The further improved model can thus be described as shown in expression (7), where the reduction factor, fCopen, depends on Copen as fCopen=(1-K′) Copen, and K′ is a coefficient to be determined empirically (in our particular case takes a value of approximately 0.5).