Dear Ms, 

Already lot ofnon-linear regression models are proposed and validated by researchers. Hence it is good to modify any existing models by incorporating influencing parameters. Experimental validation is important. 

Dear Md, 

Please find the attachments. First paper explains about the how shear strength model for shear critical beams works. Second paper explains the influencing parameters of shear critical beams. Few real life structural failures and detailed information about laboratory shear tests are detailed in both the papers.

Conference Paper Failure Modes of RC Deep Beam Panels

Article SHEAR STRENGTH OF RC DEEP BEAM PANELS – A REVIEW

Thanks both of you for your kind response. It is possible to modify existing models and also propose a new non-linear regression model. I already studied the effect of some influencing parameter on the shear strength of FRP-RC beams and based on a large experimental database, trying to propose a new non-linear model for slender beams only as deep beam behave differently. I extensively validating my proposed model with other sets of experimental results and overall it is showing a good result.  Thanks Smita Gopinath and Leon Raj for your usefull information

Smitha mam,

Yes you are correct. 

Smitha mam,

Yes you are correct. 

you can sea my works on the influence of the confinement on the shear behavior of the reinforced conrete beams.

Dear Md,

All of the existing shear models are somewhat empirical, which take into account only the "most important" influencing parameters. An empirical shear model that can fit some experimental results is not difficult to be formulated. However, for a long-term study, it is suggested to try with a mechanical-sound model and step to step reduce the emprical terms in the model. You may look into the existing shear models and create yourself the whole picture of the problem. It takes time, much time. But it is fair, if you really want to have a good shear model, which can consider all the influencing parameters correctly, especially the size effect. Note that the existing databases include only a few large test members and therefore it is not enough for the statistic evaluation of any shear model.

Best regards,