Replace Excel equipment log sheets with an offline Power App

On the first of the month, someone copies last month’s log sheet, renames the tab and clears the numbers. Operators fill in discharge pressure, bearing temperature, tank level and run hours on their rounds. Some do it in the truck, some at the end of the shift from a paper note.

That holds up until someone asks a question the sheet can’t answer. When did this compressor’s discharge temperature start creeping up? Was 68°C normal for this unit last summer? Who noticed the tank level reading was over the high limit, and when?

Why the monthly Excel log stops working

A monthly workbook is really twelve separate files a year, and that causes a few problems:

• There’s no trend. Six months of readings means opening six files and pasting them together by hand.
• There are no limits. The high and low limits live in someone’s head, or in a cell comment that got lost two copies ago.
• Baselines get lost. When a pump is rebuilt, its normal operating range changes. The sheet has nowhere to record that, so old and new readings get compared as if nothing happened.
• Nothing raises an alert. A reading over the limit sits in a cell until someone happens to look.

The fix is to treat readings as data and store the limits beside them.

How to model equipment, reading types and limits in Dataverse

In a spreadsheet, each measurement is a column. In this model each reading type is a row in a table, so adding “vibration, mm/s” to a compressor needs no new form and no developer.

In Microsoft Dataverse, that comes down to four tables:

• Equipment: each pump, compressor or tank, with its site and tag number.
• Reading types: what gets measured on each piece of equipment (pressure, temperature, level, hours) and the unit.
• Limits and baselines: the low limit, high limit and normal value for each equipment and reading type, with an effective date. A rebuild adds a new baseline row, and the old one stays for history.
• Readings: one row per value captured, with the time, the operator, and a status column set to in range or out of limit.

We used the same pattern for Bonanza Drilling. There, a single reusable checklist control replaced more than 20 separate checklist forms, because the questions live as rows in Dataverse. The readings model applies that data-driven approach to equipment rounds.

There’s a practical reason to use Dataverse as well. Offline Power Apps work only with Dataverse, so a log kept in Excel or a SharePoint list has to move before operators can do their rounds without signal. We move logs out of Excel and SharePoint lists into Dataverse.

Building an offline Power App for rounds

Operators use a standalone canvas Power App on a phone or tablet. They pick a site, the app lists the equipment on that round, and each piece shows the readings it needs.

Four choices keep the app working without signal:

• Check the limit in the app. Calculated columns aren’t re-evaluated on the device while offline, so the app compares the value against the limit itself. The operator sees a red flag on the spot, and the app writes the out-of-limit status onto the reading.
• Keep the download small. The default offline profile downloads every row the user can see. Filter it to the operator’s site and recent readings, and leave years of history to Power BI.
• Stay one level deep. Offline apps filter lookups only one level down. Site to equipment works well. Site to skid to pump to sensor, built on a parent-asset lookup, won’t filter offline.
• Run it as its own app. Offline works in standalone canvas apps, not inside Teams or embedded in another app.

Sync only happens while the app is open in the foreground with the screen on. Tell operators to open the app once they’re back in coverage.

Power Automate alerts when a reading crosses a limit

Flows don’t run from an offline app, so the alert can’t come from a button.

Instead, a Power Automate flow triggers in Dataverse when a reading row is added or modified. Set Select columns to the status column and add a row filter for out of limit, so it runs only for out-of-limit readings. The flow can email the area supervisor, post to a Teams channel or open a maintenance request.

Be clear with the field about timing. The alert goes out when the device syncs, not at the moment the reading is taken. At a site with coverage that is near real time. On a remote lease, it may be when the operator gets back to the truck or the office.

Two details keep alerts from getting noisy. First, Select columns fires even when the value hasn’t changed, and several updates mean several runs. Add an “alert sent” column that the flow checks first and sets once it has sent. A corrected reading that crosses the limit still raises an alert, and fixing a typo doesn’t send a duplicate.

Second, copy the limit onto the reading when it’s captured. Then the flow and the history both show which limit applied at the time.

Power BI trends for head office

Once readings live in one table, the questions that used to need six workbooks become one Power BI report:

Power BI connects to Dataverse directly, which suits early volumes. Readings tables grow fast, though. As history builds up, plan for scheduled import with incremental refresh, or Azure Synapse Link for large volumes, so reports stay quick.

What you need to build it, including licences

The short architecture:

• Dataverse tables for equipment, reading types, limits and readings
• A standalone offline canvas app for rounds
• A Power Automate flow on out-of-limit readings
• Power BI trend and exception reports for head office

Budget for licensing early. Standalone Dataverse apps aren’t covered by Microsoft 365 licences, so each operator using the app needs a Power Apps Premium licence, listed at $20 per user per month on a yearly commitment.

An equipment catalogue, preventive maintenance and inspections all build on the same equipment table. See our asset management work.

If your operators are still copying a monthly log sheet, we design and build this in your Microsoft tenant, from the data model through to the head office reports. Book a call.