Closed uqngibbo closed 5 months ago
Would this modification to Line 1106 of eilmer/steadystate_core.d
fix the issue?
// write out the loads
- if (!dual_time_stepping && ( (step % writeLoadsCount) == 0 || finalStep || step == GlobalConfig.write_loads_at_step )) {
+ if (!dual_time_stepping && GlobalConfig.write_loads && ( (step % writeLoadsCount) == 0 || finalStep || step == GlobalConfig.write_loads_at_step )) {
if (GlobalConfig.is_master_task) {
That's one option. We were discussing always writing the times header file or perhaps removing write_loads completely, and relying on a positive value for writeLoadsCount and dt_loads.
There wasn't much a consensus so I haven't done anything about this yet.
Fixed in eilmer 5
Volker Hanneman from DLR has found that running the steady state solver with:
config.write_loads=false
write_loads_count = 100,
Produces a malformed "times" file in the loads directory, which breaks some scripts for reading loads. Specifically this occurs because the header line is written during init_simulation, which honours the write_loads flag, but the loads files are written in steadystate_core, which does not.