The stress of PW SOC is wrong

[pxlxingliang]

Describe the bug

I have do the finite difference test on Fe2 with noncollinear and SOC, and the deviation between analytical value and finite difference value is large, and the finite difference value is close to that by QE

The stress by different methods

	11	22	33
ABACUS-Ana	60.302	60.302	60.312
ABACUS-FD	53.528	53.530	53.490
QE	53.627	53.627	53.576

Expected behavior

No response

To Reproduce

No response

Environment

No response

Additional Context

No response

Task list for Issue attackers (only for developers)

• Verify the issue is not a duplicate.
• Describe the bug.
• Steps to reproduce.
• Expected behavior.
• Error message.
• Environment details.
• Additional context.
• Assign a priority level (low, medium, high, urgent).
• Assign the issue to a team member.
• Label the issue with relevant tags.
• Identify possible related issues.
• Create a unit test or automated test to reproduce the bug (if applicable).
• Fix the bug.
• Test the fix.
• Update documentation (if necessary).
• Close the issue and inform the reporter (if applicable).

[WHUweiqingzhou]

This issue has been solved in our private branch, and will be closed until that PR merges.