Hi Siddhi,
The "user request" reason code in the RRAS event log is misleading, it's the generic termination code RRAS logs even when the session drops due to an idle timeout or an underlying TCP reset, not necessarily a manual disconnect.
Since this happens mid-session while users are actively working, first check Control Panel > Network Connections > right-click the VPN adapter > Properties > Options tab > "Idle time before hanging up," which should be set to "never" on both client and any Connection Manager profile pushed via GPO.
Next check whether NPS is issuing a Session-Timeout or Idle-Timeout RADIUS attribute in the connection request/network policy under Constraints, since a mismatched idle value there overrides local client settings and force-terminates the SSTP tunnel. Because SSTP rides entirely over TCP 443, also verify there's no intermediary firewall, proxy, or NAT device between the client and the RRAS server with an aggressive idle-TCP timeout (commonly 5–15 minutes), which will silently kill the tunnel even during active use if keepalives aren't being sent frequently enough, check HKLM\SYSTEM\CurrentControlSet\Services\PolicyAgent and confirm TCP keepalive interval isn't unusually long.
On the client side, disable "Allow the computer to turn off this device to save power" on the physical NIC/Wi-Fi adapter under Device Manager, as adapter power-saving is one of the most common causes of exactly this symptom on laptops. Finally, Windows Server 2012 is past extended support and has known unpatched SSTP TLS re-negotiation stability issues that were only addressed in Server 2019/2022 RRAS, so if the above doesn't resolve it, plan a migration to a supported RRAS version or consider Always On VPN as the long-term fix, since Microsoft won't issue further patches for 2012.
Harry.