This work reports on modeling and numerical experience in solving the long-term design and operation planning problem of the Brazilian natural gas network. A stochastic approach to address uncertainties related to the gas demand is considered. Representing uncertainties by finitely many scenarios increases the size of the resulting optimization problem, and therefore the difficulty to solve it. We calculate the value of applying a stochastic programming framework. The numerical tractability of the problem is obtained by applying state of the art techniques of optimal scenario reduction and decomposition by bundle methods. The benefits of each approach are considered. Numerical experiments in a real-life case are assessed.