Experimental–analytical method determination of the thermal power of a LED COB module for the design of cooling systems in energy-efficient lighting systems
摘要
The article presents a method for determining the thermal power of a LED COB module as a critical parameter for engineering the cooling systems of high-power LED luminaires. The research problem is the lack of simple yet accurate analytical models of thermal operating regimes for high-power LED modules that are suitable for rapid engineering assessment and embedded controllers. This is important because incorrect estimation of heat dissipation leads to excessive junction temperatures, reduced luminous efficacy, and shortened service life. The novelty lies in a physics-informed selection and comparison of four regression models—linear, with exponential and logarithmic temperature terms, and a bivariate quadratic form—with explicit inclusion of the second-order interaction between temperature T and current I. Experiments were conducted over ambient temperatures from − 30 to + 40 °C and drive currents from 1 to 5 A with fixed step sizes; regression models were fitted to measured electro-photometric quantities, and their accuracy and computational cost were assessed. The results show a monotonic increase of thermal dissipation with increasing T and I, as well as a statistically significant T⋅I interaction. The quadratic regressor provided the best accuracy–complexity trade-off, achieving RMSE≈0.8 W and a mean relative error < 1% across the entire operating window. Accounting for second-order terms and the T⋅I interaction is necessary for an adequate description of thermal losses in high-power LEDs. The proposed quadratic model is a practical surrogate for engineering use (thermal regime calculations, boundary/source terms in CFD, and LED driver firmware).