You’ve spent hours mixing test batches, only to have a gorgeous satin glaze craze on a new clay body or turn glossy when you wanted matte. Glaze development often feels like guesswork—until you apply systematic, AI-assisted iteration.
The Principle: Systematic Line Blends
The most reliable path to a new glaze is starting with a known base recipe and making controlled, single-variable changes. Instead of randomly tweaking five ingredients, you hold everything constant except one material—like a flux. By creating a line blend (e.g., Base, Base +1% new flux, +2%, +3%), you isolate the effect of that variable. AI removes the tedious math: it calculates oxide percentages, molar ratios, and predicts thermal expansion so you can focus on formulation, not spreadsheets.
One tool that embodies this is the AI Recipe Analyzer. You input your most trusted recipe (e.g., a satin glaze at 60% reflectance), along with constraints: “Must be food-safe,” “Avoid barium carbonate,” or “Match thermal expansion of Cone 6 porcelain.” The analyzer uses that base as a chemical profile, then generates a variant matrix where only one proportion changes per row—just like Columns A–D in a controlled test.
Mini-Scenario in Action
Your reliable satin glaze crazes on a new porcelain body. You load the base recipe into the AI Recipe Analyzer, set the target expansion coefficient to match the clay, and mark “satin surface (60% reflectance).” The AI outputs four recipes: Base, +1% flux, +2%, +3%—each with predicted thermal expansion values. You fire a line-blend test (including a control tile) and log ramp speed, top temperature, and hold time. One variant eliminates crazing without changing surface texture.
Three High-Level Implementation Steps
Document your base and constraints.
Write down the precise recipe (in weight or molar %) and define your functional requirements: food safety, clay body fit, surface finish (glossy, satin, matte), and material restrictions (e.g., no expensive silver, no raw barium carbonate). This is the foundation the AI will use to calculate safe variant ranges.Define the variable and increment.
Choose exactly one material to adjust—a flux, a stabilizer, or a colorant. Decide the step size (e.g., 0.5% or 1% increments). The AI will then compute a line blend matrix with each recipe’s full oxide breakdown, ensuring you never accidentally exceed a critical limit (like too much boron).Log and compare systematically.
When you fire the test, include a control tile from the original recipe. Record all kiln variables (ramp rate, top temp, hold duration, kiln location). Use the AI to track consistency across batches: it compares logged data against the predicted outcome, flagging deviations caused by firing drift or material variation.
Key Takeaways
- Systematic line blends (varying one material at a time) eliminate guesswork and produce reproducible results.
- AI handles the complex chemistry calculations, letting you focus on design intent and constraints.
- Consistent logging of firing variables and control tiles ensures your data is actionable, not anecdotal.
By pairing deliberate iteration with intelligence, you turn glaze development from a lottery into a repeatable, data-driven process.
Top comments (0)