Give all pairs of people a label. There are $N=\binom{100}{2}$ labels. Call them $1,2,3,\dots, N$. Let $X_i$ be the number of times pair $i$ repeat. The meaning of $X_i$ depends on how one counts multiple repeats by two particular individuals.
Then the total number $Y$ of repeats is given by $Y=X_1+X_2+\cdots+X_N$. By the linearity of expectation, we have
$$E(Y)=E(X_1)+E(X_2)+ \cdots +E(X_N)=\binom{100}{2} E(X_1).$$
Calculating $E(X_1)$ is an easy problem, one we know the meaning of $X_1$. For let the two people in Pair $1$ be called $A$ and $B$. The probability $B$ is paired with $A$ on any particular project is $\frac{1}{99}$.
Added: OP has given the interpretation that being together $3$ times is $2$ pairs. Presumably that means being together $4$ times is $3$ pairs, and so on.
The probability that A and B are together exactly twice is $\binom{5}{2}p^2(1-p)^3$ where $p=1/99$. This makes a contribution of $\binom{5}{2}p^2(1-p)^3$ to the expectation of $X_1$.
The probability that A and B are together exactly three times is $\binom{5}{3}p^3(1-p)^2$. Under OP's interpretation, this makes a contribution of $2\binom{5}{3}p^3(1-p)^2$ to the expectation of $X_1$.
Deal similarly with the cases A and B are together four times, five times, and write down the contributions to the expectation of $X_1$.
Finally, add up to find $E(X_1)$, and multiply by $\binom{100}{2}$.