Mechanical-design-focused engineer with hands-on experience in static FEA, topology optimisation, ROS2 navigation, and on-site MEP commissioning on the Ontario Line transit project. On track for P.Eng licensure with PEO (Oct 2026).
I'm Abdulrahman Almansy, a BEng (Hons) Mechatronics Engineering graduand at Asia Pacific University (APU) (expected Oct 2026), based in Ontario, Canada. EIT-eligible, on track for P.Eng licensure with PEO upon graduation.
My focus sits at the mechanical end of mechatronics — static FEA, topology optimisation, CAD/CAM, and vibration-aware dynamic analysis — backed by real commissioning experience on the Ontario Line transit project (Connect 6ix, Toronto). I treat every system as an integration problem: structure, motion, control, and software must all speak the same language.
From mesh-convergence studies in SolidWorks Simulation to tuning Nav2 cost-maps in ROS2 and commissioning PLC interlock logic, I apply a failure-first mindset — FMEA and RCA before the first weld, not after the first incident.
Reduced bracket mass by 28% while maintaining FoS > 2.5 via topology optimisation on a 6061-T6 aluminium component under 5 kN loading. Validated with mesh-convergence study (<5% von Mises stress variance). Generated 5-operation CAM toolpaths, cutting machining time by 15%.
Designed a custom 10 m × 8 m warehouse SDF world in Gazebo Harmonic and built the full navigation stack from scratch — SLAM Toolbox → AMCL localisation (500–2000 particles) → NavFn/Dijkstra global planner → DWB controller. Built a Python/tkinter operator GUI with live occupancy-grid map and real-time /odom robot position. Resolved 5 critical failure modes via RCA — including ROS2 Jazzy's TwistStamped cmd_vel type change, AMCL world-to-map frame offset, and Nav2 goal rejection in occupied costmap cells — achieving 92% waypoint success.
Programmed an 8-rung Omron PLC ladder-logic sequence driving double-acting pneumatic cylinders (solenoid valves), relay unit, and a magnetic pick-and-place actuator. Configured totalising timers (TTIM 001 = 20, TTIM 003 = 30, TTIM 005 = 20 units) for sub-cycle synchronisation. Applied FMEA on ladder rungs to eliminate 3 sequence race conditions, achieving ±50 ms timing consistency across 100+ cycle dry-runs with E-stop and interlock logic per safety-circuit standards.
Reduced sensor-to-actuator reaction latency by ~20% across 50 bench-test runs by tuning ultrasonic threshold logic, DC motor PWM response timing, and applying low-pass filtering on range measurements — mirroring real-time control challenges in industrial AGV platforms.
Built an end-to-end rehabilitation platform: an ESP32 glove with 5 flex sensors + MPU6050 IMU detects 5 hand gestures (2 s hold debounce, instant emergency stop) via calibrated per-finger thresholds, driving a 60-LED WS2812B NeoPixel strip. PyQt5 dashboard provides live elbow-angle plotting, circular rep counter, and CSV session logging. Voice control via ElevenLabs TTS + Google SR with fuzzy matching selects from 3 rehabilitation intensity plans with mid-session resume.
Designed a 7-phase explainable fault diagnosis framework for photovoltaic systems, combining XGBoost with Tree SHAP to attribute each prediction to the 5 monitored parameters (I, V, P, G, T) across 4 fault classes — normal, open-circuit, short-circuit, and partial shading. Benchmarks XGBoost against RF, SVM, and CNN with both Tree SHAP & Kernel SHAP — the first cross-architecture SHAP comparison in a unified PV fault pipeline. Targets macro F1 > 0.85. Aligned with UN-SDGs 7, 9 & 13.
Programmed an ABB IRB 120 industrial manipulator in RAPID to pick and place 20 coloured pegs across two pallets, forming target letter patterns. Sequenced MoveL commands with calibrated Offs() 3D offsets (±25 mm to ±150 mm), dual velocity profiles (v1000 / v500), and DO16_GRIPPER_ON Set/Reset for pneumatic gripper control. Diagnosed positioning misalignment and gripper timing errors through iterative coordinate tuning.
Designed and simulated a complete regenerative braking system for a 1500 kg EV in MATLAB/Simulink. The DC motor model (0.8 Nm/A torque constant, 8:1 gear ratio) operates as a generator during a 6-second braking event, decelerating the vehicle from 80 km/h to 15 km/h. Peak negative motor torque of −200 Nm drives 212.5 A into a 400 V battery pack, recovering 115 Wh at 82.7 kW peak. Overall system efficiency held at 85%, with regeneration approaching 100% efficiency during steady deceleration — equivalent to ~25% city range extension.
Whether you have a project in mind, want to discuss engineering challenges, or are looking for a driven mechatronics intern — I'd love to hear from you. Let's build something remarkable.
Thanks for reaching out — I'll get back to you soon.