Computational fluids dynamics (CFD) and response surface methodology (RSM) are used to optimize grooved micromixers with obstructions. The starting geometry is inspired by the work of Wang and co-workers (Wang et al., 2012) which consists of a Y-type micromixer with cylindrical grooves adjoining to the main straight channel. Obstructions are added to the grooved micromixers. To observe influences of obstruction shape, simulations are carried out for three types of geometric shapes: circular, square and diamond. The effects of obstruction dimension (OD) and offset (OF) in the range of 60-120 mu m and 0-30 mu m, respectively, are investigated on the percentage of mixing (phi), pressure drop (Delta p) and mixing energy cost (mec). The numerical experiments are designed based on a five-level CCD to develop statistical models for the optimization procedure. Results show that phi monotonically increases with the increase of OD in all values of OF for all obstruction shapes. Same trend is observed with increase in the offset. Moreover, results reveale that Delta p and mec increase with the increase of OD at almost all offset values for all shapes of obstructions. Conversely, the pressure drop and the mixing energy cost reduce as OF increases. Finally, phi, Delta p and mec are considered as the objective functions to numerically optimize the performance of grooved micromixers with obstructions. Overall desirability values of 0.796, 0.834 and 0.760 are found for micromixers with circular, diamond-shaped and square obstructions, respectively. For circular obstructions, the obtained optimal value of dimension is 110.75 mu m, while the respective value for the diamond-shaped obstructions is 116.98 mu m and for the square obstructions is 116.88. In addition, the constant value of 30 mu m is predicted for the optimum offset for all obstructed micromixers. (C) 2019 Elsevier Ltd. All rights reserved.