Changeset 2f9faf for doc/userguide/userguide.xml

Timestamp:

Mar 24, 2017, 10:12:21 AM (8 years ago)

Author:

Frederik Heber <heber@…>

Branches:

FitPartialCharges_GlobalError

Children:

a3d08c

Parents:

407d2c

git-author:

Frederik Heber <heber@…> (10/09/16 21:28:59)

git-committer:

Frederik Heber <heber@…> (03/24/17 10:12:21)

Message:

Added FitFragmentPartialChargesAction that fits partial charges to fragments in result container.

• this is one way of getting the problem of (so far) no knowing the keysets to the fragments (containing the potential grids) in the HomologyContainer. We simply don't use the HomologyContainer but the FragmentationResultContainer.
• However, this then only works for a single time step.
• FIX: Copy&pasted several functions from FitPartialChargesAction which needed to be made static to avoid duplicate symbols.

File:

• doc/userguide/userguide.xml (modified) (3 diffs)

doc/userguide/userguide.xml

@@ -1900 +1900 @@
           <note>Again, only the <productname>TREMOLO</productname>potential format is understood currently and is written.</note>
         </section>
+        <section xml:id="potentials.fit-fragment-partial-charges">
+          <title xml:id="potentials.fit-fragment-partial-charges.title">Fitting partial particle charges</title>
+          <para>There are two ways at the moment of fitting partial charges: One
+          of them uses the FragmentationResultContainer and the other uses the
+          HomologyContainer. The first, which is this action, only works for
+          a single time step, while the latter may use the fragment results
+          obtained from many time steps.
+          </para>
+          <programlisting>
+  ... --fit-fragment-partial-charges \
+      --enforce-net-zero-charge 1 \
+      --radius 1.5
+   </programlisting>
+          <para>The function call will fit partial charges to all selected
+          atoms whose fragment results are currently contained in the container.
+          It will mask the potential around each atom with a radius of 1.5.
+          Note that this is important because of the smearing of nuclei charges
+          via the "near-field-cells" parameter. Essentially, the masking radius
+          should be as large as the b-spline diameter.
+          </para>
+        </section>
         <section xml:id="potentials.fit-partial-charges">
           <title xml:id="potentials.fit-partial-charges.title">Fitting partial particle charges</title>
@@ -1927 +1948 @@
           <programlisting>
   ... --fit-partial-charges \
-      --potential-file water.particles \
       --radius 1.5
    </programlisting>
@@ -1935 +1955 @@
           from the homologies container. For each of the atoms then an average
           partial charge is computed by fitting their respective Coulomb 
-          potential to the obtained from the fragment calculations. Resulting
-          values are stored in <filename>water.particles</filename>. The
+          potential to the obtained from the fragment calculations. The
           radius is used to mask a certain region directly around the nuclei
           from the fit procedure. As here the charges of the core electrons and
           the nuclei itself dominate, we however are only interested in a good
           approximation to the long-range potential, this mask radius allows
-          to give the range of the excluded zone.</para>
+          to give the range of the excluded zone. Also, it is important to
+          remove the smeared nuclei charges via cardinal b-splines whose
+          broadened peaks are artifically and only become valid in the 
+          far-field.</para>
         </section>
         <section xml:id="potentials.parse-particle-parameters">