filtering - Matlab: How to apply low-pass filter on tf-system for faster evaluation in simulink -

i have mechanical oscillation system defined n x n matrix transfer functions tf( ... ).

w = minreal( [  tf( ... ) ... tf(...) ; ... ; tf( ... ) ... tf(...)  ]; 

in following picture can see selected frequency responses. shows various irregularites @ high frequencies.

as combine system in simulink other high-order systems, required step-size has extremely low or system not stable. simulation time tremendously high, makes impossible validate general funcionality of model.

for reason i'd apply low-pass filter on fransfer matrix, use bigger steps faster simulation time. there way implement either in matlab code or within simulink?

finally adjust threshold frequency depending on how time have , accuracy required.

i did research appropriate solvers, without success. advice regarding solvers me well.

this meager list of toolboxes have available:

control system toolbox                                version 9.3        (r2012a) simulink control design                               version 3.5        (r2012a) system identification toolbox                         version 8.0        (r2012a) 

thank in advance!

edit: picture illustrate suggestion of @am304

bode diagram reduced system 18th 4th order, low-pass in case not required anymore.

as suggested in comments, convert matrix of transfer functions state-space system , use state-space block simulink, suspect more efficient (certainly more readable if n large).

also, maybe @ reducing order of system (n) (see model simplification in documentation more details - in particular balred , balreal)