Guangcan Su logo
LCA · TEA · AI-augmented assessment

Guangcan Su

Postdoctoral Researcher · Northwestern University

I build sustainability assessment workflows that connect process simulation, life cycle impacts, techno-economics, and AI to evaluate which emerging carbon and clean-energy systems can work at scale.

LCA TEA Process Systems AI-augmented Workflows
Portrait of Guangcan Su
LCA TEA Sustainability AI-augmented Workflows

I'm a postdoctoral researcher at Northwestern University working at the intersection of process simulation, sustainability assessment, and AI. My research combines life cycle assessment (LCA), techno-economic analysis (TEA), and spatial modeling to evaluate emerging technologies for carbon conversion, direct air capture, and clean energy systems — and to understand which of them actually make sense at scale.

Selected work

Carbon conversion, waste valorization, and AI-enabled assessment.

Recent publications span reactive carbon capture, electrochemical conversion, biochar systems, and data-driven tools for sustainable process design.

Open projects

Research software for LCA, TEA, and AI workflows.

These GitHub projects turn recurring sustainability-assessment problems into reusable tools: evidence extraction, harmonization, uncertainty, regionalization, dynamic LCA, scale-up decisions, and model evaluation.

Open LCA systems stack connecting extraction, harmonization, uncertainty, GIS, TEA, dynamic LCA, and AI evaluation tools
GitHub portfolio

Open LCA Systems Stack

A connected research-software ecosystem for making LCA evidence more traceable, spatially aware, uncertainty-aware, and useful for technology scale-up decisions.

LCI extraction Evidence quality Uncertainty GIS factors TEA/LCA scale-up
View GitHub profile
ScaleBridge TEA and LCA scale-up model visualization with process simulation and decision surfaces
TEA + LCA

ScaleBridge

Joint ex-ante scale-up analysis for emerging technologies, linking process assumptions, cost, GHG intensity, scenarios, carbon price, and Pareto screening.

Open GitHub
GIS-LCA geography normalization and regional life cycle assessment factor adapter visualization
Regionalized LCA

GIS-LCA

A geography normalization and regional factor adapter that makes location handling, fallback logic, and factor provenance visible in LCA workflows.

Open GitHub
LCA-Harmonizer visualization of literature evidence cleaning, unit harmonization, provenance checks, and study comparison
Evidence layer

LCA-Harmonizer

Turns messy LCA literature into cleaner, comparable, model-ready evidence with unit, functional-unit, boundary, provenance, and quality checks.

Open GitHub
LCA-Uncertainty visualization of Monte Carlo distributions, percentile ranges, and sensitivity drivers
Uncertainty

LCA-Uncertainty

Propagates uncertainty through LCA inventories using pedigree mapping, lognormal sampling, Monte Carlo simulation, percentile summaries, and sensitivity ranking.

Open GitHub
LCA-Benchmark visualization of AI extraction scoring, evidence grounding, and recommendation evaluation
AI evaluation

LCA-Benchmark

Evaluates AI-assisted LCA extraction and recommendation workflows with numeric matching, grounding checks, top-k scoring, and machine-readable reports.

Open GitHub
DynLCA real-time grid carbon and dynamic life cycle assessment dashboard visualization
Dynamic LCA

DynLCA

Connects live grid carbon intensity and time-aware data streams to LCA calculations, supporting dynamic footprints and operational timing questions.

Open GitHub
BiocharPredictionApp machine learning visualization for biomass torrefaction inputs and predicted biochar properties
ML model

BiocharPredictionApp

A Tkinter GUI using a GBM model to predict yield, heating value, elemental composition, and proximate properties from biomass torrefaction inputs.

Open GitHub
LCA-DataExtractor document AI pipeline visualization from PDFs and EPDs to structured life cycle inventory tables
Document AI

LCA-DataExtractor

An LLM-powered pipeline for parsing papers and EPD reports, retrieving relevant chunks, and extracting structured LCI flows into reusable tables.

Open GitHub