Sounding out the health of your plant has in the past been akin to visiting your own doctor. Its something that absolutely needs to be done when suspicious symptoms start to show, but more normally is reserved for routine visits when regular check-ups are due.
These days, however, your plant comes ready equipped with its own team of doctors in the shape of the many diagnostic and monitoring systems incorporated into instrumentation and other process equipment. And, like any good doctor, todays plant diagnostics can look beyond the symptoms of poorly performing instruments and provide insights into the health of the process itself.
Diagnostic and monitoring systems have been available for many years, of course, but generally were applied in something of a piecemeal way, protecting process-critical pieces of plant and equipment rather than offering a systemic view of the process. However, the widespread adoption of digital automation and fieldbus networks now has opened the door to much more powerful and advanced levels of diagnostics.
There is a lot of activity in diagnostics in general, and certain areas in particular, notes Eddie Bridges, international product group manager with instrument maker Krohne, Peabody, Mass. Mounting pressure from end users certainly is spurring developments, he adds. NAMUR [the international association of users of process control technology in the chemical and pharmaceutical industries] has been pushing manufacturers into providing more information from the electronics in flowmeters, for example.
Plants want more diagnostics, agrees Paul Schmeling, senior marketing manager for Rosemount Pressure Products, Chanhassen, Minn., a part of Emerson Process Management. This has been a popular message with customers, who are trying to do a lot more with their instrumentation. We have a lot of engineers working on this right now, both at the device level as well as the wider system loop level. One result is its ASP Diagnostics Suite.
Making sound use of noise
Now embedded in the Rosemount 3051S series of pressure transmitters (Figure 1), ASP uses a basic principle known as Statistical Process Monitoring (SPM) and relies on the fact that virtually all dynamic processes have distinctive noise characteristics and that any significant changes in that noise can indicate that something is happening in the process or equipment (Figure 2).
The problem, however, has been in detecting and monitoring those changes. Typical process control systems generally filter out high frequency noise to improve the signal-to-noise ratio and only can read transmitter outputs every second or so. This is fast enough for most process control but its considerably slower than the speed at which many instrument sensors are actually receiving signals from the process. The 3051S, for instance, samples its process variable 22 times per second, says Schmeling. The high speed of the device together with the internal calculation of statistical parameters are key to the SPM technology.
The patented software resides on a diagnostics feature board mounted in the transmitter head and computes the mean and standard deviation of the input signals. SPM features a learning module that establishes baseline values that are considered normal for the process and installation. These values are constantly compared to the current values of the mean and standard deviation and then, based on sensitivity and user-selected settings, the SPM diagnostic can generate alarms, alerts or other action when a significant change is detected in either value.
For instance, if impulse lines to the pressure transmitter are plugged, the transmitter output may appear normal (because its effectively continuing to read the pressure at the time of the plug) and the mean will remain unchanged. However, Emersons research has shown that in a typical flow application the standard deviation at a given flow rate changes significantly if one or both impulse lines are plugged. The SPM diagnostics will pick this up, alerting the operator to the abnormal situation.
Figure 2. The effect of changing noise levels on the mean and standard deviation values for a process signal can help identify problems.
All the diagnostic information generated from the SPM-equipped 3051S now can be viewed through an enhanced EDDL-based graphical user interface much better than just showing numbers in a box, notes Schmeling.
ABB, Houston, also is keenly interested in improving the diagnostics of its instrumentation. The company is in continuous discussion with customers and end user groups to determine new sensor technology requirements for each product line, says Sean Keeping, vice president technology for ABB Instrumentation, St. Neots, U.K. The focus is on transforming the sensor data into value added information.
As one example, he cites the FSM4000 electromagnetic flowmeter, which employs a diagnostic coil to actively adjust the zero and eliminate any line noise. It effectively measures 100% of the useful signal and is said to deliver a higher signal-to-noise ratio than other systems. Keeping says the company also has patented a concept for in-situ validation of magmeters. The CalMaster 2 and CheckMaster products connect to the flowmeters and run a series of tests to verify their accuracy and repeatability against the original factory calibrations. The validation is based on advanced fuzzy logic analysis of the measurement data from the sensors.