The Most Valuable Analysis Lives in a Spreadsheet

From Signal to Standard Work, Part 3 of 4. One upstream operator, one year of multi-sensor emissions data.

Part 2 ended at a limit. “Most of the compressor emission volumes sit in one region” is a slide, and slides don’t fix valves. This part is about the document that actually turned the corner, and what it costs to keep it alive.

The program runs on a hand-built workbook

Underneath the operator’s executive rollup there’s a hand-built workbook that ranks 65 facilities by repeat detections. The top site got hit six times in five months. Ten facilities got hit three or more times. It is the list the program actually works from.

For each high-risk site, somebody assembled the full story of what happened, pulling from every system that touched the event. One row pulls together the flyover hits, the satellite hits, the OGI inspections, the work orders, and the SCADA history, then lands on a physical root cause and a specific fix. That consolidation is needed for every event, and it is expensive. Each event requires hours of work, and days of waiting, because all the sources live in different systems, different inboxes, and arrive on different schedules. The payoff is a complete story that the crew can act on, instead of treating each data point individually.

The root-cause work behind one row

Three examples.

1. A valve on a high-pressure drain cracked in the cold and started leaking. The site had already been flagged by flyover a dozen times, but the event that finally got someone’s attention came in by phone. Nothing was wrong with the detection. Nobody had ever pulled those twelve flags into one story about one asset.

2. Dump valves on a compressor scrubber washed out. A flyover caught it on a Sunday afternoon, and the finding took two days to reach someone who could act on it, moving between systems that do not talk to each other.

3. A unit ran with its blowdown valve open, downstream of a gas-supply constraint the schedule never surfaced. The detection data showed the release but not the cause. That came from the SCADA suction-pressure trace.

None of these root causes are visible in a single source. Each one took cross-referencing to reach, and each one ended in a specific repair a crew could carry out.

The status quo, in one file

This sheet is the most useful emissions analysis in the program, because it is the one artifact that connects a detection to a cause to a fix a crew can act on.

At this operator it lives in a spreadsheet, outside the systems of record, because none of those systems were built to assemble it. Somebody builds it for their collection of facilities and rebuilds it by hand as the picture changes. And the operation underneath one of these programs never holds still. Contractors rotate, vendors change, detection providers get swapped in and out, and every change has to be chased back into the sheet or the analysis goes stale. When the person who maintains it leaves, the reasoning leaves with them.

Whether this is one process problem across sixty-five facilities, rather than sixty-five separate equipment problems, is the question that would change how the program is run. It goes unanswered because nobody has time left to assemble the data it would take.

That same spreadsheet is also a blueprint, drawn by the field, for a tool that did not exist yet. Part 4 is about building it.

If you have your own version of this spreadsheet, bring one row of it to a working session and watch it assemble itself.

Earlier in this series: Part 1 · Part 2