Orbit propagation - modified initial orbit?

Well I didn’t understand it either but seeing the values of the acceleration, it may be the reason of the observable differences.

There is however still a small difference between a maneuver the 01/09 and one the 10/09:

Here is an updated version of the code :
ManeuverPropagation_mastermode.java (11.4 KB)

With a maneuver of 0m/s the 01/09/2018 at 01h00, the console output is the following :

Orbite initiale :
Date de début : 2018-08-31T00:00:00.000
Cartesian parameters: {2018-08-31T00:00:00.000, P(3.77279251691374E7, 1.8861471230199E7, -26502.8594459892), V(-1372.8424195657, 2750.00801092554, 0.04512254183365766), A(-0.20039105190141654, -0.10018444972438904, 1.4169550626880686E-4)}
Orbite en fin de propagation :
Date de fin : 2018-09-01T01:00:00.000
Cartesian parameters: {2018-09-01T01:00:00.000, P(3.10710382907986E7, 2.8535097918909054E7, -25533.85799639652), V(-2077.6794344539444, 2264.4437171303443, 0.6330152605735903), A(-0.16496233177155464, -0.15150153045018774, 1.3772286843900033E-4)}
dV : {-0; 0; 0}
Orbite après manoeuvre :
Date de la manoeuvre : 2018-09-01T01:00:00.000
Cartesian parameters: {2018-09-01T01:00:00.000, P(3.10710382907986E7, 2.8535097918909054E7, -25533.85799639652), V(-2077.6794344539444, 2264.4437171303443, 0.6330152605735903), A(-0.16496328350926823, -0.1514993932261263, 1.35565120686435E-4)}
Orbite initiale :
Date de début : 2018-09-01T01:00:00.000
Cartesian parameters: {2018-09-01T01:00:00.000, P(3.10710382907986E7, 2.8535097918909054E7, -25533.85799639652), V(-2077.6794344539444, 2264.4437171303443, 0.6330152605735903), A(-0.16496233177155462, -0.15150153045018772, 1.377228684390003E-4)}
Orbite en fin de propagation :
Date de fin : 2018-09-10T00:00:00.000
Cartesian parameters: {2018-09-10T00:00:00.000, P(3.396374607230927E7, 2.5014699678586394E7, -43029.03873951323), V(-1821.2222727507792, 2475.7268082216992, 0.6779891048675524), A(-0.18037992553756985, -0.13286191399002403, 2.2722304957728922E-4)}
Orbite finale :
Date de fin : 2018-09-10T00:00:00.000
Cartesian parameters: {2018-09-10T00:00:00.000, P(3.396374607230927E7, 2.5014699678586394E7, -43029.03873951323), V(-1821.2222727507792, 2475.7268082216992, 0.6779891048675524), A(-0.18037992553756982, -0.13286191399002403, 2.272230495772892E-4)}

With a maneuver at the end of the propagation, the output is the following :

Orbite initiale :
Date de début : 2018-08-31T00:00:00.000
Cartesian parameters: {2018-08-31T00:00:00.000, P(3.77279251691374E7, 1.8861471230199E7, -26502.8594459892), V(-1372.8424195657, 2750.00801092554, 0.04512254183365766), A(-0.20039105190141654, -0.10018444972438904, 1.4169550626880686E-4)}
Orbite en fin de propagation :
Date de fin : 2018-09-10T00:00:00.000
Cartesian parameters: {2018-09-10T00:00:00.000, P(3.39637460890208E7, 2.5014699658136625E7, -43029.03891703195), V(-1821.222271099359, 2475.7268093313883, 0.6779890964996327), A(-0.18037992560927554, -0.13286191386885463, 2.2722305049791068E-4)}
dV : {-0; 0; 0}
Orbite après manoeuvre :
Date de la manoeuvre : 2018-09-10T00:00:00.000
Cartesian parameters: {2018-09-10T00:00:00.000, P(3.39637460890208E7, 2.5014699658136625E7, -43029.03891703195), V(-1821.222271099359, 2475.7268093313883, 0.6779890964996327), A(-0.1803804399785667, -0.1328523219564699, 2.285259391400916E-4)}
Orbite finale :
Date de fin : 2018-09-10T00:00:00.000
Cartesian parameters: {2018-09-10T00:00:00.000, P(3.39637460890208E7, 2.5014699658136625E7, -43029.03891703195), V(-1821.222271099359, 2475.7268093313883, 0.6779890964996327), A(-0.1803804399785667, -0.1328523219564699, 2.285259391400916E-4)}

Any idea of where this difference comes from ? Is it from the infinitesimal difference seen at the maneuver on the 01/09 ? If so, would it be possible not to have this difference ?

Thanks for your answers

Nicolas