R.K. Diagnostics – Diagnostic Algorithms

Diagnostic Algorithms

R.K. Diagnostics delivers software solutions that integrate knowledge and expertise from several disciplines. These algorithms and associated software are designed for easy adaptation to the specific system architecture selected by each customer.

The algorithms integrate concepts of model-based systems, data driven algorithms and expert knowledge. Generic model-based algorithms for diagnostics of mechanical components permit rapid development of new applications to meet each customer’s machinery diagnostic needs.

The existing algorithms' infrastructure allows analysis of large amounts of data from various sources - including vibrations, acoustics, temperatures, pressures, and other environmental and operating parameters.

The system allows integration of human expert knowledge with advanced learning algorithms providing high-accuracy diagnostic and prognostic solutions. The initial system starts with a "generic" knowledge base and is continually refined as more data is acquired.

Based on the vibro-acoustic signatures of the monitored machinery, the diagnostic algorithms reliably and accurately detect and distinguish among a diverse set of mechanical failures such as:

Damage in bearings (damages of the surfaces as well as deformations of the inner race, outer race, rolling elements, or cage)
Damage in gear wheels (distributed and localized damages, surface wear, partial or total tooth break, cracks)
Misalignment or imbalance
Oil whirl or whip
Pump malfunctions
Rotor blades partial or total loss
Rotating stalls
Installation problems

Architecture design

The software building blocks are designed so that they can be adapted to fit different types of machinery, any range of frequencies, rotating speeds, combination of rotating or reciprocating components and geometries. The open architecture provides maximum flexibility and simplifies further development and deployment of additional data manipulation procedures. Designed to support diversity of complex machinery, the system can manage and analyze very large amounts of data from various sources.

The Process

The vibration based Diagnostics & Prognostics process is divided into four major stages:

Data acquisition and raw signals validity check – sensor setup and positioning, signal conditioning, data acquisition and recording parameters, data validation, signal quality
Signal processing – separate the effects of the possible failure modes of each component from among the numerous excitation sources
Feature extraction – assign values to a set of health indicators that are correlated with various failure modes of the system components
Artificial intelligence methods for Classification, Pattern recognition and Fusion – analyze and combine all extracted data for automatic recognition of evolving failure patterns and generation of detailed alerts.


	Benefits
	Highest quality analysis – identify the exact fault type based on vibro-acoustic data to pinpoint the problematic component Predictive prognostics lets you know how much time is left before failure so you can plan maintenance schedule Reliability and accuracy of detection minimizes false alarms GUI application graphically displays data in near real time Fully automated process – from data acquisition thru decision process – for timely decision-making Expert support in adapting diagnostic algorithms and software to each customer’s particular hardware (sensors and CPU)