5G technology is transforming the industry by enabling drones to operate with greater precision and in real-time, improving the efficiency and safety of operations, allowing more effective and reliable inspections and monitoring. Additionally, 5G connectivity allows multiple users within a company, spread across the globe, to conduct remote construction meetings, facilitating collaboration and real-time decision-making.
5G SA Impact in the Use Case:
Low Latency and High Speed - 5G SA networks provide ultra-low latency and high-speed data transmission, which is crucial for real-time control and monitoring of drones
Extended Range and Coverage - With 5G SA, drones can operate over larger areas without losing connectivity
Enhanced Safety - The reliable and stable connection provided by 5G SA networks ensures that drones can be operated safely even in challenging environments
Increased Efficiency - Drones equipped with 5G connectivity can perform inspections faster and more efficiently than traditional methods
Automation and Remote Operation - 5G SA networks support the use of fully automated drones, which can be programmed to follow specific flight paths and perform tasks without human intervention
Beyond Visual Line of Sight Operations - 5G SA enables BVLOS operations, allowing drones to fly beyond the operator’s line of sight
Fast, Accurate and Safe with 5G
• Indoor heavy-duty drones to do instant, quick and safe inspections
• Scanning and rendering while inspecting
• Measuring stock levels in stockholes and silos
• Shorten the stoppage periods and costs by accessing complicated and risky zones easily
Use cases:
• Kiln Inspections
• Cyclones inspections
• Stock measurement Piles inside silos or stock polar
• Remote thickness measurements
• Internal silo inspections
• Inspections internal structures and gas ducts
These drones use thermal imaging technology to precisely identify equipment defects and efficiency anomalies. Perform daily autonomous inspections along preset flight paths without human intervention. This allows inspectors to remotely monitor and compare defects in solar farms and temperature abnormalities in cement plant equipment, significantly reducing the labour and time costs associated with manual inspections.
Fast, Accurate and Safe with 5G
• Outdoor inspections and real-time thermal monitoring
• 3D rendering and project management
• Autonomous audits to check deviations and anomalies
Use cases:
• Thermal inspection in preheater tower
• Thermografical work in main substation of plants
• Solar Photovoltaic System Inspection
• Stock measurements piles in coal and quarry
• Precisly inspection of some equipments that are not accessible
• Clear image of mettalic, concrete strutures and roofs and evolution degradation by comparing the sequence of images
• Leack detection
• Made the 3 D model design when is necessary
• Digital twins project to industry 4.0
Engineers can use voice commands to access necessary information, which is then displayed via augmented reality (AR).
If equipment malfunctions or other issues arise, they can instantly connect with remote experts for technical support.
Through AI, engineers receive clear, step-by-step guidance to precisely complete equipment maintenance, significantly reducing errors and ensuring smooth operation. Furthermore, the AI smart glasses can display real-time equipment health data, allowing engineers to anticipate issues before they occur.
Remote Advanced Support and Effective Field Works
• Handsfree control system with AI-powered voice recognition
• Communicate and access information without interrupting tasks
• Remote assistance and training usages
• Follow digital workflows connection with SAP PM
• Visualize IoT data
• Ultra-fast data transfer rates
• Multi language
• Real-time video streaming
• Data processing without delays
• Secure and protected against external threats
Remote Advanced Support and Effective Field Works
• Real-time high-definition video transmission
• Low-latency and ultra fast video transfers
• Surveillance, monitoring and live streaming from the field
• AI analytical integrations: instant detections, interactive warnings and notifications
• AI supported dynamic load and anomaly analytics with motion amplification
Smart factories and next-generation solutions with 5G MPN
• Sophisticated IoT sensors
• Edge computing and AI
• Indoor & Outdoor Drones
• UHD Video Analytics
• Digital Twin Integrations
• Autonomous Operating Technologies
• Remote Support & Access without cables
Digital Transformation and Its Effects
• Minimized risks and eliminated events with effective monitoring
• Improved reliability with early detections of failures
• High productivity and best quality with less interruptions
• Energy savings with efficient processes and operations
• Decreased maintenance costs and optimized inventories with predictive failure potentials and lifetime analyses
• Higher efficiency and low emissions in quarries: autonomous mining operations
• Talent attraction: new generation to develop sustainable culture and further innovations
Smart Sensors Integrated with AI to build a cloud-based system for continuous remote monitoring of equipment. Abnormalities can be detected before failures occur, enabling predictive maintenance.
CIMPOR, in collaboration with FIZIX, is implementing a large-scale AI-powered smart monitoring system across 14 plants in 4 countries. As part of this initiative, smart sensors have been successfully deployed; Türkiye: 8,100 sensors, Portugal: 2,070 sensors, Cameroon: 228 sensors, Ivory Coast: 202 sensors.
This system utilizes AI to continuously monitor the health of cement plant equipment, enabling real-time monitoring and advanced data analysis. It acts as a smart brain for production operations, predicting and preventing potential equipment failures in advance.
By enabling predictive maintenance, the system ensures stable factory operations, minimizes unplanned downtime, and prevents losses caused by production interruptions.
Project Numbers: IIoT Integrations
Project Architecture: Sensors, Communication and Analytics
IIoT Sensors Specifications
Vibration: By utilizing a 3-axis sensor and collecting data up to 2.6.6 kHz, it's possible to simultaneously monitor all faults (such as unbalance, looseness, misalignment, bearing/gear faults, cavitation, etc.) at the same time.
