DEVELOPMENT OF AN ELEVATOR MONITORING SYSTEM USING ARTIFICIAL INTELLIGENCE

АННОТАЦИЯ

В этой статье рассматривается, как искусственный интеллект (ИИ) улучшает системы мониторинга лифтов, обеспечивая предиктивное обслуживание и обнаружение неисправностей в реальном времени.

Благодаря машинному обучению и интеграции IoT системы на основе ИИ повышают безопасность лифтов, сокращают время простоя и оптимизируют эксплуатационные расходы. Мы сравниваем подходы к мониторингу на основе ИИ с традиционными системами, подчеркивая повышение эффективности за счет предиктивной аналитики.

ABSTRACT

This paper explores how artificial intelligence (AI) enhances elevator monitoring systems by enabling predictive maintenance and real-time fault detection. Through machine learning and IoT integration, AI-driven systems improve elevator safety, reduce downtime, and optimize operational costs. In next be compare AI-based monitoring approaches with traditional systems, highlighting the efficiency gains from predictive analytics.

Ключевые слова: ИИ, предиктивное обслуживание, обнаружение неисправностей в реальном времени, безопасность лифтов, аналитика данных.

Keywords: AI, predictive maintenance, real-time fault detection, elevator safety, data analytics.

1. Introduction

Elevators are essential components of modern infrastructure, transporting daily millions of individuals across various building types—residential, commercial, and industrial. However, despite their importance, elevators frequently encounter challenges concerning safety, dependability, and operational efficiency [1][2].

The intricate nature of elevator systems, coupled with their high usage rates, underscores the necessity of consistent monitoring and maintenance to prevent malfunctions, delays, and potential hazards.

Traditionally, elevator maintenance has been reactive, addressing issues post-failure or according to set schedules, which leaves room for unexpected breakdowns [3] and other issues.

Recent technological advancements have led to the creation of elevator monitoring systems capable of continuously tracking the condition and performance of elevators [4][5].

Using a variety of sensors and communication tools, these systems monitor elevator operations, identify anomalies, and alert maintenance personnel about possible issues. Although these systems have enhanced elevator management, they still lack the real-time predictive insights that could substantially cut downtime and bolster safety [6].

The integration of artificial intelligence (AI) has introduced transformative potential to elevator monitoring. AI-enabled systems are engineered to process large datasets in real-time, foresee potential issues, and fine-tune elevator performance [7]. This proactive approach not only reduces unexpected failures but also enhances the overall efficiency and safety of elevator operations. By embedding AI within elevator monitoring frameworks, building managers and service providers can achieve significant improvements in reliability while lowering maintenance costs and minimizing interruptions [8].

This paper examines the present state of elevator monitoring technologies and the transformative role of AI within this field. It delves into the various monitoring systems available today, AI's contributions to these systems, and the advantages and limitations of AI-driven monitoring.

 Additionally, the study provides an in-depth analysis of the architecture, data collection techniques, and AI training models applied in elevator monitoring, followed by an assessment of these systems' effectiveness based on test outcomes [9].

2. Literature Review

2.1 Elevator Monitoring Systems:

Elevator monitoring systems have seen substantial development over the years. Earlier systems offered basic alarms and fault detection, whereas modern systems track multiple parameters that impact elevator performance, safety, and usage [10].

1. Basic Fault Detection Systems: These are the most straightforward monitoring systems, utilizing sensors to identify issues such as door malfunctions, overloads, and misalignment [11]. These systems trigger alarms when a fault is detected, alerting maintenance staff to resolve the issue. Although affordable, they are reactive, addressing faults only after they arise.
2. Remote Monitoring Systems: Employing Internet of Things (IoT) technology, these systems gather real-time elevator data and transmit it to a central server, enabling remote monitoring of elevator performance and rapid response to any irregularities [12]. While these systems offer enhanced control and real-time insights, they tend to be more costly and require reliable internet access.
3. Predictive Maintenance Systems: These advanced systems leverage sensors and algorithms to anticipate potential elevator failures. By analyzing historical data on usage patterns, wear, and environmental conditions, they predict failures and schedule preemptive maintenance [13]. Though highly effective at reducing downtime and boosting safety, these systems are complex and costly to implement.

