Development of an Artificial Neural Network to Predict Sulphide Capacities of CaO–SiO<sub>2</sub>–Al<sub>2</sub>O<sub>3</sub>–MgO Slag System
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Depletion of the high quality ores around the world has forced ferronickel producers to extract metal values from low-grade ore bodies with significant amounts of impurities. Under this condition, maintaining alloy quality is of utmost importance for the smelters; however still, accessibility of a reliable sulphide capacity model for FeNi refining processes is an issue. Many of the current models, such as those incorporating optical basicity, have proven to be erroneous and unreliable for wide ranges of composition and temperature. These models are typically developed and tested without a proper validation method thus allowing for great correlations which may not fare well with the introduction of new data. Models built from fundamental thermodynamic data perform much better in predicting sulphide capacities but are not only complicated to formulate but also too complicated to be used by operators on a day to day basis as multitude of inputs are needed. Hence, development of a reliable model based on fundamentals, which can also be directly used by plant operators is very much demanded by the industry. In the current study, an artificial neural network (ANN) approach has been used to predict sulphide capacities of slag compositions in the CaO–SiO2–Al2O3–MgO system with an objective to be used in ferronickel refining processes. The resulting models are evaluated on: 1) coefficient of multiple determination (R2), 2) correlation strength (r), 3) root mean square error (RMSE) and 4) computation speed. The ANN based model has shown to be superior in predicting sulphide capacities to current models.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.001 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.000 | 0.002 |
| Open science | 0.003 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it