The shape of a real PCB trace will be most likely trapezodial, in so far the model is accounting the facts for better accuracy. But w = w1 is possible of course.
You get 94.8 ohms differential impedance and 1.742 pF/inch.
15 cm corresponds to 6" respectively 6000 mils, by the way.
I found some calculators that present 3.486 pF/inch (but it seems to be characteristic capacitance).
From what you sent (1.742 pF/inch) - it's looks like i have to divide the char capacitance by 2 to get the differential load capacitance caused by the traces?