Simple approximation for Lambert W function for x >= 2, should only be used for large input values as error decreases for increasing input values (Batzelis, 2016)
function Wsimple input Real x(min = 2); output Real W; end Wsimple;