Attributes of an Effective Maintenance Program for Data Center Physical Infrastructure

Overview

The paper places maintenance of UPS, switchgear, cooling and other physical infrastructure on a four-step maturity ladder: reactive, calendar-based (time-based), condition-based and risk-informed. In the calendar-based model that is common today, preventive visits happen quarterly, semi-annually or annually depending on device type; the visit takes place when the date arrives regardless of the device's actual condition. The reactive model acts only after a failure.

Connected equipment and analytics are what make condition-based maintenance possible: because the device reports its own state, work is done when needed and becomes less invasive. The end of the roadmap is risk-informed maintenance, where decisions are driven by risk instead of fixed intervals. The paper names connected equipment as the basis for this move. This cuts unnecessary invasive visits and surfaces a real deviation early through remote monitoring.

Five attributes are listed for choosing a provider: personnel expertise, quality assurance, on-site response time, remote monitoring and thorough inspections during preventive visits. Response time feeds directly into MTTR and therefore availability. The paper dates from 2021 and contains no quantified benefit; the related standards are NFPA 70B, EN 50600-3-1 and the Uptime Tier Standard: Operational Sustainability. Two of the five, remote monitoring and response time, tie directly to the condition-based model and to MTTR.

Key points

1. Calendar-based programs schedule preventive visits quarterly, semi-annually or annually per device type.
2. Maturity progression: reactive, calendar-based, condition-based, risk-informed maintenance.
3. Five provider attributes: expertise, QA, on-site response time, remote monitoring, comprehensive PM inspections.
4. Connected equipment and analytics enable less invasive condition-based maintenance.
5. Response time feeds directly into MTTR, so it is an input to the availability model.

Using it in design

• In the operations module attach a maintenance strategy and default PM interval (quarterly/semi-annual/annual) to each catalog asset class, with a condition-based monitoring option.
• In the availability model tie MTTR to the response-time SLA and feed opex estimates.
• Add a maintainability check that every redundant component can be isolated for PM (concurrent maintainability).

Who should read it

Operations managers selecting service contracts and designers building availability models.

Related standards

• NFPA 70B
• EN 50600-3-1
• Uptime Tier Standard: Operational Sustainability

Publication details

Authors

Ken Belanger; Eric Brun; Prasanna Kanchikere; James Martinec; Wendy Torell

Publisher

Schneider Electric

Year

2021

Document number

WP 264

Document language

English

Source type

Manufacturer

Access

Free to read

Licence

Copyright retained; citation only — the publisher’s terms apply to the original document

Citation