Machinery Lubrication magazine published by Noria Corporation
Issue link: https://www.e-digitaleditions.com/i/229047
OIL ANALYSIS "Technology cannot make up for a bad sample," Underwood warns. Incomplete or illegible information can lead to data-entry errors and limited testing that yields suboptimal reporting. Once restricted to whatever information could be scribbled on a small label, the latest software programs allow maintenance technicians to input critical information, including equipment (make, model, identification number, location, etc.), hours of operation, maintenance activities, drain interval and more. While incomplete information doesn't affect the test results, it significantly impacts the analyst's ability to draw conclusions or detect trends. Therefore, it is essential that users provide repetitive, information-rich and credible samples to ensure quality and meaningful reports. When more data points are given during the sampling process, laboratory analysts can deliver more comprehensive reports. With consistent and complete sample information, labs can ensure normalization of results based on the organization's result history. Take Advantage of Data Mining A highly technical area of computer science, data mining extracts information from a set of data and transforms it into understandable and actionable information. In oil analysis, this process uses data management and complex metrics to detect abnormalities in single samples or groups of samples. While laboratory experts excel in extracting comprehensive information from a sample, end users may find it difficult to put technical information into practical terms. The average manager typically isn't interested in particle counts or the presence of iron or metals in a single piece of equipment. However, the ability to recognize trends across a population of equipment can signal a bigger problem that could result in lost revenue from downtime or expensive repairs. According to Underwood, data mining is particularly helpful when managing fleets. "The ability to compare units and equivComparison graphing offers a visual comparison of equipment alent services helps companies determine performance against a population of data, allowing plant personnel to determine which makes and models are best suited for each site. what the best product on the market is for their particular business," he says. It is important to note that data mining is not the end user's responsibility but rather an important and integrated component of any effective software program. Use Graphical Comparisons Graphing sample conditions enables users to easily spot trends in specific units, equipment types, makes or models. 38 November - December 2013 | www.machinerylubrication.com Graphs and other visual representations highlight the severity of non-conforming data far better than tables and spreadsheets. Keep in mind that if a report isn't readable, it won't get read. "A picture is worth a thousand words," Underwood says. "People understand a graphical data presentation much more readily than a bunch of numbers, so it is a critical component to any software program." Users should be able to select different graphing styles (line, bar, area, spider, etc.) based on preference and need. Especially helpful in comparing a pre-defined set of