I believe the answer should be one bridge.
By the time it comes to building a fully functional bridge in itself, the engineer should have studied all the possible options, variations, inconsistencies and problems that may arise during the entire construction process. The entire project should have been reproduced by computational and small scaled physical models at the very least.
Building a bridge requires lots of work, time, money and materials, and therefore, the construction of a bridge cannot be considered an “attempt” nor “trial” since its possible failure is not one of the goals.