CYBERWARD

A reliability analysis was comducted to monitor performance and predict potential failures across the haul truck fleet. The study identified patterns linking increased operating hours and delayed maintenance to higher failure probability. Predictive insights helped flag high-risk trucks early, reducing unplanned downtime and improving Mean Time Between Failures (MTBF). Implementing condition-based maintenance and data-driven scheduling is expected to enhance fleet reliability and minimize costly breakdowns.

Objective:
Analyse truck performance and monitor and predict failure & haul truck maintenance issues using telemetry, operating hours, and sensor data to reduce unplanned downtime and improve fleet reliability.

Data Collection and Prepares:

• Dataset columns: Truck_ID Date Run_Time_Hours Idle_Time_Hours Breakdown_Flag Maintenance_Type Oil_Temperature (°C) Hydraulic_Pressure (bar) Vibration_Level (mm/s) Fuel_Rate (L/hr) Ambient_Temperature (°C) Predicted_Failure_Prob (%) Next_Maintenance_Due (hrs) Actual_Downtime_Hours
• Number of records: 500
• Source: Simulated dataset (python script)

<!-- wp:syntaxhighlighter/code {"language":"python"} -->
<pre class="wp-block-syntaxhighlighter-code">import pandas as pd
import numpy as np
from datetime import datetime, timedelta

# Parameters
num_records = 500
num_trucks = 10
start_date = datetime(2025, 1, 1)

# Generate random data
np.random.seed(42)
data = {
    "Truck_ID": [f"TRK{str(np.random.randint(1, num_trucks+1)).zfill(2)}" for _ in range(num_records)],
    "Date": [start_date + timedelta(days=int(i)) for i in np.random.randint(0, 120, num_records)],
    "Run_Time_Hours": np.random.uniform(3, 10, num_records).round(2),
    "Idle_Time_Hours": np.random.uniform(0.5, 3, num_records).round(2),
    "Breakdown_Flag": np.random.choice(["Yes", "No"], num_records, p=[0.1, 0.9]),
    "Maintenance_Type": np.random.choice(["None", "Oil Change", "Brake Check", "Engine Tune", "Hydraulic Repair"], num_records, p=[0.5, 0.2, 0.15, 0.1, 0.05]),
    "Oil_Temperature (°C)": np.random.uniform(70, 110, num_records).round(1),
    "Hydraulic_Pressure (bar)": np.random.uniform(150, 250, num_records).round(1),
    "Vibration_Level (mm/s)": np.random.uniform(1.5, 5.5, num_records).round(2),
    "Fuel_Rate (L/hr)": np.random.uniform(20, 60, num_records).round(2),
    "Ambient_Temperature (°C)": np.random.uniform(-10, 35, num_records).round(1),
    "Predicted_Failure_Prob (%)": np.random.uniform(1, 90, num_records).round(1),
    "Next_Maintenance_Due (hrs)": np.random.uniform(20, 500, num_records).round(1),
    "Actual_Downtime_Hours": np.random.uniform(0, 6, num_records).round(2),
}

# Create DataFrame
df = pd.DataFrame(data)

# Sort for realism
df.sort_values(by=["Truck_ID", "Date"], inplace=True)

# Save to Excel
output_path = "Truck_KPI_Dataset.xlsx"
df.to_excel(output_path, index=False)

print(f"Dataset generated successfully: {output_path}")
print(df.head())
</pre>
<!-- /wp:syntaxhighlighter/code -->

Sample of the dataset in Power BI:

Dataset columns

KPIs Calculation

Data Analysis & Visualizations

based on the dataset, we are going to represent three related projects, analyzed separately, for more on depth analysis for the datasets.

1. Truck Reliability & Failure Analysis

Project Overview

This analysis evaluates the reliability, availability, and failure behavior of a fleet of mine haul trucks using predictive failure probabilities, downtime records, and reliability engineering metrics such as MTBF (Mean Time Between Failures) and MTTR (Mean Time To Repair). The objective is to understand failure patterns, identify high-risk trucks, and optimize maintenance strategies for improved uptime and operational efficiency.

The dashboard visualizes daily failure probability trends, monthly downtime behavior, and breakdown rates per truck—providing a comprehensive view of both predictive and historical reliability performance.

Key Insights

• Avg Predicted Failure Probability: 49.75%
  • Indicates nearly half of the truck fleet is at moderate to high risk of failure within the prediction window.
  • Predictive models or maintenance thresholds should be fine-tuned to reduce this number below 35%.
• MTBF (Mean Time Between Failures): 60.09 hours
  • Trucks, on average, operate for about 60 hours between failures — showing moderate reliability.
  • Industry benchmark for heavy-duty haul trucks is usually 80–100 hours between breakdowns under optimal maintenance.
• MTTR (Mean Time to Repair): 1.69 hours
  • Maintenance crews are performing efficiently, with quick recovery times once failures occur.
  • Indicates a responsive maintenance team and good spare parts availability.

