Blog

The Value of PID Tuning, CLPM, and System Identification Software

PID tuning optimization, monitoring, and advanced process control are increasingly important today.  Many plants do not use software and tools in these areas because the software is often too complex, expensive, not robust or not practical enough in the control room environment.  Colleges continue to teach process control theory in the ancient Laplace, frequency domain or the complicated Z (discrete) domain- all not practical and not effective in the control room environment.  Subsequently, few plants tune PID loops using modern effective methods.  Even fewer plants monitor the control quality of PID loops in online or offline manner.  Many PID loops are not optimally tuned.  They may work “satisfactorily”, but with tuning optimization, the performance could be much better, but there is lack of awareness of the potential and the resulting benefits.

Another related area of high value and benefits to chemical plants is system identification.  System identification involves identifying the dynamic response between manipulated variables and controlled variables.  System identification using transfer functions is a practical and effective method.  Unfortunately, almost all system identification- transfer function identification is done using the age-old ARMAX method –  autoregressive moving average with exogenous inputs.  ARMAX method is too complex for use in the control room by busy engineers and technicians.  Subsequently, system identification (transfer function identification) skills and usage is poor in chemical plants.  A few Ph.D. control engineers in core engineering or in research departments may use ARMAX but it is not widely used otherwise.

For successful and effective process control in industry, two areas are critical and essential – first is system identification and second is controller design and tuning. Accurate system identification is a prerequisite for good controller design.  Due to the challenges and limitations on the practical success of system identification, dynamic models are rarely available and this precludes the mathematically scientific and precise design of the closed-loop controller.  Often the age-old trial-and-error method of tuning or determination of other APC (advanced process control) parameters is used with little chance of correctness and good control quality performance.  Guess work on PID controller loop tuning and APC parameters results in poor control quality, often oscillations or sluggish control followed by the control room operator turning off the advanced process control scheme or putting the PID control loop in manual mode.  A plant then could run in this mode with many PID loops and APC schemes in manual/inactive state for years and even decades.

In recent times, many experienced process control engineers have retired.  Many new, young and inexperienced process control engineers, DCS technicians, PLC technicians and control room operators have entered chemical plant control rooms.  Lack of modern, effective and practical process control tools has resulted in many missed opportunities in maximizing plant performance and plant profits.  Production rates are maximized effectively, waste and utility consumption are not minimized.  Avoidable mistakes and plant shutdowns due to poor process control impacts plant performance metrics.

PiControl Solutions LLC is a global leader in the development and sales of modern software for PID tuning, system identification, APC (advanced process control) design/implementation and process control training for process control engineers, DCS/PLC technicians, operators and managers.  PiControl has four key software tools that address the current needs of the industry and trains on deficits left by academic colleges.  The process control software algorithms are new, novel and modern and work better than ARMAX method and better than currently known and used PID tuning methods.  The same software used in control rooms for making PID control improvements and APC implementation can be used for training plant personnel.  Furthermore, the same tools can be used in academic colleges for teaching semester based process control courses at both undergraduate and graduate process control curriculum and also for process control academic research.

The new PiControl software offers the following four main functions:

1. PITOPS Closed-loop system identification (closed loop transfer function identification) with PID control loops in automatic or even cascade modes without any step changes.  This is a great improvement over the current tools that require open-loop step tests on the OP (output) with PIDs in manual or step tests on the SP (setpoint) in auto mode.  PiControl algorithm can identify open-loop transfer functions with completely closed-loop data with controllers in cascade mode. This is a modern-day breakthrough invention and a new capability previously considered difficult if not impossible.
2. The closed-loop system identification software PITOPS described above is multivariable – so multiple open-loop transfer functions can be simultaneously identified with completely closed-loop data with controllers in cascade mode.  Most existing tools that advertise closed-loop transfer function identification are only SISO (single input single output) and even then the final results show significant uncertainty.  PiControl technology is MIMO (multiple input multiple output) and completely closed-loop without the need for step tests.
3. Many of the PID tuning criteria used currently like IMC (internal model control), Ziegler Nichols, Cohen-Coon, Lambda tuning and others are based on an equation for calculating the proportional gain, integral tuning parameter, and derivative tuning parameter.  In contrast, PITOPS from PiControl technology allows custom configuration of the PID control loop with conditions as seen in the real DCS followed by PID tuning optimization.  In this method, a typical setpoint change is configured in the actual engineering units seen in the DCS (not percent units) plus typical noise is configured and also ramp, pulse and sinusoidal disturbances mimicking the real unmeasured disturbances and PID loop interactions and potential fighting with neighboring PID loops.  PiControl’s PID tuning optimizer then minimizes the error for this custom control configuration comprising the SP change and the various disturbances.  Furthermore, the optimizer allows imposing a ROC (rate of change) on the control valve movement or the slave PID controller’s setpoint movement to avoid jerking and disturbing the downstream unit, equipment, and process.  These functions are novel, new and unmatched by known technology, algorithms, and competitor software.
4. PiControl has developed new algorithm and software called APROMON (advanced process monitoring) for quickly and easily monitoring the PID control quality of all the PID controller loops in any factory or manufacturing plant.  The monitoring tool can be installed in just half a day using direct OPC connectivity to any DCS or PLC equipped with an OPC server.  PiControl’s new algorithm – TAD (true amplitude detection) is a major breakthrough in its ability to reliably detect oscillations, instability, and hunting without getting fooled by the noise and squiggles that are present in process data.  The monitoring tool reports over 25 statistical and process control quality criteria including integrated error, standard deviation, variance, control valve diagnostics, stiction, hysteresis, PV signal problems, excessive control valve movement, low onstream factor, PID control loop interaction, PP, PPK, and others.

All PiControl software works entirely in the time domain with no complex or advanced math and can be learned and installed in just a few minutes to no more than a few hours.  PiControl vision is to offer amazing new, novel and modern tools to help control room personnel and help bridge the gap between academic colleges and industry.  PiControl also offers a wide variety of practical industrial process control courses covering PID tuning, system identification, advanced process control (APC), model predictive control (MPC), plant safety, cyber security, OPC (OLE for process control), industrial statistics, DCS and PLC. If you have any further questions, contact us online or give us a call at 832-495-6436.