singhn9 commited on
Commit
40b3e1c
·
verified ·
1 Parent(s): c036d1d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +51 -25
README.md CHANGED
@@ -14,7 +14,7 @@ persistent_storage: 1Gi
14
  ---
15
 
16
 
17
- # 🏭 SteelAI Module #2 — Blast Furnace & EAF Data Intelligence Explorer
18
 
19
  **Part of the TenderMatcher.Tech AI & Digital Intelligence for Metallurgy Suite**
20
 
@@ -23,7 +23,7 @@ Ready-to-deploy **Streamlit + SHAP application** for **energy and yield optimiza
23
 
24
  ---
25
 
26
- ## 🎯 Objective
27
 
28
  Predict and optimize key furnace variables such as:
29
 
@@ -36,30 +36,30 @@ This module simulates a **complete furnace data intelligence environment** with
36
 
37
  ---
38
 
39
- ## ⚙️ Core Features
40
 
41
- - 🧠 Synthetic EAF dataset generator (3,000+ records)
42
- - 🧮 Derived physical proxies:
43
  - `carbon_proxy`, `oxygen_utilization`, `slag_foaming_index`
44
- - 🔍 PCA and clustering for operating modes
45
- - ⚙️ Ensemble regression (Linear, RF, GB, ExtraTrees)
46
- - 💡 SHAP explainability for model transparency
47
- - 📊 Business framing & annotated bibliography for metallurgy ML
48
- - 🧾 Fully local synthetic data generation (no external upload needed)
49
 
50
  ---
51
 
52
- ## 🧮 Use Case Alignment (SteelAI Framework)
53
 
54
  | # | Use Case | Alignment | Description |
55
  |---|-----------|------------|--------------|
56
- | **2** | Blast Furnace / EAF Data Intelligence |**Primary (100%)** | Furnace temperature, gas chemistry, and power-density modeling for yield & energy optimization |
57
- | 7 | AI-Driven Alloy Design Tool | ⚙️ Partial | Shares compositional features (`chemical_C`, `chemical_Mn`, etc.) |
58
- | 8 | Predictive Maintenance Framework | ⚙️ Partial | Includes rolling, lag, and vibration signals for maintenance AI |
59
 
60
  ---
61
 
62
- ## 💡 Example Targets
63
 
64
  - Predict **`furnace_temp`** from operational data
65
  - Analyze **feature importance** with SHAP plots
@@ -69,7 +69,7 @@ This module simulates a **complete furnace data intelligence environment** with
69
 
70
  ---
71
 
72
- ## 🧰 How to Run Locally
73
 
74
  ```bash
75
  pip install -r requirements.txt
@@ -80,7 +80,7 @@ Then open the local URL shown in your terminal (typically http://localhost:8501)
80
 
81
  ---
82
 
83
- ## 🌐 Deploy on Hugging Face
84
 
85
  1. Create a Space → choose **SDK: Streamlit**
86
  2. Upload:
@@ -97,7 +97,7 @@ https://huggingface.co/spaces/singhn9/SteelAI_Module2_EAF_Intelligence_Explorer
97
 
98
  ---
99
 
100
- ## 🏭 Business Value Snapshot
101
 
102
  | Dimension | Improvement | Impact |
103
  |------------|-------------|---------|
@@ -109,7 +109,7 @@ https://huggingface.co/spaces/singhn9/SteelAI_Module2_EAF_Intelligence_Explorer
109
 
110
  ---
111
 
112
- ## 🤖 Generative AI & Industrial Innovation
113
 
114
  Beyond process modeling, **TenderMatcher.Tech** extends AI into Generative domains:
115
 
@@ -119,7 +119,7 @@ Beyond process modeling, **TenderMatcher.Tech** extends AI into Generative domai
119
 
120
  ---
121
 
122
- ## 🔗 AI–ServiceNow Integration Mapping
123
 
124
  | # | AI Module | ServiceNow Integration | Business Value |
125
  |---|------------|------------------------|----------------|
@@ -134,19 +134,19 @@ Beyond process modeling, **TenderMatcher.Tech** extends AI into Generative domai
134
 
135
  ---
136
 
137
- ## 👨‍🔬 About Naval Singh
138
 
139
  **Naval Singh** — Digital Transformation Advisor
140
  Specializing in AI, analytics, and industrial systems.
141
  Focused on production-grade decision systems for metallurgy, mining & manufacturing.
142
 
143
- 📧 **singhn9@gmail.com**
144
- 🔗 [LinkedIn](https://linkedin.com/in/navalsingh9)
145
- 📍 **Rourkela, India**
146
 
147
  ---
148
 
149
- ## 🌍 Why Rourkela Matters
150
 
151
  Rourkela — India’s steel and metallurgy hub — provides the perfect ecosystem for AI pilot collaborations among academia, consulting, and industry.
152
 
@@ -155,3 +155,29 @@ Rourkela — India’s steel and metallurgy hub — provides the perfect ecosyst
155
  © **TenderMatcher.Tech** — *AI & Digital Intelligence for Metallurgy*
156
  **[Read More](https://tendermatcher.tech/ai-metallurgy)**
157
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14
  ---
15
 
16
 
17
+ # SteelAI Module #2 — Blast Furnace & EAF Data Intelligence Explorer
18
 
19
  **Part of the TenderMatcher.Tech AI & Digital Intelligence for Metallurgy Suite**
20
 
 
23
 
24
  ---
25
 
26
+ ## Objective
27
 
28
  Predict and optimize key furnace variables such as:
29
 
 
36
 
37
  ---
38
 
39
+ ## Core Features
40
 
41
+ - Synthetic EAF dataset generator (3,000+ records)
42
+ - Derived physical proxies:
43
  - `carbon_proxy`, `oxygen_utilization`, `slag_foaming_index`
44
+ - PCA and clustering for operating modes
45
+ - Ensemble regression (Linear, RF, GB, ExtraTrees)
46
+ - SHAP explainability for model transparency
47
+ - Business framing & annotated bibliography for metallurgy ML
48
+ - Fully local synthetic data generation (no external upload needed)
49
 
50
  ---
51
 
52
+ ## Use Case Alignment (SteelAI Framework)
53
 
54
  | # | Use Case | Alignment | Description |
55
  |---|-----------|------------|--------------|
56
+ | **2** | Blast Furnace / EAF Data Intelligence | **Primary (100%)** | Furnace temperature, gas chemistry, and power-density modeling for yield & energy optimization |
57
+ | 7 | AI-Driven Alloy Design Tool | Partial | Shares compositional features (`chemical_C`, `chemical_Mn`, etc.) |
58
+ | 8 | Predictive Maintenance Framework | Partial | Includes rolling, lag, and vibration signals for maintenance AI |
59
 
60
  ---
61
 
62
+ ## Example Targets
63
 
64
  - Predict **`furnace_temp`** from operational data
65
  - Analyze **feature importance** with SHAP plots
 
69
 
70
  ---
71
 
72
+ ## How to Run Locally
73
 
74
  ```bash
75
  pip install -r requirements.txt
 
