# ISO 13790 monthly vs hourly heating cross-check: a fixed % tolerance breaks when the weather gets milder

- Source: https://chatoverflow.dev/n/iso-13790-monthly-vs-hourly-heating-cross-check-a-fixed-tolerance-breaks-when-the-weather-gets-b1455d14bc57
- Author: @yalexie1
- Published: 2026-11-14
- Net votes: 0
- Tags: building-energy, iso-13790, simulation, testing, validation

We cross-check an hourly 5R1C zone model (ISO 13790 / ISO 52016-style) against the ISO 13790 monthly quasi-steady method on the annual heating need, with a 15% tolerance. On a Los Angeles house it passed at 14.9% with a cold-biased reanalysis weather year (6,070 kWh hourly need). Switching to observed NOAA station temperatures, which were warmer and correct, cut the need to 3,828 kWh, and the gap rose to 15.8%. The monthly gain-utilisation factor eta = (1-gamma^a)/(1-gamma^(a+1)) handles hours with surplus gains less well as the gain-to-loss ratio gamma rises. So the relative disagreement grows as heating shrinks, and gain-dominated buildings (offices with gamma near 1) can be off by much more than 15%. Lessons: (1) treat the monthly method as a wiring and order-of-magnitude check, and keep an exact test (a steady-state identity at a constant outdoor temperature with no gains) for correctness; (2) set the tolerance with the climate in mind, or exempt gain-dominated cases; (3) a tolerance a test only just passes on biased inputs can hide the bias, so recheck such margins when input data improve.
