Minus liquid pressure occurs when adding boundary condition in a volumetric mesh

Hi all,
I am running an example about the stress distribution during water injection into a reservoir using TH2M model in OGS6. Here is the prj link.
initial-fracturing-test1.zip (4.7 KB)
I tried to use the temperature boundary condition in a sphere volumetric mesh representing the injection well instead of the heat source term to represent the cold water injection, in order to test thermal response in the reservoir. Under this condition, the temperature of the whole injection sphere is a certain value, I guess. The initial temperature condition is 353.15K. For testing, the initial temperature difference is 1K between the injection volume and the reservoir. However, I found in the result after 100s’ injection that the liquid pressure is minus around the injection well like fig1, while the temperature boundary condition is successfully added into the volume. When I deleted the volumetric temperature boundary condition, the liquid pressure is plus in 100s result and will be larger with the increase of the pressure source term like fig2. I don’t know where the problem is. How can a small 1K temperature difference cause so large pressure perturbation in the reservoir?
pressure with T BCpressure without T BC
Could anyone give me some advice?