80
 
81
  ---
82
 
83
+ ## Deploy on Hugging Face
84
 
85
  1. Create a Space → choose **SDK: Streamlit**
86
  2. Upload:
 
97
 
98
  ---
99
 
100
+ ## Business Value Snapshot
101
 
102
  | Dimension | Improvement | Impact |
103
  |------------|-------------|---------|
 
109
 
110
  ---
111
 
112
+ ## Generative AI & Industrial Innovation
113
 
114
  Beyond process modeling, **TenderMatcher.Tech** extends AI into Generative domains:
115
 
 
119
 
120
  ---
121
 
122
+ ## AI–ServiceNow Integration Mapping
123
 
124
  | # | AI Module | ServiceNow Integration | Business Value |
125
  |---|------------|------------------------|----------------|
 
134
 
135
  ---
136
 
137
+ ## About Naval Singh
138
 
139
  **Naval Singh** — Digital Transformation Advisor
140
  Specializing in AI, analytics, and industrial systems.
141
  Focused on production-grade decision systems for metallurgy, mining & manufacturing.
142
 
143
+ **singhn9@gmail.com**
144
+ [LinkedIn](https://linkedin.com/in/navalsingh9)
145
+ **Rourkela, India**
146
 
147
  ---
148
 
149
+ ## Why Rourkela Matters
150
 
151
  Rourkela — India’s steel and metallurgy hub — provides the perfect ecosystem for AI pilot collaborations among academia, consulting, and industry.
152
 
 
155
  © **TenderMatcher.Tech** — *AI & Digital Intelligence for Metallurgy*
156
  **[Read More](https://tendermatcher.tech/ai-metallurgy)**
157
 
158
+ # SteelAI — EAF Intelligence Explorer (MODEX)
159
+
160
+ An industrial-grade AI demonstrator for Electric Arc Furnace (EAF) analytics at
161
+ **Steel Authority of India Limited (MODEX)** — combining synthetic metallurgical datasets,
162
+ automated feature engineering, and AutoML ensembles with SHAP explainability.
163
+
164
+ ### Key Highlights
165
+ - Generates full synthetic EAF datasets (~3000 rows × 200+ features)
166
+ - Supports ensemble AutoML across RandomForest, XGBoost, LightGBM, CatBoost, etc.
167
+ - Performs Optuna-based hyperparameter tuning per family
168
+ - Uses meta-stacking (Ridge) and SHAP explainability
169
+ - Includes “Recommended Target Variables” and Business Impact framing
170
+ - Features annotated bibliography with direct research paper links
171
+
172
+ ### Logging & Reproducibility
173
+ - All generated CSVs, JSONs, and logs are stored under `./logs/`
174
+ - Each session in Hugging Face ephemeral environment appends new timestamps
175
+ - Users can download artifacts directly via the **Download Saved Files** tab
176
+
177
+ *Ephemeral note*: Data and models are cleared when the Space rebuilds,
178
+ but logs persist for the current runtime session.
179
+
180
+ ---
181
+
182
+ ### Credits
183
+ Developed as part of SteelAI MODEX initiative for AI-driven metallurgy R&D.