2.2 Comparison of Elevator Monitoring Systems:

Table 1.

Comparison system

Based on the analysis of this comparison, we can clearly see that each of the systems has its own advantages, as well as a variety of choices for certain needs and technical tasks of the customer.

2.3 Artificial Intelligence in Elevator Monitoring:

AI holds significant potential for improving elevator monitoring by offering predictive insights and optimizing operations in real-time. Integrating AI into these systems often involves machine learning algorithms that process large volumes of data gathered from elevator sensors [14].

2.4 Examples of AI-powered Elevator Monitoring Systems:

Several companies are already incorporating AI in their monitoring systems

- KONE 24/7 Connected Services: This system uses sensor data to monitor elevator performance, predict failures, and optimize maintenance schedules in real-time, presenting insights via a cloud-based platform [15].

- Otis eView: Otis’s system applies machine learning to forecast potential failures based on sensor data, also offering real-time video monitoring for enhanced security [16].

- Schindler Ahead: Schindler’s solution gathers data from connected elevators and escalators to support real-time monitoring, predictive maintenance, and performance enhancements [17].

3. Methodology

3.1 Architecture of an elevator monitoring system based on artificial intelligence:

An elevator monitoring system using Artificial Intelligence (AI) algorithms is a complete solution that includes several key components:

• Sensors: The main source of data for the system are sensors that monitor various elevator performance parameters. These include travel speed indicators, door operation, motor temperature, and vibration levels. The sensors provide a continuous stream of data, which allows even minor deviations from the norm to be recorded [18].
• Data processing unit: The data received from the sensors is pre-processed in real time. This block performs the functions of noise filtering, data normalization and its subsequent transfer to the AI model for detailed analysis [19].
• AI model: The system is based on a machine learning model designed to analyze the collected data and identify patterns indicative of potential faults. Deep learning algorithms (e.g., convolutional and recurrent neural networks) and decision tree-based methods (e.g., gradient boosting) are most used. These approaches can identify complex multivariate dependencies and predict failure probabilities [20].
• Cloud platform: An important part of the system is the cloud infrastructure that provides big data storage, model training, and access to analysis results. It enables building managers and technicians to receive immediate notifications of potential failures and analyze historical data for preventive maintenance planning [21].

Data Collection and Processing:

The process begins with the continuous collection of data from sensors installed on the elevator.

The information collected covers key operating parameters such as speed, vibration levels, temperature and door mechanism status. The data is transferred to a processing unit where it is cleaned and prepared for further analysis [22].

Training and utilization of AI models:

Historical data of elevator operations including breakdowns and maintenance performed are used to create a robust fault prediction system. Machine learning models are trained using different approaches:

Deep learning: Suitable for identifying complex patterns in data, especially in the presence of large amounts of information.

Decision trees and ensemble methods: Applied to build interpretable models that deal effectively with heterogeneous data.

After the training phase, models are validated and tested to assess their accuracy and ability to identify potential problems. In operation, the trained models analyze real-time incoming data and predict potential failures to enable proactive elevator maintenance [23].

4. Results

4.1 Testing of the AI-based elevator monitoring system:

As part of the Digital Elevator project, METEOR Lift entered into an agreement with Sber Business Soft LLC to develop an AI solution using computer vision and speech analytics.

The AI-based system provided timely alerts, reducing downtime and preventing critical failures. Experimental testing demonstrated high accuracy in detecting failures, including door malfunctions, motor overheating, and abnormal vibrations.

