Updated documentation with addition of Brouwer-Lyddane model.

src/design/available-propagators-class-diagram.puml

@@ -76,6 +76,7 @@
         AbstractAnalyticalPropagator <|-- AdapterPropagator
         AbstractAnalyticalPropagator <|-- EcksteinHechlerPropagator
         AbstractAnalyticalPropagator <|-- KeplerianPropagator
+        AbstractAnalyticalPropagator <|-- BrouwerLyddanePropagator
 
         AbstractAnalyticalPropagator <|-- Ephemeris
         BoundedPropagator            <|.. Ephemeris

src/design/propagator-conversion-class-diagram.puml

@@ -57,6 +57,7 @@
       class KeplerianPropagatorBuilder
       class TLEPropagatorBuilder
       class EcksteinHechlerPropagatorBuilder
+      class BrouwerLyddanePropagatorBuilder
 
       interface PropagatorConverter {
         +Propagator convert(Propagator source, double timeSpan, int nbPoints, Collection<String> freeParameters)
@@ -87,6 +88,7 @@
       PropagatorBuilder <|.. KeplerianPropagatorBuilder
       PropagatorBuilder <|.. TLEPropagatorBuilder
       PropagatorBuilder <|.. DSSTPropagatorBuilder
+      PropagatorBuilder <|.. BrouwerLyddanePropagatorBuilder
       PropagatorBuilder <|.right. EcksteinHechlerPropagatorBuilder
 
       PropagatorConverter <|.. AbstractPropagatorConverter

src/site/markdown/architecture/propagation.md

@@ -280,6 +280,19 @@ If users need a more definitive initialization of an Eckstein-Hechler propagator
 should consider using a propagator converter to initialize their Eckstein-Hechler
 propagator using a complete sample instead of just a single initial orbit.
 
+### Brouwer-Lyddane propagation
+
+At the opposite of the Eckstein-Hechler model, the Brouwer-Lyddane model is
+suited for elliptical orbits. In other words, there is no problem having a small
+(or big) eccentricity or inclination. Lyddane helped to solve this issue with
+the Brouwer model by summing the long and short periodic variations of the mean anomaly
+with the ones of the argument of perigee. One needs still to be careful with
+eccentricities lower than 5e-4. For singularity reason, the computation should not
+be done for the critical inclination : 63.4°.
+
+The Brouwer-Lyddane model considers J2 to J5 potential zonal coefficients, and uses the
+mean and short periodic variation of the keplerian elements to compute the position.
+
 ### GNSS propagation
 
 There are several dedicated models used for GNSS constellations propagation. These

src/site/markdown/index.md

@@ -77,6 +77,7 @@
     * analytical propagation models
         * Kepler
         * Eckstein-Heschler
+        * Brouwer-Lyddane
         * SDP4/SGP4 with 2006 corrections
         * GNSS: GPS, QZSS, Galileo, GLONASS, Beidou, IRNSS and SBAS
     * numerical propagators