I am modeling submerged circular jet (6mm dia), which is injecting at velocity of 70 m/s along x-axis. Domain size is (x-1.5m, y-0.75m, z-0.75m). I am using k-e model and elliptical-straggered equation for steady state solution.

Solution is not converging, it starts oscillating I have given relaxation (P1= -0.75, U1= 0.3,V1= 0.3, W1= 0.3 and K.E= -0.75) I have meshed jet into 3 along x-plane and y-plane

You do not say if the automatic convergence wizard CONWIZ=T or F, which is the default for recent versions of PHOENICS. If CONWIZ=F, try false time step relaxation on the velocities equal to 0.02/NX seconds, thereby making the false-transient relaxation terms the same order of magnitude as the convection terms in the momentum equations. I would use linear relaxation of 0.5 on the turbulence transport equations, but ensure that the turbulence linearisation parameter KELIN=3.

i have make the automatic convergence wizard CONWIZ=T i am using PHOENICS 3.6 even then i am not able to get the convergence

There may not be a steady-state solution - kind of like the vonKaraman vortices shed from a flagpole. Have you tried a transient solution with few sweeps (~50) per time step? Also, you didn't say whether it simply doesn't converge of if earexe.exe crashes, which sometimes happens if the solution is highly unstable and generates math errors (I see that a lot).

If earexe.exe is not crashing, you can change relaxations on-the-fly while earexe.exe is running and see if the changes affect the convergence rate.