Tested System:

An AI-driven elevator monitoring prototype utilizing deep learning and ensemble machine learning was tested. It was compared with a traditional monitoring system based on manual inspections and scheduled maintenance.

Comparison Criteria:

• Failure prediction accuracy
• Response time to anomalies
• Reduction in unscheduled breakdowns
• Overall reliability and downtime reduction

Key Testing Examples:

1. Motor Overheating Detection:

• Initial Data: Normal motor temperature: 60°C. Increased load raised it to 69°C (+15%).
• Result: AI detected abnormal temperature rise and sent alerts, preventing failure. Traditional monitoring relied on manual inspections, causing delays.

2. Abnormal Vibration Detection:

• Initial Data: Normal vibration: 2.5 mm/s. Simulated wear increased it to 3.0 mm/s (+20%).
• Result: AI identified deviations and predicted maintenance needs within 24 hours, preventing major failures.

3. Door Mechanism Malfunction:

• Initial Data: Normal closing time: 3 sec. Simulated fault increased it to 8 sec.
• Result: AI detected the issue and sent alerts, preventing service delays. The traditional system relied on user complaints.

4. Failure Prediction Accuracy:

• Initial Data: 150 historical failures: 30% overheating, 40% mechanical wear, 20% door faults, 10% other.
• Result: AI predicted 92% of failures at least 48 hours in advance, significantly reducing downtime [24].

4.2 Performance Evaluation:

To quantify the effectiveness of the system, a comparative analysis of elevator failure rates was conducted before and after its implementation.

 The results showed a significant reduction in elevator downtime: the system successfully predicted about 90% of potential failures at least 48 hours before their actual occurrence.

Additionally, implementation of the system allowed to optimize the maintenance schedule, reduce the cost of unscheduled repairs and increase the safety of elevator operation.

These results confirm the feasibility of using artificial intelligence technologies in elevator equipment monitoring systems, especially in conditions of high operational load.

4.3 Comparison with Traditional Monitoring Systems:

Compared to traditional systems, the AI-powered solution more effectively reduced downtime, provided precise predictive maintenance alerts, and improved overall elevator reliability [24].

5. Conclusion

In conclusion, AI-based elevator monitoring enhances reliability and safety, though at the cost of higher expenses and complexity.

The predictive and real-time monitoring capacities of AI systems provide substantial benefits over traditional methods, reducing downtime and enabling proactive maintenance. However, challenges like high implementation costs and integration difficulties must be addressed for broader adoption.

References:

