This paper applies Co-flow Jet (CFJ) active flow control (AFC) on NASA hump and numerically studies its flow control performance to suppress flow separation at low energy expenditure. The effects of the CFJ injection and suction locations are studied. The experimental results of the baseline hump and the hump with steady blowing and suction are used to validate the accuracy of numerical method. The high fidelity in-house CFD code FASIP with the 2D Unsteady Reynolds averaged Navier-Stokes (URANS) equations with one-equation Spalart-Allmaras model is utilized. The validation is overall in very good agreement with the experiment for the baseline and steady blowing cases. The numerical simulation indicates that using co-flow jet for the hump separation control is very effective and energy efficient. With injection location at 50%C and suction location at 70%C, a full flow attachment is achieved at Cµ=0.0077 with the CFJ power coefficient (Pc) of 0.0032 and the energy coefficient (CE ) of 0.0034. This is consistent with the previous observation for CFJ airfoil that placing the injection at the suction peak location minimizes the AFC energy cost. The present study also discovers that the optimum location for CFJ suction is at the location the hump surface slope reaches the minimum, which gives the lowest energy consumption. A study is also conducted to move the CFJ injection location downstream to the baseline flow separation onset position at 67%C and move the suction location into the adverse pressure gradient area. Such configuration significantly increases the energy consumption. The study indicates that applying the CFJ upstream of the flow separation with mixing under favorable pressure gradient is much more efficient than applying it at the baseline flow separation area under the adverse pressure gradient. To further investigate the role of suction in the CFJ active flow control, the injection-only flow control is designed and numerically studied. Compared with the CFJ case, the minimum energy consumption for the injection only case with the separation removed is increased by 57% with CE of 0.0055. For the injection only flow control, the injection location for the minimum energy consumption is not at the suction peak location of 50%C as for the CFJ, but at the onset of separation location of 67%C. These results indicate that the suction of the CFJ flow control is very beneficial for two reasons: 1) It energizes the boundary layer and makes the CFJ energy consumption significantly lower than the injection only flow control; 2) It provides the source of the mass flow for injection to make the CFJ self-contained zero-next-mass-flux flow control to avoid introducing mass flow from other source, which will incur extra energy consumption.