Worldwide benchmarking of cost-effective radiometers for direct and diffuse irradiance

Solar energy projects can benefit from direct normal irradiance (DNI) and diffuse horizontal irradiance (DHI) measurements during all project phases. Several commercial measurement systems for DNI and DHI are available. Sun trackers with pyranometers and pyrheliometers can provide highly accurate measurements but are often impractical in solar energy applications. For less expensive and more robust sensors, it is often unclear which accuracy can be expected under a project site’s specific atmospheric conditions. We address this challenge through our dedicated experimental comparison of relevant sensor systems (rotating shadowband irradiometer [short RSI], Delta-T SPN1, EKO MS-90, PyranoCam, Sunto CaptPro, Kipp & Zonen CSD3) at up to six sites worldwide. The RSI systems ( rRMSD 3 to 8.6 % , DNI; 4.8 to 7.6 % , DHI) and PyranoCam ( rRMSD 2.6 to 5.2 % , DNI; 4.4 to 5.8 % , DHI) exhibit similar error metrics and are the most accurate systems in the test. Delta-T SPN1 and EKO MS-90 ( rRMSD 6.8 to 15 % , DNI; 10.6 to 20.1 % , DHI) but especially Kipp & Zonen CSD3 and Sunto CaptPro show significant deviations ( rRMSD 17.7 to 20 % , DNI; 33 to 58 % , DHI). We evaluate the influence of relevant atmospheric parameters on the sensors’ accuracies by a rather unique measurement setup. MS-90’s DNI errors depend on DNI itself, with overestimations for low reference DNI. The deviations of SPN1’s DHI and DNI measurements increase sharply in situations with high circumsolar irradiance. Also CaptPro and CSD3’s increased measurement errors are related to circumsolar irradiance. For RSI and PyranoCam, only moderate influences on the measurements are identified, indicating a general applicability of these instruments. • Accuracy of practically feasible sensors for direct and diffuse irradiance evaluated. • EKO MS-90, Delta-T SPN1, RSI/RSP, PyranoCam, Sunto CaptPro, Kipp & Zonen CSD3 tested. • Six sites worldwide included with approximately one year of measurements per site. • Influences of atmospheric and site conditions on the sensors’ accuracies investigated. • Strong differences in accuracy found between the radiometer types.