The best way to approach this sort of problem, of course, is with a free body diagram of the arm curling the weight. For a biceps curl, your problem is to solve for the arm angle that produces the greatest force on the bicep. This may seem like an easy thing to do. It is not.
This type of problem is tricky since you the more realistic you make it (a curl also depends on the brachioradialis and other muscles, not just the bicep; is close to but not strictly in one plane; the elbow may move slightly during the movement), the less likely you are to have enough data to solve the equations.
How can you set up these equations? Some clues can be found here. This is how to draw a free-body diagram, then set up equations, for an arm holding a weight using a straight arm, which is not moving.
http://www.engin.umich.edu/class/bme456/muscleforce/static1.htm
This site covers similar ground.
http://64.233.167.104/search?q=cache:a1KhXBgAJesJ:www.usc.edu/dept/LAS/kinesiology/exsc408l/lab/408_Lab7_JointKin_F04.doc+"free+body+diagram"+biceps&hl=en
In practice, this problem should first be approached by drawing a free-body diagram of a bent arm (at elbow angle theta) holding a weight, assuming motion is in one plane, minimizing the number of muscles involved, etc. These assumptions are not always trivial – but without them there is usually insufficent data to go ahead (see http://www.jneurosci.org/cgi/content/full/17/6/2128 for an example of a more complex approach).
Muscle attachment points can be used to determine the moment arms, as shown in the first site. The weight of the arm and forearm can be estimated from anatomic data, if desired, and these forces (mg) applied at their centres of mass. Rotation occurs around the elbow with an angular acceleration alpha, which is zero if the weight is lifted with a constant angular velocity. The sum of the moments around the centre of mass of the system is Icm*alpha, where Icm is the moment of inertia of the system. While there are papers that estimate I for the elbow joint (q.v. Kearney et al., J. of Biomechanics), things are made much easier by assuming that the angular acceleration about the centre of mass of the system is zero. In practice, the position of the system centre of mass is NOT constant, nor is the angular acceleration about the centre of mass is not zero, nor is the position of the system centre of mass the same as the point of rotation around the elbow.
“Hold on a freakin’ minute”, you say. I’m not trying to build an artificial elbow. I just want to do a rough-and-ready calculation. Back to our static diagram. We draw a bent arm at angle theta, similar to the “line diagram” in the first website cited, except the forces are functions of theta. If we ignore a couple muscles, we can find the biceps force as a function of theta and see where it is maximized.