Page 43 - Root Cause Failure Analysis

34 Root Cause Failure Analysis

It is necessary to develop a test sequence that isolates and directly compares each sus-
pect variable. For example, a series of vibration tests can be conducted to measure the
actual change in the machine or process system’s response to discrete changes in
incoming-product parameters, such as viscosity or temperature. Another sequence of
tests can be conducted to measure the effects of variations in operating practices or
ambient environment.
All vibration testing should be based on a well-thought-out test plan that defines its
purpose and the methods to be used. Good, detailed documentation is essential. Each
test in the sequence must be fully detailed to clearly define the exact parameters to use
during data acquisition. This documentation is essential for proper analysis and com-
parison of the data.

In most cases, process or ambient environment data must be acquired along with the
vibration data. Since these tests are designed to quantify the effects of changes in the
variable, it is necessary to verify the actual change caused by each.

Real-time data can be used to accurately quantify the cause-and-effect relationship
between suspected variables and the operating dynamics of the machine or system.
Multichannel data-acquisition techniques provide the best means of acquiring this
type of data. Multichannel techniques record all variables at the same time, elimi-
nating any transients that may be missed with single-channel data acquisition
methods.

Process Parameters
Evaluating variations in process parameters, such as pressures flow rate, and tem-
perature, is an effective means of confirming their impact on the production system.
In some cases, the Level I computer monitoring systems used to operate the process
system provide much of the data needed. The monitoring systems normally
acquire, report, and store key process variables throughout the normal production
process. This information is useful for both the initial analysis and confirmation
testing.

Tests should be developed that measure the impact of process-parameter changes on
the system to determine if specific deviations contributed to the event being investi-
gated. In most cases, these tests should be used in conjunction with vibration analysis
to measure the cause-and-effect relationship between specific process setups and the
operating dynamics of the system.

Visual Inspection

Visual inspection is one tool that always should be used. Such inspections should be
performed on measurement devices, such as pressure gauges, and failed machine
components.