Thank your reply dear Mr. Achieng

First one question, are you talking about suspended sediment or bed load or both?

For bed load:

In order to evaluate the accuracy of sediment transport results comming from HEC-RAS (and actually for any hydraulic software for engineering software) you need to consider the "hydraulic part" and the "sediment part" separetely.

If you are performing a 2D modelling the main issue is the depth averaged approach followed by 2D modells. Trying to keep it simple and practical, this means that the shear stress is actually an average for the whole grid cell. First approximation.

Usually and in order to estimate the shear stress at the boundary (where it is relevant for bed load transport), the "law of the wall" equation is "fed" with the average shear stress and proper height above the bed to estimate shear stress being applied to the sediment. Second approximation.

The more regular your channel, the better this two approximations will perform.

Now to the bed load transport equations. The first thing one needs to be clear about bed load transport equations is that they represent "transport capacity equations" and this  means, that they actually predict sediment transport under unlimited sediment supply conditions (partly because they were developed under such conditions). Such conditions are only found on sand bed rivers, wich are usually large alluvial rivers. For most of the other rivers, bed load is considered to be "supply limited", this means that the actual ammount of transported sediment maybe less than what the river can transport. Third approximation.

Bed load transport equations have been for the most part empirically developed. There is a large number of publications comparing bed load predictions with bed load measurements and in most cases the accuracy of the equations is rated  between 50 to 200 % the measured values. In most cases differences are attributed to not considering sediment supply conditions. Fourth approximation.

Since, and as far as I know, HEC-RAS performs a continous simulation on a dynamic river bed (changes in river bed configuration after each time step are considered), implicitely accounting for armoring (I do not know about hiding or bed form effects)  then the issues treated above are to some extent dealt with, though not completely.

Summarizing. HEC-RAS will not provide results with a much higher accuracy than the one "intrinsically contained" in these approaches and will likely fall also within the same range 50% - 200%.

One final word of caution regarding parametrization. There are some parameters regarding bed load, like the active layer depth, wich are of a conceptual nature, thus making it difficult to set "the correct value"  other than by trial an error, more or less  like it is done with the Gauckler-Manning-Strickler parameter (or simply Manning coefficient) for the hydraulics. And finally take into account that such modelling approach, whether with HEC-RAS or any other software (MIKE, TELEMAC) is highly parametrized and likely to suffer from overparametrization.

In following answers I will attach some papers that might be usefull.

here some papers

and an abstract of my master thesis where I indirectly worked with the subject. If you would like the whole thesis, let me know.

 Dear Dr. Anleu

Thank you so much for your detailed responses. 

Dear Dr. Farzin I also need to know about that in my job