HOW PRECOG analyses periodic motion waveforms

Telemetry data process

It is important to read this article once there is familiarity with how PRECOG can be used for continuous-operation machines.
Only one additional step needs to be added when using the same approach: PRECOG must be calculated for each moment in time along the curve.

For this, select a single moment, and gather all the values at that moment across each curve, as shown in the figure below.
Once the calculations for each moment are completed, they will provide a clear picture of which segments of the curve show anomalies.

Since the curves of continuous-operation machines often contain numerous values, and periodic motion curves are not measured as frequently, the nature of the data may result in multiple anomalies being detected. Counting these anomalies can provide a good representation of the machine’s general health.

In our experience, it is initially worth setting the target values for each moment to the maximum values recorded during proper operation under normal operating conditions, with the tolerance based on the standard deviation of the data. Naturally, the manufacturer’s specifications, such as maximum power consumption, must also be taken into account. After this, various test cases can be conducted to further refine the target and tolerance values to achieve the desired results.

It is important to note that the accuracy, sensitivity, and sampling frequency of the measuring instrument can significantly influence the types of errors that can be detected in a device. Therefore, it is essential to carefully review the technical specifications of the measuring instrument and perform tests on a sample device before deploying it for widespread use.

For example, in the case of 3-phase AC motors, an important consideration is the type of low-pass filter used by the instrument, as it determines how much the measured power curve is smoothed.

PRECOG uses “recipies” to devide data sets into sections

For more sophisticated analysis and to differentiate between various potential errors, the curve should be divided into smaller sections based on the test results.
By analyzing the data from these tests, we can identify the specific time periods that require monitoring.
Additionally, a “recipe” can be developed to allow for quick and easy evaluation using the data provided by PRECOG.
The recipe can be fine-tuned as needed based on experience and, in most cases, can be applied immediately to multiple units of the same type of equipment.

The recipe

It’s sufficient to use just the PRECOG approach for all faults

The recipe