Oil‑Club → rules → OCM
Real-world scenarios
Public Oil‑Club topics are used as a research corpus: VOA/UOA, engine hours, average speed, top-up, cold starts, short trips, laboratories, driveline fluids, coolants and repeated sample series.
Deterministic utility
A recurring practical question becomes a small input → calculation → rule → output module. There is no model training in runtime and no hidden coefficients.
Standards layer
Methods and terminology are linked to ASTM, ISO, API, ACEA, IEC, IEEE and OEM context. The forum helps identify the use case but does not become a normative document.
Trend instead of magic
Where one sample is insufficient, the system uses baselines, normalization per 100 hours/1000 km, deltas and sequential measurements. This is sufficient for most OCM tasks and remains explainable to engineers.
Oil‑Club research import
A separate importer is intended only for developer-side corpus preparation: public HTML pages without authentication → structured facts and source links. The web portal does not crawl the forum or process forum attachments during user execution.
Why laboratory and method are mandatory
Oil‑Club contains same-sample comparisons across laboratories, so laboratory method and baseline remain part of the context. Measurement dispersion must not be hidden behind a 'smart model'.
ML on request, not inside Auto OCM
If a customer accumulates a high-quality labelled history, a separate ML application can be developed. Base Auto OCM stays small, transparent and suitable for a server without a GPU.