Predictive maintenance

Predictive Maintenance

Striving to determine the likely time and amount of failures

Predictive maintenance (PdM) is a proactive effort that uses large data sets and advanced analytical methods for the purpose of identifying the time (when question) and the volume of failures (how many question). It uses innovative and complex advanced analytics and technologies, powered by Artificial Intelligence and Machine Learning. The advancements in collecting large amounts of data, and making those data sets available for predictive solutions, have significantly enhanced the application of PdM tools.

There are two categories of PdM, each with distinct applicability and value proposition: Definitive/ Deterministic and Stochastic/ Probabilistic.

Definitive/ Deterministic Predictive (dPredictive)

____________________________________

In essence, dPredictive assesses and identifies the most likely time of failure. Its value centered around the premise that knowing the time of failure as accurately as possible allows just-in-time performance of maintenance before breakdown taking place and creating operational interruptions.

The following are needed for dPredictive:

1.

Identification of all possible causes/ symptoms making the equipment inoperative.

2.

Condition-monitoring system to track the equipment operational/ health status.

3.

Systems for collecting and keeping large volume of operational status and condition-monitoring data.

4.

Advanced analytics to discover patterns and predict failure time.

Stochastic/ Probabilistic Predictive (sPredictive)

____________________________________

sPredictive assumes that failures are random, hence the probability that a failure takes place at a specific time is zero. However, there is a probability that a failure can happen in a specific time interval. A survival cure is then generated, and used to assess probabilities.

The key requirements for sPredictive include:

1.

Historical failure data.

2.

Status data of equipment in operation.

3.

Survival curve via application of stochastic models for assessing failure probability within a time interval.

4.

Systematic approach to compile failure probabilities, calculated individually, to determine failure volume.

Please contact us for further information regarding our PdM practice.

Our AEMPS and M24R technologies offer sPredictive
solutions that provide exceptional benefits.

Advanced Engine Maintenance

Planning Systems (AEMPS)

Materials Management for

Repairable Parts (M24R)

In essence, dPredictive assesses and identifies the most likely time of failure. Its value centered around the premise that knowing the time of failure as accurately as possible allows just-in-time performance of maintenance before breakdown taking place and creating operational interruptions. The following are needed for dPredictive:

	1.	Identification of all possible causes/ symptoms making the equipment inoperative.

	2.	Condition-monitoring system to track the equipment operational/ health status.

	3.	Systems for collecting and keeping large volume of operational status and condition-monitoring data.

	4.	Advanced analytics to discover patterns and predict failure time.

sPredictive assumes that failures are random, hence the probability that a failure takes place at a specific time is zero. However, there is a probability that a failure can happen in a specific time interval. A survival cure is then generated, and used to assess probabilities. The key requirements for sPredictive include:

	1.	Historical failure data.

	2.	Status data of equipment in operation.

	3.	Survival curve via application of stochastic models for assessing failure probability within a time interval.

	4.	Systematic approach to compile failure probabilities, calculated individually, to determine failure volume.

Please contact us for further information regarding our PdM practice.