2. Monthly Downtime Analysis

• Highest downtime: April (88.4 hrs) and February (80.4 hrs)
  • Suggests seasonal strain, possibly due to cold-weather effects or delayed preventive maintenance cycles.
• Lowest downtime: March (32.4 hrs)
  • Could coincide with recent maintenance overhauls or new fleet optimization measures.
• Trend insight: Fluctuating downtime indicates inconsistent maintenance scheduling or variable truck utilization intensity.

3. Breakdown Rate by Truck

• Trucks TRK06, TRK08, TRK17, and TRK18 show breakdown rates above 25%.
  • These are outliers that need condition monitoring (oil analysis, vibration, temperature tracking).
• Trucks TRK02, TRK04, TRK20 are the most reliable with rates below 10–12%.
  • Their maintenance logs should be reviewed to identify best practices for replication.

4. Predicted Failure Probability Trend

• The failure probability line chart shows spikes mid-month (around Jan 19).
  • Suggests possible maintenance backlog or high operating stress in that period.
• Fluctuations indicate that the failure model could benefit from including operating conditions (load, idle time, temperature) for higher accuracy.

Key Recommendations

Operational

1. Introduce condition-based monitoring (CBM):
   • Install sensors or use existing telemetry (engine temp, vibration, oil condition) to trigger proactive maintenance before predicted failure > 50%.
2. Reallocate workload:
   • Balance truck assignments; trucks TRK06, TRK08, TRK17, and TRK18 appear overused or poorly maintained.

Maintenance

3. Increase preventive maintenance frequency during high-risk months (Feb–Apr).
4. Target MTBF improvement to at least 75+ hours by analyzing root causes of frequent stoppages.
5. Audit maintenance logs of high-performing trucks (TRK02, TRK04) to identify maintenance best practices.

Summary

The Mine Truck Reliability and Failure Analysis reveals meaningful opportunities to enhance uptime, reduce breakdown frequency, and support more predictive maintenance planning. By integrating predictive insights with historical downtime data, fleet operators can shift from reactive repairs to proactive reliability management—ultimately improving productivity and reducing operational costs.

2. Truck Operational Performance

Project Overview

This analysis evaluates the operational performance of a fleet of trucks using telemetry data such as run time, idle time, fuel consumption, and oil temperature. The goal is to understand efficiency patterns, detect areas of operational waste, and support data-driven decision-making to improve fuel economy, reduce idle time, and maintain optimal engine health.

Interactive dashboards visualize multi-dimensional data to uncover relationships between run hours, fuel rate, idle behavior, and seasonal performance trends across the fleet.

Key Insights

1. Fuel Rate vs. Run Time

• The scatter plot shows a positive correlation between Run_Time_Hours and Fuel_Rate (L/hr).
• Trucks with longer run times consume proportionally more fuel — this is expected.
• However, a few trucks (e.g., top-right of the scatter) show higher fuel usage than average for similar run times, indicating inefficient fuel performance.

Insight: Fuel efficiency varies across trucks; some trucks may require calibration or maintenance.

2. Truck Utilization (Run vs Idle Time)

• From the bar chart by Truck_ID, most trucks show high Run_Time_Hours compared to Idle_Time_Hours, which is good.
• A few (e.g., TRK09, TRK15) have significantly higher total hours — possibly over-utilized or doing double shifts.
• Trucks with high idle time ratios are underperforming and increasing operational cost.

Insight: Idle time should be reduced; balancing workloads across fleet can improve efficiency.

3. Oil Temperature Trend

• Oil Temperature (°C) fluctuates month-to-month.
• Peaks around May and June, which might be linked to warmer ambient temperatures or heavier operational demand.
• Some high readings may approach risk thresholds.

Insight: Seasonal or workload effects influence oil temperature — trucks might need more frequent oil checks in warmer months.

4. Fuel Rate & Truck Count by Month

• Monthly Fuel_Rate shows spikes around May–June, matching oil temperature trends.
• Indicates increased fuel consumption — could be due to longer shifts or hauling heavier loads.

Insight: Monitor monthly consumption trends to identify overconsumption periods.

5. Idle & Run Time by Month

• Run_Time_Hours remain consistent, but Idle_Time_Hours slightly rise in some months (e.g., March–April).
• This might point to waiting times, logistics delays, or driver behavior issues.

Insight: Consistent production but minor inefficiencies during certain months — opportunity for process optimization.

