Page 305 - Design of Simple and Robust Process Plants

8.2 Instrumentation 291
were higher than the off-line analysis. The disadvantages of the analyzers were still
the sample taking and handling, and the related long response times, and this forced
development effort in the direction of in-situ (or in bypass loops) installed optical
measurements including UV, IR, near IR (NIR) and Fourier transformed IR (FTIR).
A good example of this technique is flue gas analysis with IR. Sample taking was
the main handicap of this application, the sample needing to be sucked from the
stack. Condensation in the sample lines was problematic, and the response times
were long. In-situ measurement overcame these handicaps. Only a few of the avail-
able techniques are mentioned, but the trend is towards greater accuracy and reliabil-
ity, and shorter response times. Thus, these units can be included in a control loop.

Although the development of the most common measurement techniques have
been discussed in the previous section, many more measurements have been sub-
ject to the same development cycle. The improvements made have addressed the
major failures of instruments as detailed in Table 8.1. The improvements in the
reliability of instruments are important for robust operation, as nuisance trips are
often caused by instruments and also by ªover-instrumentationº. We mentioned ear-
lier the objective of a two-engine airplane flying over the ocean at the same level of
reliability as a four-engine counterpart. The same approach is applicable to instru-
mentation, and therefore process plants will benefit from more reliable, but fewer,
instruments. Some data were collected for nuisance trips concerned with compres-
sor trains. Four multi-stage centrifugal compressor trains with drivers of around
20 MW were monitored for 4 years during the late 1990s, the number of trips being
totaled and categorized:
. Instrumentation
. Process
. Mechanical
. External cause
The results are shown in Table 8.2. Although the number of data are limited, it may
still be concluded that the number of trips due to instrument failure are high in this
data set (50%). The mechanical failure is a clearly identified other cause. The other
ªtripsº were registered as process and externally caused trips, though the data did
not indicate the original cause of these interruptions. These data illustrate the
importance of the reliable instruments that are superseding mechanical failures,
though it should be noted that instrumental trips (in general) result in short process
stops. Mechanical failures might result in considerably longer outages. The failure
rate data for the improved measurements as discussed above require collection, as
these are needed for instrumental design.

Table 8.2. Causes of nuisance trips of four major multi-stage compressor trains (including the
drivers) over a 4-year period.

Instrumentation Process Mechanical Outside cause
Number of trips 10 5 1 4