Base Driven Time History Collapse Analysis

I have been conducting some sample seismic time history collapse analysis with a simplistic model to check how realistic the results are from the SACS Collapse module. The SACS model I created was a simplistic monopile structure with a length of 30m, 2m diameter and 50mm wall thickness. The monopile is fixed at the bottom. I conducted a time history seismic analysis with a point mass of 100T at the top of the monopile and a constant only horizontal base acceleration of 0.2*g. 

Acceleration time graph: A constant horizontal base acceleration of 0.2*g

Base shear graph: This is how the base shear graph looks like from the dynamic response, which is something we would expect; oscillations at the beginning (transient response) and then a constant value (steady state response).

Corresponding Base shear graph from the Collapse Module:

From the base shear graph, it can easily be seen that the (dynamic inertia) loads get accumulated at each load step. For example, the maximum base shear value is equal to the number of load steps (2500) times the inertia load (196 kN), which is equal to 490,000 kN.

 Such result does not look realistic since the constant acceleration should lead to a constant base shear without any accumulation. How is it possible to do the analysis without such accumulation?

If the base shear is accumulated, every structure would fail after a while, even with a very small constant acceleration. For me it does not make any sense. Or do you think the accumulation is correct?