1. Ahmed Nabih Zaki Rashed; ManasaYarrarapu. Connected smart elevator systems for smart power and time saving. Scientifc Reports (2024) 14:19330, 2024
2. Abioye, S.; Oyedele, L.O.; Akanbi, L.; Ajayi, A.; Delgado, J.M.D.; Bilal, M.; Olugbenga, O.A.; Ahmed, A. Artificial intelligence in the construction industry: A review of present status, opportunities and future challenges. J. Build. Eng. 2021, 44, 103299.
3. Law Jing Shen; Jacqueline Lukose Lau; Chee Young. Predictive maintenance on an elevator system Using machine learning. Journal of Applied Technology and Innovation (e -ISSN: 2600-7304) vol. 5, no. 1, (2021)
4. Kesheng Wang; Guohong Dai; Lanzhong Guo. Intelligent Predictive Maintenance (IPdM) for Elevator Service- Through CPS, IOTaS and Data Mining 2021.
5. Svetlana Ivanova; Aleksandr Kuznetsov; Roman Zverev; Artem Rada. Artificial Intelligence Methods for the Construction and Management of Buildings. Journal, 2023, 23, 8740.
6. Nitasha Hasteer; Jasmine Awatramani; Gaayan Verma. Implementing Deep Learning Model to Predict the Maintenance of an Elevator System. 2nd International Conference for Emerging Technology (INCET), (10.1109/INCET51464.2021.9456133) 2021.
7. Chen, H.-P.; Ying, K.-C. Artificial Intelligence in the Construction Industry: Main Development Trajectories and Future Outlook. Appl. Sci. 2022, 12, 5832.
8. Pan, Y.; Zhang, L. Roles of artificial intelligence in construction engineering and management: A critical review and future trends. Autom. Constr. 2021, 122, 103517.
9. Chien, C.F.; Dauzère-Pérès, S.; Huh, W.T.; Jang, Y.J.; Morrison, J.R. Artificial intelligence in manufacturing and logistics systems: Algorithms, applications, and case studies. Int. J. Prod. Res. 2020, 58, 2730–2731.
10. Regona, M.; Yigitcanlar, T.; Xia, B.; Li, R.Y.M. Opportunities and Adoption Challenges of AI in the Construction Industry: APRISMA Review. J. Open Innov. Technol. Mark. Complex. 2022, 8, 45.
11. Vasilios Zarikas; Nurislam Tursynbek. Intelligent Elevators in a Smart Building. Future Technologies Conference (FTC) 2019.
12. You Zhou; Kai Wang; Hongxia Liu. An Elevator Monitoring System Based On The Internet Of Things. 8th International Congress of Information and Communication Technology (ICICT-2018).
13. Ziliang An1; Dingsong Bai; Ningli Gan; Ning Wang; Jiyu Deng; Yuxin Huang. The Enlightenment on Prediction and Health Management Technology of Elevator System. Journal of Physics. Conference Series 1939 (2021) 012071, doi:10.1088/1742-6596/1939/1/012071
14. Michael M. Gichane; Jean B. Byiringiro.  Digital Triplet Approach for Real-Time Monitoring and Control of an Elevator Security System. Designs 2020, 4, 9; doi:10.3390/designs4020009.
15. Electronic resource URL: https://www.kone.com/en/
16. Electronic resource URL: https://www.otis.com/en/
17. Electronic resource URL: https://www.schindler.com/en.html
18. Lan Zhang; Jian Lu; Hideki Takagi; Sohei Matsumoto; Eiji Higurashi. An Ultra-Compact MEMS Pirani Sensor for In-Situ Pressure Distribution Monitoring. Micromachines 2022, 13, 1686. https://doi.org/10.3390/mi13101686
19. Zayed, S.M.; Attiya, G.; El-Sayed, A.; Sayed, A.; Hemdan, E.E.-D. An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems. Int. J. Comput. Intell. Syst. 2023, 16, 69.
20. Xi-Yang Jiang, Xiao-Chen Huang, Jian-Peng Huang, Yi-Fei Tong. Real-Time intelligent Elevator Monitoring and Diagnosis: Case Studies and Solutions with applications using Artificial Intelligence. Computers and Electrical Engineering. 2022, 100, 107965.
21. Duc-Minh Pham; Van Tung Tran; Vedisha Vasudha Roopun; Junhong Zhou. Cloud-based Data Analytics Framework for Real-time Lift Monitoring and Diagnostic System. Proc. of the 9th Intl. Conf. of Asian Society for Precision Engg. and Nanotechnology (ASPEN 2022)  2022, Singapore.doi://10.3850/978-981-18-6021-8 OR-08-026
22. Babovi´c, Z.; Bajat, B.; Barac, D.; Bengin, V. Teaching computing for complex problems in civil engineering and geosciences using big data and machine learning: Synergizing four different computing paradigms and four different management domains. J. Big Data 2023, 10, 89.
23. Maria Raluca Raia; Andrei Ailincai; Alexandra Băicoianu; Calin Husar. Condition Monitoring of Industrial Elevators Based on Machine Learning Models. Conference: 2023 IEEE 28th International Conference on Emerging Technologies and Factory Automation (ETFA). 10.1109/ ETFA54631.2023.10275563
24. URL: https://www.cnews.ru/news/line/2024-0606_blagodarya_sberbanku_u_ kompanii