There are different channel models that can be applied, one usual approach (which makes all the calculation easier) is what Marzetta has used in his seminal paper. Assuming G_{lj}  the channel matrix from all users in cell j to BS in cell l, which can be written as:

G_{lj} = [g_{lj1},...,g_{ljK}], where the N*1 vector g_{ljk} is the channel from user k in cell j to BS in cell l.

Then one can write  g_{ljk}  as follows

g_{ljk} = \beta_{ljk}^{0.5}  *  h_{ljk} 

where \beta_{ljk}  is a scalar representing the Large Scale Fading (path loss and shadowing) and h_{ljk} is an N*1 vector for small scale fading (usually assumed to be zero-mean and unit-variance).

Now note that 

\beta_{ljk} = z_{ljk} / r_{ljk}^{\gamma}

where z_{ljk} is a log-normal random variable (also in some papers it is common to consider this one as one) and r_{ljk} is the distance between user k in cell j and the BS in cell l. Note that \gamma is called the decay exponent (I would say path loss exponent).

Note that in this modelling the distance of users should not be less than a certain threshold value. For detailed information refer to "Noncooperative Cellular Wireless with Unlimited Numbers of Base Station Antennas"

Thank you Meysam. After calculating g_{ljk},  it is to estimate the user SINR in case of single cell scenario. But in case of multi cell scenario, do we estimate the inter cell interference the same way we do in conventional LTE systems? precoding will not reduce the intercell interference? and is doing power control is necessary to achieve the same SINR for each user?

First of all note that gljk is a cross channel and hence it is for cellular scenario. If we talk about single we don't need such a notation.

The second issue, for me to be able to answer your question correctly I need a more exact definition of your question. But as Far as I can understand you want to know how to calculate the SINR in muticell massive MIMO? Am I right? If not please correct me.

Also do not combine different stuff, forget about LTE and standards. You have a mltiuser MIMO system which can be single cell or multi-cell. I dont see where the LTE comes to the picture ??

uch a notation.

Yes I mean the SINR in multicell massive MIMO, I meant conventional cellular system by LTE where we estimate the interference as the sum of received power from the interfering cells.  To be more precise, I need to calculate the forward link SINR of a user K  which is located in the cell J and receiving interference from cell I (1 to 6 for example).

So you need to calculate:

1. the intended user signal power

2. the intra cell interference, which is due to multi-user scenario and imperfect CSI (if it is perfect CSI and you use for example ZF it is zero)

3.the inter-cell interference from other cells

4. Noise

now the simultaneous SINR is 1/(2+3+4).

If you want ergodic rate and as the users don't have any channel estimates you need to use Theorem 1 in Jubin Jose paper.

Thank you Meysam, my doubt  was mainly about  calculating 3 (inter-cell interference) , is it enough to subtract  (in dB)  large scale fading pathloss (Distance+Shadowing) from the cell total transmitted power?  Does number of antennas in interfering cell, or  the used precoding method have any impact on the intercell interference or not?