S. Mcbul, OQ Specialty Chemicals, Masqaţ, Oman
When addressing process controls, the immediate interest is to invest in advanced process control (APC) technology without knowing the performance and stability of complex wide primary (PID) control. An APC typically sits on top of the PID control—if the PID control is unstable, the APC cannot do much to resolve the instability because there are no known shortcuts in process control. There is a tendency in APC projects to tune controllers that are identified for APC control while ignoring the rest; however, the process interactions creep and affect those controls that APC needs directly or indirectly. The lesson learned is that there are no shortcuts.
Organizations tend to jump into APC mainly due to advertised benefits by vendors. However, the value of well-performing PID receives only partial attention mainly due to the difficulty in identifying the value of the PID tuning stabilization.
It has always been challenging to address the business value of improving PID controls because they only reduce the variability or standard deviation around the mean value. With improved PID control, the operation continues to operate at its previous average value (FIG. 1).
For example, if the average production rate were 100 throughputs per labor hour (tph) with a standard deviation of 3, a PID would be performed, producing 100 tph with a standard deviation of 0.5. The finance team will not confirm any improvement in the bottom line as the production rate remains the same.
In certain cases, PID stabilization directly improves the plant's bottom line or improves reliability; in various other cases, it enables moving closer to optimum production (FIG. 2). The three phases of process control are:
This case is an example from a propylene fractionator in a residue fluid catalytic cracking (RFCC) unit. The operator kept a low reboiling duty due to high reboiler steam variability, mainly because of an unstable valve coupled with poor tuning, leading to increased bottom impurity, propylene and propane loss. When the controller tuning was optimized, the variability was reduced, allowing the operator to increase the reboiling, reducing propylene loss. The reboiling duty increased, leading to increased bottom purity and improved recovery of higher value olefin from the column top product. FIG. 3 shows that the standard deviation was reduced, and the mean was moved closer to the column flooding limits.
FIG. 4 shows the amine top pressure control in oscillation, leading to higher than normal amine losses. Once stabilized, it leads to a significant reduction in amine losses. This is an example of how the tuning could sometimes lead to reduced losses.
On another occasion, instability in the hydrogen (H2) header led to higher-than-normal flaring, and the tuning optimization of pressure control led to operating conditions below flaring limits and increased plant profits by conserving H2 produced from the steam methane reformer (FIG. 5).
FIG. 6 is an example of loss to the plant’s bottom line due to lower PID control utilization. In this case, the air-to-fuel ratio control of the furnace remained in manual mode. When taken in auto, it adjusted the air-to-fuel ratio so the same energy was delivered to the process with lower fuel consumption.
Takeaway. Dozens of such examples occur daily; however, FIG. 6 illustrates the concept. It is recommended to continuously monitor the utilization and stability of all controllers, identify if the issues remain in the controller, valve or instrument, and resolve them as soon as possible to improve the business impact.
An average refinery can have more than 1,000 PID loops, while an integrated petrochemical complex can exceed 3,000 loops. It is hard to monitor each of them, so it is common practice to use a PID monitoring tool to automatically identify and rank the best- to worst-performing PID loops and tune the poor-performing loops, valves or transmitters daily. Once the process is stabilized, there will be loops affected by multivariable interactions (causing oscillatory behavior) that will be the target of the APC. Tuning alone may not completely resolve multivariable interactions without detuning, which may again not be preferable.
The well-tuned plant will ensure greater stability and reliability of process equipment and valves, allowing direct and indirect benefits by shifting the mean closer to constraining limits. Once this is done, the next step will be to use APC to stabilize multivariable interfacing loops and allow it to automatically shift to the optimum levels across the plant. HP
SAQIB MCBUL is the Head of process control and digitization at OQ Specialty Chemicals’ Oman refinery and petrochemicals complex. He has served in various plants helping owner-operators in the hydrocarbon processing industry increase their bottom lines using regulatory and advanced process control, manufacturing execution and digital information systems. Mcbul is the founder of Apex Digitization.