I have been having a problem with my results for quenching a spin-1/2 XXZ model in the TEBD code and both my colleagues and I are stumped. I am attempting to study a quench from Delta = 4 to 2 in the XXZ model, starting from the ground state of the original Hamiltonian, and looking at the local magnetisation and correlation functions that result. I tried this out in small systems (around L=32), following the tutorials very closely and adapting them to my situation, and looking at the local magnetisation. When I included S_z (the total S_z) as a conserved quantum number in my parameters, I got no time dependence in my results, with each S_z being 1/2 everywhere along the chain for every time, which seems to be nonsensical. However, when I removed this constraint, I got (after a much longer simulation) more sensible results, with a time-dependence and the results being closer to 0. I am fairly certain that S_z should be conserved here, so my question is why does this not seem to work? I followed the tutorials very closely, and those scripts seem to work well enough. The parameter section of my script is as follows:

import pyalps
import matplotlib.pyplot as plt
import pyalps.plot
import numpy as np
import copy

parms = [ {
    'L'                : 32,
    'MODEL'                : 'spin',
    'local_S'            : 0.5,
    'Jxy'                : 1,
    'Jz'                : 4.0,
    'ITP_CHIS'            : [40, 40, 40],
    'ITP_DTS'            : [0.05, 0.05, 0.025],
    'ITP_CONVS'            : [1E-9, 1E-9, 1E-10],
    'INITIAL_STATE'            : 'ground',
    'CHI_LIMIT'            : 40,
    'TRUNC_LIMIT'            : 1E-12,
    'NUM_THREADS'            : 1,
    'TAUS'                : [0.0, 50.0],
    'POWS'                : [0.0, 1.0],
    'GS'                : ['Jz', 'Jz'],
    'GIS'                : [4.0, 2.0],
    'GFS'                : [2.0, 2.0],
    'NUMSTEPS'            : [100, 2500],
    'STEPSFORSTORE'            : [5, 5]
    } ]

baseName = 'Jz4to2Quench1L32'
nmlname = pyalps.writeTEBDfiles(parms, baseName)
res = pyalps.runTEBD(nmlname)

Data = pyalps.load.loadTimeEvolution(pyalps.getResultFiles(prefix = 'Jz4to2Quench1L32'), measurements = ['Local Magnetization'])