This a long story which I am trying to make it short here. However, if the following response is not satisfying, I can provide you with theories behind each every one of those parameters from appropriate references.
When you do have a turbulent flow you need to specify some turbulent parameters that represent the real conditions. In addition, they are also being used for the convergence of your numerical simulation. Turbulent Kinetic Energy (K) and Turbulent Dissipation rate (Epsilon) are from the nature of a turbulent flow. This is being well explained in distribution of turbulent energy in wave number space [turbulence spectrum] (you can refer to Landahl and Mollo-Christensen (1986)). Another provided choice in SWFS is Turbulent Length Scale (L) and Turbulent Intensity (I). (Let's not to get into definitions of them at this point.) All of K,e, L, and I can be measured in a turbulent flow, however, measuring them is so hard and an expensive experimental test. Having said that it does not mean that we should leave those values as they are and do not care about them in SWFS. In fact, we can refer to literature to find some relevant cases to have a good estimation of them to setup our CFD problem. For instance, a flow with turbulent intensity of 5-10% considered as highly turbulent flow while a turbulent intensity of about 0.1 is a low turbulent flow. As a tip for turbulent length scale for simple cases, such as a turbulent flow inside a pipe, you can consider the hydraulic diameter of the geometry as turbulent length scale, which in pipe case it would be the pipe's diameter. For internal flows a good estimation for turbulent intensity would be 0.16*(ReH)^(-1/8), wherein ReH is the Reynolds number based on hydraulic diameter. On the other hand, having a good idea about K and e is a bit harder. I believe the following article would be very helpful for you: FLUENT 6.3 User's Guide - 7.2.2 Determining Turbulence Parameters.
Hope you find this helpful. Please let me know, should you have any questions.
Thank you so much, it was really helpful.