Robust regression technique to estimate the global radiation

V. Sivamadhavi, R. Samuel Selvaraj

2012 · 11 citations · 19 references

Concepts

Abstract

In the present paper, the monthly mean daily global radiation on a horizontal surface has been correlated with meteorological parameters. Sunshine duration was proved to be the best predictor. Overall and month-wise regression equations have been developed using sunshine duration to estimate monthly mean daily global radiation on a horizontal surface at Chennai, India. For this, data analysis was done for a period of twenty years from 1990 to 2009. Overall regression coefficients for linear, quadratic and cubic polynomials were determined using linear least squares and robust regression methods. Robust method yielded a better set of values for the regression constants. The second order polynomial equation has been found to be an optimum fit for global radiation. Month-wise second order regression equations have been developed using robust method. An excellent agreement has been found between the values estimated from the proposed model and the measured values.

References

19