Drude ff for polymers

Hi. I have been trying to use Drude ff to simulate PEG-water solution. I have installed the Drude-compatible Gromacs following the instructions on MacKerell Lab, and prepared a Drude .rtp file for a PEG 9-mer molecule. I obtained the partial charges and the atomic polarizabilities of the core and Drude particles using the prediction tool found at MacKerell Lab.

I have then generated a system which consists of 185 PEG 9-mers and 315 SWM4-NDP water molecules, and applied energy minimization, nvt and npt equilibration steps at 273K. The equilibration was successful. However, I found out that temperatures did not equilibrate at exactly 273K, but around 270K. The deviation in temperature became larger when I raised the system temperature to a higher value, e.g., 378K, and the temperature ended up equilibrating at 370K. I used a timestep of 1fs, and the temperature deviation was larger when a timestep of 2fs was used. I have checked the density, pressure, and the configuration, they all looked ok except that the temperature was a bit lower than the thermostat temperature.

This is the .mdp file I am using for npt equilibration. I have tried changing the tau_t value, but the deviation in temperature still existed.
integrator = md
dt = 0.001
nsteps = 5000000
comm-grps = SYSTEM
comm-mode = Linear
energygrps =
;
nstxout = 0
nstvout = 0
nstfout = 0
nstlog = 500
nstenergy = 500
nstcalcenergy =1
nstxout-compressed = 2000
compressed-x-grps = system
;
annealing = single
annealing_npoints = 2
annealing_time = 0 50 ;ps
annealing_temp = 0 273 ;K
;
pbc = xyz
cutoff-scheme = Verlet
coulombtype = PME
rcoulomb = 1.2
vdwtype = cut-off
rvdw = 1.2
DispCorr = EnerPres
;
Tcoupl = V-rescale
tau_t = 0.1
tc_grps = system
ref_t = 273
;
Pcoupl = Berendsen
pcoupltype = isotropic
tau_p = 0.5
ref_p = 1.0
refcoord-scaling = all
compressibility = 4.5e-5
;
gen_vel = no
gen_temp = 298.15
gen_seed = -1
;
freezegrps =
freezedim =
constraints = h-bonds

; DRUDE PARAMETERS
drude = yes
drude-mode = SCF
drude-hyper = yes
drude-khyp = 16736000.0
drude-r = 0.02
drude-pow = 4

I was wondering if anyone has experience with this problem. Any help would be greatly appreciated. Thanks in advance for your time!