Acoustics: Detect issues with your bearings, gears, and oil at a very early stage with the help of ultrasound data collection speed up to 80kHz.
Magnetic Flux: By collecting data up to 32 kHz, it's possible to detect motor specific problems like broken bar faults, eccentricity, short circuit, load unbalance, misalignment and winding issues.
RPM: Accurately predicts machine RPMs with the combination of vibration and magnetic flux sensor data.
AI-powered predictive maintenance platform
Data Collection & Data Lake Storage
AI-Powered Predictive Software
Predictive Maintenance - Approaches
Rule-based Approaches
Rule-based approaches in predictive maintenance use predefined thresholds and conditions for detection. This method is straightforward to implement and easy to understand, making it accessible for teams without advanced technical expertise. However, it lacks flexibility and may not adapt well to changing conditions or complex scenarios. This rigidity can lead to higher false positives or negatives, as the system may not account for all possible variations in the data
* Easy to implement and adjust
* Requires domain expertise
* Rigid structure
ML Anomaly Detection
ML anomaly detection employs machine learning algorithms to identify patterns and anomalies. This approach offers higher accuracy and adaptability to changing conditions, as the algorithms can learn from new data and improve over time. However, it requires more technical expertise to implement and maintain, and the initial setup can be resource-intensive. Additionally, the complexity of the models can make them less transparent, which might be a concern for stakeholders who need to understand the decision-making process
* High accuracy
* Earlier detection
* Requires data science expertise
Health Condition Score
The score level for all sensors is determined according to the alarm limits established, ranging from normal status (green) to critical status (red).
The overall machine health status is determined by combining all sensor scores according to mathematical rules. The result is represented on a scale from green to red.
The aggregate of each machine's health score will determine the overall system health score.
Ultimately, health condition scores will be available at every level—from individual sensors and machines up to the entire plant.
Machine Health Monitoring Benefits
Machine Learning Models Building Concept
Historical sensor data from the PI System is utilized to characterize typical asset operations and to develop a corresponding model.
Erroneous or noisy data is omitted, and appropriate filters are implemented to ensure analysis is limited to steady-state operations.
The system employs machine learning algorithms to analyse standard asset behaviour, establishing a baseline model through clustering and regression methodologies.
The machine learning model identifies discrepancies between actual values and predicted values.
Alerts, Notification & Validation
Each time Predictive Analytics creates an alert, the Predictive team receives an automatic email notification..
The categorisation of machine learning model outputs into true positives and false positives are essential for effectively assessing and enhancing the mathematical models and performance of predictive systems.
Smarter Insights, Every Day, our generative AI brings together three specialized agents—a Statistical Analyst, a Technical Expert, and an Optimizer—working in parallel to deliver clarity and control. Every day, automated reports align laboratory strength values with model predictions, enriched with AI-generated commentary and actionable guidance.
The real game-changer is the Optimizer, which focuses on the top five strength predictors such as, SO₃/alkali balance, C₃A cubic-orthorhombic ratio, MS, etc. to give optimized working ranges, ensuring stable clinker reactivity and consistent cement performance. By transforming complex data into practical guardrails, operators gain the power to act early, reduce variability, and maintain product quality with confidence.
The result is lower deviations, more consistent strengths, and stronger customer trust—all driven by intelligence built in-house and tailored to the realities of cement production.
AI-Powered F.CaO Predictions for Stable Clinker
Our internally developed Artificial Intelligence models deliver daily predictions of free lime (F.CaO), combining operational signals and quality measurements to ensure stable clinker performance. By continuously learning from kiln conditions, raw mix chemistry, and laboratory data, the system provides reliable forecasts that reduce trial-and-error and give production teams a forward view of clinker reactivity.
A key innovation is the direct link between early strength predictions and optimized F.CaO targets. By analyzing how parameters such as LSF, SM, AM, burning zone conditions, and cooling rates affect both F.CaO and downstream strengths, the AI defines optimal working ranges that prevent over-burning, minimize variability, and support consistent cement quality.
The outcome: greater process stability, lower energy costs, and reliable clinker reactivity—all achieved with in-house AI, built for the realities of cement production.
Automated Reporting, Seamless Integration
Our reporting system transforms data into decisions by fully automating the flow of production, energy, and quality information. All operational and energy data streams are collected in the central datalake and transferred directly to the reporting platform. Calorific values are automatically retrieved from SAP and the quality database, ensuring accuracy and eliminating manual entry.
Once consolidated, the plant team reviews the reported values, and with a single step the validated data is sent back to SAP and directed into BW (Business Warehouse) for corporate reporting. This closed-loop process ensures consistency, transparency, and real-time visibility—delivering reliable insights while reducing workload and risk of error.
Beyond Conventional APC – Towards Autonomous Plants
At CIMPOR, improving plant efficiency and effectiveness means pushing the boundaries beyond traditional Advanced Process Control (APC). While APC stabilizes key process variables, we are enhancing this foundation with end-of-the-line AI technologies that combine predictive analytics, machine learning, and optimization engines. The goal is not just to react to process changes, but to anticipate them, recommend proactive adjustments, and automate decisions across the entire value chain.
This approach brings us closer to the concept of the autonomous plant—an operation where kiln, mill, and quality processes interact seamlessly, guided by AI that continuously balances energy efficiency, product quality, and equipment reliability. By embedding intelligence at every stage, from raw mix preparation to finished cement delivery, CIMPOR is shaping a future where plants can self-optimize, reduce manual intervention, and achieve consistent performance with greater agility and resilience.
