Instrumentation generates a large amount of data that can quickly become overwhelming. For example, 1,440, 86,400 and 864,000 measurements would be archived if they were stored every minute, second and one-tenth of a second. Fortunately, plant historians use various techniques to reduce the number of measurements stored, so the historian does not balloon in size. When there is an issue to be addressed, determining which time frame to use is often more art than science, and sometimes short-sighted.
An analysis of a building owner’s quarterly water flowmeter was requested after the invoices increased ten-fold, and then by as much as 40-fold when compared to previous invoices. The first order of business was to obtain as many previous invoices as possible and any documents associated with this issue.
The owner provided invoices for both meters feeding the building and recent photographs of the flowmeter taken every few days. This information allowed calculation of the average daily flows during each quarter before the increase that could be compared with current average daily flows spanning a few days. Stated differently, the quarterly invoices provided a historical perspective of the amount of water consumed, whereas the photos provided more granular measurements for comparison. Historically, the less occupied part of the building consumed more water than the more occupied part. However, current measurements showed the opposite.
As sometimes happens, the invoices for the two flowmeters located in the same building had the same title, which created confusion. At some point, the titles were changed to reflect both addresses and provide some clarity. The meters reported on their respective invoices were incorrect since installation, so they were switched to appear correctly on recent invoices. It appears that this switch, plus corrections based on inconsistent meter readings, resulted in invoices that were higher than the actual amounts of water consumed.
Despite the complications, sorting this out was facilitated not by determining which time frame to analyze, but rather which time frames to analyze.
