Dear All,

I know lamps provides two ways of calculating thermal conductivity, Green Kubo and RNEMD.

I realized that another possible way of calculating thermal conductivity is possible, “Patrick K. Schelling, Simon R. Phillpot, and Pawel Keblinski. Comparison of

atomic-level simulation methods for computing thermal conductivity. Phys. Rev. B ,65(14):144306, 2002.”

Their idea is adding and removing the same amount of energy from hot and cold part of the system. This is achieved by rescaling velocities of the atoms in the layers by same factor. This will result in heat flux and temperature gradient and hence, thermal conductivity can be obtained.

Maybe I can use all the three methods and make some comparisons.

```
Briefly, this is what I think it might work in lamps.
```

```
1.Group atoms in two geometrical regions.
```

```
2. Adding and removing same amount of energy by using fix ID group-ID heat N eflux
```

```
3. get a space and time average time gradient by using fix ave
```

```
```

```
Is this viable?
```

```
Best regards
```

```
Haiming
```