A completed local pilot

The pilot took place at an air-conditioned office in Shah Alam, Malaysia. Roof work was carried out in February 2026, followed by wall and window work in May. The analysis covers one air-conditioned office.

Electricity and temperature records were collected during January–July 2026 at five-minute intervals. The power dataset contains about 110 days and 31,300 samples. Temperature and humidity logging had a gap between 26 February and late May; this was not uninterrupted six-month monitoring.

What the analysis found

For comparable periods with a 26°C set point, the analysis estimated AC power by subtracting non-AC loads and adjusting for outdoor temperature. Estimated AC power was about 35% lower after the combined roof and wall intervention. A day-level bootstrap gave a 95% interval of approximately 24–50% reduction.

Recorded indoor temperatures while AC was operating were similar: 24.9°C before and 24.6°C after. These are observations in a before-and-after analysis, not proof that all of the difference was caused by the retrofit.

Why this is not a 35% bill-saving claim

AC operating time increased from about 4.65 to 12.42 hours per day. Estimated daily AC energy therefore increased from 9.1 to 13.8 kWh. Lower condition-adjusted power and higher total energy can coexist when use changes.

For another facility we need its roof and wall construction, cooling area, operating hours and baseline data before estimating potential savings.

What remains unresolved

  • The roof-only effect cannot be determined because temperature data is missing for that period.
  • Roof and wall effects cannot be separated.
  • There was no untreated comparison building. Solar and seasonal differences are not fully controlled.
  • AC maintenance history and sensor placement remain unverified. They may affect the interpretation.
  • The confidence interval quantifies uncertainty in the analysis; it does not remove these potential confounding factors.

How this changes the next project

Agree the measurement plan before work, keep temperature and power logging operating throughout the project, and record set points, operating hours and equipment maintenance. A larger factory needs its own diagnosis and measurement plan. Ask us how we would assess your facility.