Projects — Miguel Cuaycong

Passive Temperature Compensation Device: Senior Design (UC Davis), Sandia Labs

I was part of a senior design team that was tasked to conduct a trade study for Sandia National Laboratories. We investigated thermomechanical solutions to regulate fluid flow. We produced a computational prototype for a Passive Temperature Compensation Device (PTCD): wherein a bimetallic actuator drives a diamond-shaped valve orifice. My contributions centered on the computational side — building and running the CFD, and multi-body simulations used to verify the design — and on early architecture work, including the hexagonal-diaphragm orifice concept which was one of the team's first design iterations. The members of the team are: Sam van der Veen, Jon Pizarro, Nicholas Metta, Isaiah Mathew, John Bodine, and myself.

Design Evolution

Hexagonal diaphragm concept — Initial concept: A hexagonal diaphragm which could transfer a linear translational motion to a linear change in orifice area. Abandoned for having too many moving parts and error sources to hit the required micron-scale tolerance.

Rack-and-pinion actuator interface — Second iteration: a double rack-and-pinion actuator interface, chosen to let the actuator sit on its side and save space — later dropped for adding mechanical complexity without improving precision.

First iteration of the direct mount — Third iteration: an A-frame direct mount replacing the geared interface — simple in principle, but difficult to machine.

Final PTCD architecture — Final architecture: a compact cylindrical ALLVAR mount paired with the diamond-overlap orifice, meeting the linearity and precision requirements with a single moving part.

Computational Verification

CFD pressure map of the orifice, used to recover a boundary load for the multi-body simulation.

Multi-body displacement simulation verifying weld and contact-surface behavior under a rigid orifice translation.

Volumetric strain field from the coupled thermal-structural simulation.

von Mises stress field from the coupled thermal-structural simulation.

Poster

Read the 5 page overview (PDF)

nanoHUB: Topology Enumeration for Phase-Transforming Cellular Materials

This tool exhaustively enumerates every compatible arrangement on a grid based on a set of compatibility rules. Then, the enumerated data is handed back in a text file, along with a rendering of any arrangement you want to look at. The idea is that the data can be used to train a neural network that will eventually predict good topologies directly. The tool is published on nanoHUB and runs in a browser, so a researcher can use it without installing anything. This tool was developed by Sam Benemerito and myself.

The Material

A phase-transforming cellular material (PXCM) gets its properties from geometry rather than composition. In this case, a sheet is tiled with small curved cells that snap between stable shapes, and the arrangement of those elements affects properties such as mechanical energy dissipation. Four distinct elements, but they only fit together certain ways.

Hierarchy from a single cell to the full lattice — A visual of a hierarchical material.

The four constituent building blocks — Four constituent blocks. Every topology the tool produces is some arrangement of these.

Representing a Topology

A rendered topology beside its integer matrix — The picture and the grid of numbers are the same object. Reducing a topology to a matrix of block IDs is what makes both the enumeration and the downstream machine learning tractable.

Full curved lattice render — A complete lattice. Continuity as a constraint can be easily seen in this figure. Adjacent blocks have to physically join, so most matrices you could write down describe something that cannot be built.

How It Works

Three separate runtimes cooperate, with no single entry point: a shell layer that tells the nanoHUB platform how to start a session, a Python notebook which is the interface, and a compiled MATLAB core which contains the algorithm that generates the training data.

End-to-end control flow diagram — End-to-end control flow, from session launch through to the rendered image the user sees.

Enumeration kernel diagram — The enumeration itself: build the candidate rows once, then walk down the grid one row at a time, keeping only what still joins up.

Call graph and file contracts diagram — Call graph. The two halves talk through files as much as through arguments.

Click any diagram to open it full size.

Poster

My nanoHUB profile — The tool on nanoHUB

Manufacturing

A collection of parts manufactured using a combination of manual and CNC machining. Each one started as a drawing and a stock blank, and the work was as much about reading the print and planning the operation order as it was about running the machines.

Gyroscope

Totem

Totem, opposite end — The opposite end, showing the stepped slot cut in the second setup.

View the full totem drawing (PDF)

Engineering & Robotics Club

Robot chassis with drive wheel installed — A 3D-printed chassis for the club's competition robot. This is the bottom side of the chassis.

Robot chassis opened up — The same chassis flipped right side up.

Tita Neng's Food Manufacturing, Philippines

Tita Neng's is a start-up company whose mission is to produce the best quality peanuts. It is intended to be a modest outlet for us to learn how to run a company and develop leadership skills. It also functions as a test bed for implementing general manufacturing concepts. Tita Neng's is co-founded by Leo Estrada and myself, whose backgrounds are in accounting and engineering respectively.

In place of an ERP (Enterprise Resource Planning), the company relies on independent software written in-house that handles payroll, logistics, and production.

The photo above is one such software written in JavaScript named the "Inventory system". It can calculate an x amount of finished goods produced by x amount of raw materials. It takes into account projected waste and has a built in "material conservation" functionality which makes cost calculations easier.

To maintain consistency, the company only produces two products, roasted peanut and the candied variety seen in this photo.

In full swing, the company can produce a minimum of 6 kg of finished goods a day.

The raw peanuts go through a labor-intensive filtering process to get rid of "bad" peanuts and ensure consistent size and quality.

Currently, the company is preparing for a big push around late 2026 to early 2027. This is a soft reset which will launch the company with a stronger infrastructure consolidating everything we learned so far.

CAD

Collection of CAD designs

Halverail(WIP)

WIP