HDR图像测量天空亮度及在动态采光模拟中应用

为解决在通用天空模型下进行建筑采光模拟所得到的结果准确程度不高的问题,尝试使用实测的天空亮度分布HDR图像作为光源进行建筑采光/动态采光模拟并对该方法进行了验证.首先,说明了通过高动态范围(HDR)图像测量天空亮度分布的方法;进而,介绍并验证了基于图像的采光/动态采光模拟(IBL):既使用HDR图像作为光源输入数值进行采光模拟的方法.验证实验于2017年7月4日至7月10日及2017年8月3日至8月12日进行,通过实测与模拟结果的比较分析验证了HDR图像测量天空亮度分布以及基于图像的采光模拟的准确程度.最后,文中讨论了基于天空数学模型的采光模拟以及基于图像的采光模拟的优缺点.结果表明:通过HDR图像可以准确地测量天空亮度分布;基于图像的采光模拟准确度良好,同一时刻模拟值与实测值之误差集中分布在-15%~47%范围之内;基于图像的动态采光模拟所求取的100 lxUDI3 000 lx指标值最大误差限制于±25%以内.

HDR Image in measurement of sky luminance distribution & application in dynamic daylighting simulation

To solve the problem of low accuracy of the results generated by building daylighting simulations under the generic sky models, this paper attempts to use the on-site measured sky luminance distribution HDR images as the light source to conduct the building daylighting / dynamic daylighting simulation and has validated the method. First, the method of measurement of sky luminance distribution via HDR image is indicated, and the image-based daydaylight lighting / dynamic daylighting simulation (IBL) and the method that uses HDR image as the light source input value in daylighting simulation are introduced and validated. The validation experiments are carried out from: 4^th to 10^th Jul. and 3^rd to 12^th Aug. 2017. The interior illuminance and luminance distribution are measured via instruments and meanwhile, the sky luminance distribution (including the exterior environment) is measured simultaneously. After comparing and analyzing the measured and simulated results, the accuracy of HDR image in measurement of sky luminance distribution and image-based daylighting simulation is validated. Finally, the pros and cons of sky mathematical model based daylighting simulation and IBL image based daylighting simulation are discussed. The results show that the sky luminance distribution could be accurately measured via HDR image, the accuracy of IBL simulation results of IBL is well performed, the the errors between simulation and measurement in one instant are concentrated focused within the range of -15%~47%, and the max error of 100 lx < UDI < 3000 lx generated from IBLimage-based dynamic simulation limits within ±25%.