design stages of a project, but like the vast majority of engineering mistakes, poor communication and “siloed” working lies at the root cause of the problem.
To my limited experience, I have seen quite frequently these two errors:
1. Lack of operator friendliness for the setting of reverse/direct at control loop components when involving a failed open (FO) (air to close) control valve.
Though the settings can be correct (achieving the final negative feedback control loop), it is not necessarily intuitive. For an intuitive display to the operator in control room, the controller output (CO) signal command to the control valve should align with valve position—0% and 100% CO represent a shut and fully open valve position, respectively.
For control loop with FO valve, we usually see the opposite, whereby the controller is configured in such a way that it’s CO indication on human-machine interface (HMI)/DCS graphics is not aligned with valve position. For example, 20% CO means commanding the valve to be 80% open. In the event that intervention by an operator is required, when the valve is demanded to be fully closed, the operator may fully open the valve if he accidentally keys in a CO of 0%.
There are many great guidelines written on how to correctly configure the controller actions, such as Appendix C Checklist for PID Controller Features from 101 Tips for a Successful Automation Career, Table 1 of Good Tuning: A Pocket Guide Fourth Edition, and others (I would also use simple guides from Harold Wade). For a negative feedback control loop to work, there must be an odd number of sign reversals around the loop. Let’s use flow control loop using a FO (air to close) valve as an example.
The flow process is naturally direct. To achieve negative feedback control, the loop may be configured as per Figure 2. As we have one sign reversal (odd number), the loop will work. With only the valve element reversed (increase-close), the rest of the loop components can be set as direct. This setup, however, is not so intuitive for the operators.
For an operator-friendly approach, the loop can be configured with additional sign reversal in the controller and AO block. Finally, we have three sign reversals which still works as we have an odd number. Better yet, our CO now aligns with valve position (see Figure 3). Note that instead of an AO block, reversal setting can also be done at the positioner. However, this is not advisable as this setting could potentially be lost when someone else replaces/upgrades the positioner later.
2. Wrong configuration of override control scheme
Instead of using normal high/low selectors on the output of controllers, some DCS configurators misunderstand by configuring unnecessary conditional logic which doesn’t work. Figure 4 is the wrong implementation that I have seen for override control. Most of the time, the selector switch is being forced manually by the operator at one selected position only.
This mistake could be due to a lack of understanding on how override control works. Figure 5 is showing the correct implementation.