Trane Technologies

Graduate R&D Engineer projects and initiatives

Master's Thesis WP1: Simulink Model of Thermo King Advancer VCC

M.E. Mechanical Engineering, University of Galway

Master's Thesis WP1: Simulink Model of Thermo King Advancer VCC preview 2

Built real-time Simulink/Simscape model of Thermo King Advancer A-500 for HiL testing:

• 19.5 kW cooling capacity at 0°C validated against bench data

• Real-time execution at 0.1s timestep on commodity hardware

• Full evaporator, trailer thermal network, and Cycle-Sentry controller

• First HiL simulator demonstrated for transport refrigeration

Simulink

Simscape

HiL

Digital Twin

VCC

Real-Time Simulation

MATLAB

Master's Thesis WP2: Arduino-based HiL Simulator for Transport Refrigeration

M.E. Mechanical Engineering, University of Galway

Master's Thesis WP2: Arduino-based HiL Simulator for Transport Refrigeration preview 2

Master's Thesis WP2: Arduino-based HiL Simulator for Transport Refrigeration preview 3

Two-Arduino HiL rig bridging Simulink VCC model with real embedded controller:

• 0.92°C RMSE temperature emulation via MCP4261 digital potentiometer

• PT1000 RTD sensor emulation with SPI control and voltage divider

• Validated On-Off control switching at −10°C setpoint

• First demonstration of HiL for transport refrigeration controls

Arduino

HiL

Embedded Systems

SPI

Control Systems

Sensor Emulation

C++

Master's Thesis WP3: Experimental Validation of Prototype HiL Simulator

M.E. Mechanical Engineering, University of Galway

Master's Thesis WP3: Experimental Validation of Prototype HiL Simulator preview 2

Master's Thesis WP3: Experimental Validation of Prototype HiL Simulator preview 3

Ground-truth validation of HiL system against real Advancer A-500:

• 6-hour bench test at Thermo King Galway with MultiDAC data acquisition

• Behavioural validation confirmed at programme level

• 4.32× modulation acceleration root-caused to Simscape assumption

• ~50% cooling overshoot attributed to controller and refrigerant differences

Experimental Validation

Data Acquisition

Root Cause Analysis

MultiDAC

Bench Testing

Physics based real time ML model - Patent Pending

Trane Technologies

Real-time autoregressive LSTM for thermo-fluid simulation in HiL environment:

• 40MB RAM footprint, stable across 150,000+ autoregressive iterations

• 95% dynamics correlation, 90% magnitude — before large-scale fine-tuning

• Passed Trane IP council first review in 4 months as graduate

• Clean ONNX export path to embedded C++ deployment

Machine Learning

LSTM

Python

TensorFlow

ONNX

Patent

Digital Twin

Physics Simulation

On-Prem Cursor

Trane Technologies

Self-hosted agentic coding platform for Fortune 500 R&D team:

• Ollama + open-source VS Code agent + Open WebUI stack

• Adopted by 2–5 engineers on departmental ML pipelines

• Cursor-equivalent functionality with zero data egress

• Triggered internal review into team-wide Claude adoption

Ollama

LLM

VS Code

Open WebUI

MLOps

Agent Orchestration

Python

TK Telematics

Trane Technologies

Led OTA telematics validation campaign as graduate, saving $2M+ in manual dealer campaigns:

• 100% failure → 100% pass rate across 20+ tests in 2 months

• Diagnosed server prioritization blocker via CAN bus log analysis

• Sole communication point between Connected Solutions and R&D Lead

• Shipped on time for flagship V1 release

OTA Updates

Telematics

CAN Bus

Testing

Project Management

HPE ALM