assume(-0.008 <= delta_1 & delta_1 <= 0.008)
assume(-0.002 <= delta_2 & delta_2 <= 0.002)
CL = 0.331 + delta_1 + (0.15+delta_2)*ALPHA
The matrix approach is not correct, as it gives specific values, whereas you need a range of values. For example you might have 1/(CL-0.15*ALPHA-0.331) as part of an expression, and if you only work with the edge cases you miss the fact that this passes through 1/0 .