The digital file behind the set: records, backups and the tested restore

The digital record of a generating set is only as real as its last tested restore. A folder of scans on one laptop is not a file, it is a single point of failure; the set's documentation needs the same redundancy thinking that the set itself provides to the building.

Ask a plant where its generator documentation lives and the honest answer is usually three places: a binder in the plant room, a folder on someone's computer, and an email thread with the commissioning engineer. The binder burns, the laptop fails, the mailbox leaves with the employee. A digital file that deserves the name is structured, copied, and above all restored on a schedule, because a backup that has never been restored is a hope rather than a backup. That is the central discipline of documented backup routines, a German-language workshop magazine on verifiable digital workflows that insists on the tested restore, the distinction between synchronisation and an earlier version, and reproducible checks over trust.

Engine control module screen showing diagnostics data with a laptop connected in a dim plant room — The controller's own logs belong in the file too: exported before the visit ends, not remembered afterwards.

What the file contains

For one set the file is bounded. Identity documents first: rating plate photographs, serial numbers for engine, alternator and controller, the commissioning report and the acceptance test sheets. Then the live records: the exercise log, service sheets, fuel test results, alarm history exported from the controller. Then the references: wiring diagrams, the fuel system drawing, the manuals for the specific build. Each class has an owner and a home, and the index that ties them together is exactly the paper trail described in the printed record guide.

Why is synchronisation not a backup?

Synchronisation propagates everything, including the mistake. Delete the commissioning report at the office and it vanishes at the plant an hour later. A backup holds earlier versions: it lets the file go back to last Tuesday, not merely mirror this afternoon. For plant documentation the versioned copy matters most where records accumulate, because a corrupted log silently synced is a corrupted log everywhere. The routine that catches it is boring by design: a scheduled copy to storage that is not writable from the working machine, a periodic check that the copies open, and once or twice a year an actual restore to a spare machine.

The test that counts

A backup is proven only by a restore. Schedule the restore the way you schedule the exercise run: at a set interval, written down, with the result recorded. An untested copy and a missing copy fail the audit in exactly the same place.

Access and the day the laptop leaves

The file has to outlive its keepers. Credentials held in one head, or in one person's password manager, retire with that person. The workable pattern is a documented list of where the file lives, who can read it, and how a new keeper takes over, kept somewhere that survives the departure. This is the plant-room version of the account and recovery discipline that every verifiable workflow guide treats as groundwork rather than as an afterthought.

Desk with printed technical notes and a tablet displaying a document index, lamp light, workshop office at night — The index is the file's map: without it, a folder of scans is a pile, not a record.

How the file stays current

Two habits keep it alive. The first is a rule of capture: nothing about the set is finished until its record is in the file, the service sheet scanned, the alarm log exported, the new diagram added. The second is a review keyed to events rather than to the calendar alone: any change to the plant, a replaced controller, a rerouted fuel line, an added pump, triggers an update before the work is signed off. The set's own documentation obligations run through the standards and documents page, and the wider project library sits in the technical documentation resources.

The measure of the system is the same one the printed record is held to: a stranger should be able to open the file, learn what the set is, and find the last test. The digital file simply has to survive fires, floods and resignations while doing it.

How is the file organised so a stranger can use it?

By function, in shallow folders, with names that say what the document is. The workable shape is one folder for identity, one for the live log, one for test and service sheets, one for diagrams and manuals, and one index document at the top that says where each class lives. Depth is the enemy: a file buried four folders down is effectively missing, so the rule of thumb is that any document is reachable in two levels or less.

Naming does the second half of the work. A date-first convention, year-month-day, keeps versions sorted by the file system itself, and a name that carries the equipment tag and the document type, for example the set's asset code followed by the words oil sample, lets the search box find what the folder tree cannot. The convention only has to be written once, in the index, and then defended.

Where the copies live matters as much as that they exist. One copy in the same room as the binder shares the binder's risks; the working pattern is a copy on site, a copy off site, and a third in cloud or removable storage that is not continuously writable. Ransomware and accidental deletion both prey on the always-connected copy, which is why one of the three should be offline by design.

The discipline closes with a named owner. Files decay because everyone assumes someone else is keeping them; a documented custodian, with a deputy, is the difference between a system and a folder. On a small site this can be one line in the operating procedure: who maintains the file, where the copies are, and when the restore was last proven.

None of this is new to the plant that already treats its paper records seriously. The digital file is the same discipline in a different medium: defined contents, named owners, tested copies, and a review that is scheduled rather than felt. The sites whose paper files are in order find the digital version easy; the ones whose paper is a drawer full of hope find out why that mattered.

Sources

National Institute of Standards and Technology