Traditional sensor nodes are powered by batteries. The limited battery capacity, however, constrains the lifetime of the wireless sensor networks. Wireless power transfer technology allows energy transfers from a charger to sensor nodes via wireless, and thus solves the problem completely. One fundamental issue in wireless rechargeable sensor networks is the wireless charger placement problem, i.e., how to effectively place the chargers to maximize the overall charging utility of the network. Existing works mainly focus on the deployment issues of omnidirectional chargers, which are confined to positions such as the end point of triangles or lattice point in a grid. These works inevitably have their limitations. This study is to consider the general placement problem in which the charging area of chargers is a sector and the charger can be deployed at any position in the field with arbitrary orientation. First, a charging model for directional chargers is constructed based on trace data. Then, a series of novel techniques is proposed to transform the problem to develop an effective algorithm, CDG (charger deployment-greedy), with approximation ratio (1-1/e)/(1+e) to solve this problem. The simulation results demonstrate the effectiveness of the CDG algorithm. Compared with other two random algorithms, the CDG algorithm has performance gains of nearly 300% and 100%, respectively.