Turn the evidence into a maintenance plan

Combine HI, RUL, hazard, and asset criticality, then schedule the first jobs within budget and crew constraints.

Question for this chapter

How do HI and RUL become an executable order within budget and crew limits?

Why this matters now

The asset with the lowest HI is not always the one with the shortest RUL. Planning needs one visible rule for current condition, remaining time, failure tendency, criticality, cost, and crew capacity.

Try it

Open the maintenance_planning pipeline and select Run.
Sort maintenance_schedule by priority_score in descending order.
Run the periodic_report pipeline.
Open the equipment_health_overview dashboard.

Successful result

The schedule has ten rows, with the top three assets all assigned to scheduled day 0.

Rank	Asset	HI	RUL	Priority	Maintenance type	Estimated cost
1	`EQ-008`	77.0742	1.9643 years	0.3501	`MAJOR_OVERHAUL`	KRW 24,879.63
2	`EQ-005`	76.2767	1.9813 years	0.3427	`MAJOR_OVERHAUL`	KRW 21,112.47
3	`EQ-003`	79.1367	2.8840 years	0.2405	`PREVENTIVE`	KRW 12,498.28

The total estimated cost is KRW 216,371.74, below the KRW 500,000 budget. No asset is excluded by the budget. The maximum of three jobs per day spreads the work across scheduled days 0–3: three assets on each of days 0, 1, and 2, then one asset on day 3.

Read the result

EQ-008 and EQ-005 both have voltage-based criticality 3. EQ-005 has slightly lower HI, but EQ-008 has shorter RUL and the highest hazard rate among all ten assets, leaving it 0.0074 ahead on the final priority score. This run therefore places EQ-008 first and EQ-005 immediately after it.

Deep dive — four factors in the priority score

The current code multiplies four factors.

priority = voltage-based criticality
         × (1 - HI / 100)
         × hazard rate normalized by the maximum across ten assets
         × (1 / RUL)

A higher score schedules earlier. The rule turns unlike evidence into an order, but it does not include statutory inspection periods, actual outage windows, crew qualifications, or material availability.

Confirm the same result in the operations view

The eight dashboard widgets map one-to-one to the source JSON:

관리 설비 수 — 10 assets
평균 건전성 지수 — 71.9
이상 탐지 설비 수 — 5 assets
잔여수명 2년 미만 설비 — 2 assets
설비별 건전성 지수 — HI bars in ascending order
설비별 잔여수명(년) — RUL bars in ascending order
건전성 등급별 설비 현황 — six B and four C assets
정비 최적화 일정 — jobs ordered by descending priority score

The monthly report row records 10 assets, average HI 71.86, six B and four C grades, two assets below two years RUL, five thermal anomalies, ten scheduled jobs, and KRW 463,298.13 in historical maintenance cost. The dashboard and report do not create a new decision; they regroup the pipeline outputs for operational questions.

Next decision

You have a first planning order. Next, check the observed values, applied rule, and remaining field checks in four questions before preparing the final maintenance-meeting report.