dhani10's picture
Deploy Engine Condition Monitoring App with Docker
44a1e2b verified
raw
history blame
6.86 kB
import os
import joblib
import pandas as pd
import numpy as np
import streamlit as st
from huggingface_hub import hf_hub_download, login
# Configuration
HF_TOKEN = os.getenv("HF_TOKEN") # optional if model is public
HF_MODEL_REPO = os.getenv("HF_MODEL_REPO", "dhani10/engine-condition-model")
MODEL_FILE = os.getenv("MODEL_FILE", "best_engine_model.joblib")
# Use a writable cache (esp. on Spaces)
os.environ.setdefault("HF_HOME", "/tmp/huggingface")
os.environ.setdefault("HF_HUB_CACHE", "/tmp/huggingface/hub")
os.makedirs(os.environ["HF_HUB_CACHE"], exist_ok=True)
if HF_TOKEN:
try:
login(HF_TOKEN)
except Exception:
pass
@st.cache_resource
def load_model():
p = hf_hub_download(
repo_id=HF_MODEL_REPO,
filename=MODEL_FILE,
repo_type="model",
token=HF_TOKEN,
cache_dir=os.environ["HF_HUB_CACHE"],
)
return joblib.load(p)
model = load_model()
# Engine sensor features
ENGINE_FEATURES = [
"Engine rpm",
"Lub oil pressure",
"Fuel pressure",
"Coolant pressure",
"lub oil temp",
"Coolant temp"
]
st.set_page_config(page_title="Engine Condition Monitor", layout="centered")
st.title("🏭 Engine Condition Monitoring System")
st.caption("Enter sensor readings to predict engine condition (Normal/Faulty)")
# Add information about the model
with st.expander("ℹ️ About this Model"):
st.write('''
This model predicts engine condition based on real-time sensor readings:
- **Normal (0)**: Engine operating within normal parameters
- **Faulty (1)**: Engine showing signs of potential failure
**Typical Operating Ranges:**
- Engine RPM: 600-2500
- Lub Oil Pressure: 2.0-4.0 bar
- Fuel Pressure: 8.0-15.0 bar
- Coolant Pressure: 1.5-4.0 bar
- Lub Oil Temp: 75-110°C
- Coolant Temp: 70-100°C
''')
# Sensor input form
with st.form("engine_predict_form"):
st.subheader("🔧 Sensor Readings")
col1, col2 = st.columns(2)
with col1:
engine_rpm = st.slider(
"Engine RPM",
min_value=0,
max_value=3000,
value=1800,
help="Engine rotations per minute"
)
lub_oil_pressure = st.slider(
"Lub Oil Pressure (bar)",
min_value=0.0,
max_value=6.0,
value=3.1,
step=0.1,
help="Lubricating oil pressure in bar"
)
fuel_pressure = st.slider(
"Fuel Pressure (bar)",
min_value=0.0,
max_value=20.0,
value=12.0,
step=0.1,
help="Fuel system pressure in bar"
)
with col2:
coolant_pressure = st.slider(
"Coolant Pressure (bar)",
min_value=0.0,
max_value=5.0,
value=2.9,
step=0.1,
help="Cooling system pressure in bar"
)
lub_oil_temp = st.slider(
"Lub Oil Temp (°C)",
min_value=0,
max_value=150,
value=92,
help="Lubricating oil temperature in °C"
)
coolant_temp = st.slider(
"Coolant Temp (°C)",
min_value=0,
max_value=150,
value=89,
help="Coolant temperature in °C"
)
submitted = st.form_submit_button("🔍 Predict Engine Condition")
if submitted:
# Build input data
ui_row = {
"Engine rpm": float(engine_rpm),
"Lub oil pressure": float(lub_oil_pressure),
"Fuel pressure": float(fuel_pressure),
"Coolant pressure": float(coolant_pressure),
"lub oil temp": float(lub_oil_temp),
"Coolant temp": float(coolant_temp)
}
# Create DataFrame with expected columns
row = pd.DataFrame([ui_row])
try:
# Make prediction
prediction = model.predict(row)[0]
probability = None
if hasattr(model, "predict_proba"):
probability = model.predict_proba(row)[0]
# Display results
st.subheader("🎯 Prediction Result")
if prediction == 1:
st.error("🚨 **FAULTY CONDITION DETECTED**")
st.warning("⚠️ Engine shows signs of potential failure. Immediate maintenance recommended.")
if probability is not None:
st.metric(
"Confidence Score",
f"{probability[1]:.1%}",
delta=f"Faulty probability",
delta_color="inverse"
)
else:
st.success("✅ **NORMAL OPERATION**")
st.info("🌡️ Engine operating within normal parameters.")
if probability is not None:
st.metric(
"Confidence Score",
f"{probability[0]:.1%}",
delta=f"Normal probability",
delta_color="normal"
)
# Display detailed probabilities
if probability is not None:
col1, col2 = st.columns(2)
with col1:
st.progress(probability[0], text=f"Normal: {probability[0]:.1%}")
with col2:
st.progress(probability[1], text=f"Faulty: {probability[1]:.1%}")
# Show input values
with st.expander("📊 Sensor Readings Used"):
st.dataframe(row.T.rename(columns={0: "Value"}))
# Add maintenance recommendations for faulty conditions
if prediction == 1:
st.subheader("🔧 Recommended Actions")
issues = []
if lub_oil_pressure < 2.5:
issues.append("Low lubricating oil pressure")
if fuel_pressure > 13.0:
issues.append("High fuel pressure")
if lub_oil_temp > 105:
issues.append("High lubricating oil temperature")
if coolant_temp > 95:
issues.append("High coolant temperature")
if issues:
st.write("Potential issues detected:")
for issue in issues:
st.write(f"• {issue}")
st.write('''
**Immediate Steps:**
1. Check oil levels and quality
2. Inspect cooling system
3. Verify fuel system components
4. Consult maintenance manual
''')
except Exception as e:
st.error(f"❌ Prediction failed: {e}")
st.write("Expected features:", ENGINE_FEATURES)
# Add footer
st.markdown("---")
st.caption('''
**Engine Condition Prediction System** |
Predictive Maintenance Model |
[View Model on Hugging Face](https://huggingface.co/dhani10/engine-condition-model)
''')