Do C.TShen HChan Y.-CLiu X.YING-CHIEH CHAN2021-08-052021-08-05201925222708https://www.scopus.com/inward/record.uri?eid=2-s2.0-85107473283&partnerID=40&md5=5f45cb72cf65c6886bbe3703aa795ac4https://scholars.lib.ntu.edu.tw/handle/123456789/576104Illuminance prediction is critical for accurate daylighting simulation which is essential for daylit space designs. However, illuminance data are unavailable for many locations. There are several widely used luminous efficacy models to estimate illuminance from irradiance such as Perez, Littlefair, Muneer, and Chung models. To validate the performance of these models when applied to different locations, we recorded global and diffuse irradiance, global illuminance, and some meteorology parameters in Taipei and Kingsville. Subsequently, we provided luminous efficacy models for these specific locations by using regression and neural network. The results showed that Muneer and Perez models provide good estimation. Water content described in Perez model does not show significant influence in most of the analysis. Additionally, luminous efficacy does not seem to vary due to different locations, at least for temperature and humidity. ? 2019 by International Building Performance Simulation Association (IBPSA) All rights reserved.Lighting; Diffuse irradiance; Luminous efficacy; Model evaluation; On-site measurement; Optimization techniques; Space design; Specific location; Temperature and humidities; Locationconference paper2-s2.0-85107473283