Following the developments in wireless communications, the use of intelligent reflecting surfaces (IRS) to extend radar illumination to non-line-of-sight (NLoS) scenarios has spurred research interest of late. Initial works have assumed ideal propagation conditions based on the radar equation to assess the signal noise ratio (SNR) enhancement. In this paper, we consider a realistic target position estimation of an IRS-aided radar system framework. Firstly, a time-code-space (TCS) IRS array model was proposed, where each sub-unit array can work independently as a TX IRS unit or Rx IRS unit. Then, the signal model of the time division multiplexing (TDM) IRS array based on the frequency-modulated continuous waveform (FMCW) is derived. Thereafter, the developed signal model is used to emulate the radar performance utilizing a Blender-based simulator in NLoS scenarios, where the various assumptions commonplace in literature and their suitability in realistic scenarios are considered. The simulation result shows the validation of the proposed IRS-aided framework in target localization. Further, the trade-off between angle resolution and energy/ time consumption is also discussed.