Key Recommendations

Fleet Operations

• Optimize scheduling: Balance run hours across trucks to prevent overuse and uneven wear (TRK09, TRK15).
• Reduce idle time: Implement driver awareness programs or automatic engine shutdown systems to reduce idling fuel waste.

Maintenance & Performance

• Fuel Efficiency Check: Investigate trucks with unusually high fuel rates for possible fuel leaks, tire pressure issues, or poor engine tuning.
• Monitor Oil Temperature: Schedule predictive maintenance for trucks showing consistently high oil temps.
• Predictive Maintenance KPI: Use the Predicted_Failure_Prob (%) to prioritize trucks for inspection before breakdowns.

Seasonal & Strategic Actions

• Temperature-Aware Scheduling: During high-temperature months, reduce continuous run time to avoid overheating.
• Data-Driven Maintenance Calendar: Use your dashboard trends to plan maintenance when utilization is low to reduce downtime impact.

Summary

This operational performance analysis provides actionable insights that can significantly improve fleet efficiency, reduce fuel costs, minimize idle waste, and enhance engine health. By combining telemetry trends with predictive insights, organizations can adopt a more proactive and cost-effective fleet management approach.

3. Failure Prediction & Risk Analysis

Project Overview

This project focuses on predicting truck failure probability and analyzing operational risk using historical run-time data, maintenance records, and model-generated failure probabilities. The goal is to help fleet managers proactively identify high-risk trucks, optimize maintenance scheduling, and reduce unplanned downtime.

A predictive analytics model was used to estimate the likelihood of each truck experiencing a mechanical failure. The results were visualized through interactive dashboards that highlight average failure probabilities, monthly trends, and the relationship between run-time hours and failure risk.

Key Insights

1. Overall Fleet Risk Level

• The average predicted failure probability across the fleet is ~49.75%, indicating a moderate risk level.
• This suggests nearly half of the fleet shows warning signs that could lead to failures without timely intervention.

2. High-Risk Trucks Identified

Certain trucks have significantly higher predicted failure probabilities:

• TRK20 (62.22%), TRK13 (55.99%), TRK15 (56.28%), TRK06 (53.56%), TRK10 (53.83%), TRK14 (47.55%)
• These trucks should be prioritized for diagnostics and preventive maintenance.

Low-risk trucks include:

• TRK05 (33.10%), TRK17 (42.08%), and TRK09 (42.79%), showing relatively stable performance.

The variance across trucks highlights inconsistent wear patterns—likely influenced by route types, driving conditions, or asset age.

3. Monthly Trend Analysis

The monthly average failure probability fluctuates between ~45% and 60%, with notable peaks:

• March & July show the highest risk levels (~55–60%).
• February, May, and December have comparatively lower values (~45–47%).

These seasonal patterns may correspond to:

• Increased operational load.
• Weather impacts (temperature extremes).
• Maintenance cycles misaligned with peak usage months.

4. Run-Time Hours vs Failure Probability

The scatter plot reveals:

• A positive correlation between run time and failure probability—trucks with higher cumulative run hours generally show higher failure risk.
• However, some outliers exist:
  • Trucks with high run hours but moderate risk (good maintenance history).
  • Trucks with relatively low run hours but high risk (potential underlying mechanical issues).

This suggests that while run time is a strong predictor, other factors such as component age, maintenance delays, or route difficulty also contribute to failure risk.

Recommendations

1. Prioritize High-Risk Trucks for Immediate Maintenance

• Create a high-priority maintenance list for trucks with risk >55%.
• Conduct full mechanical diagnostics on TRK20, TRK13, TRK15, TRK10, and TRK06.

2. Align Maintenance Scheduling With Monthly Risk Trends

• Increase preventive maintenance during February (before March peak) and June (before July peak).
• Review workload distribution during peak-risk months to reduce strain.

3. Optimize Fleet Utilization

• Rotate usage of low-risk trucks to balance run-time distribution.
• Reduce load on high-usage, high-risk trucks to avoid sudden breakdowns.

4. Implement Predictive Maintenance Alerts

Based on model outputs:

• Trigger alerts when a truck’s predicted failure probability exceeds 50%.
• Use real-time telematics data to continuously refresh predictions.

5. Investigate Outliers

• Trucks with low run time but high risk may require deeper mechanical inspection.
• Trucks with high run time but low risk can be used as benchmarks for best-practice maintenance.

Summary

This truck failure prediction system provides actionable insights that support data-driven maintenance planning and fleet risk management. By leveraging predictive analytics and visual dashboards, fleet operators can reduce unscheduled downtime, improve safety, and extend asset life cycles.