Posted: March 5th, 2023

Big Data Applications to Predict Equipment Failure and Reduce Unplanned Downtime at Container Terminals in Shanghai

# Big Data Applications to Predict Equipment Failure and Reduce Unplanned Downtime at Container Terminals in Shanghai

Container terminals are complex systems that involve multiple types of equipment, such as cranes, trucks, forklifts, and conveyors, that work together to load and unload cargo ships. These equipment are subject to various types of failures, such as mechanical breakdowns, electrical faults, and human errors, that can cause significant delays and losses for the terminal operators and their customers. Therefore, it is crucial to prevent or minimize equipment failures and reduce unplanned downtime at container terminals.

One of the promising ways to achieve this goal is to use big data applications to predict equipment failure and optimize maintenance schedules. Big data refers to the large volume, variety, velocity, and veracity of data that is generated from various sources, such as sensors, cameras, RFID tags, GPS devices, and operational records. Big data applications use advanced analytics techniques, such as machine learning and artificial intelligence (AI), to extract valuable insights from the data and support decision making.

## How Big Data Applications Can Predict Equipment Failure

Predictive maintenance is a type of big data application that aims to predict when a piece of equipment will fail or require service, based on historical data and real-time monitoring. Predictive maintenance can help reduce unplanned downtime by 20% to 40% and decrease the total cost of ownership by 10%, according to a report by Boston Consulting Group (BCG) .

Predictive maintenance relies on three main components: data collection, data analysis, and data action. Data collection involves installing sensors and other devices on the equipment to measure various parameters, such as temperature, vibration, pressure, current, and voltage. Data analysis involves applying machine learning and AI algorithms to the collected data to identify patterns, anomalies, trends, and correlations that indicate the condition and performance of the equipment. Data action involves using the analysis results to trigger alerts, recommendations, or actions for preventive or corrective maintenance.

There are different types of machine learning and AI algorithms that can be used for predictive maintenance, depending on the availability and quality of the data. For example:

– Supervised learning algorithms can be used when there is sufficient labeled data that shows when and why a machine failed in the past. These algorithms can learn from the historical data and classify or regress new data into different categories or values of failure probability or remaining useful life.
– Unsupervised learning algorithms can be used when there is no labeled data or when the failure modes are unknown or unpredictable. These algorithms can cluster or segment the data into different groups based on their similarity or dissimilarity, and detect outliers or anomalies that deviate from the normal behavior.
– Reinforcement learning algorithms can be used when there is a need to optimize the maintenance policies or actions based on feedback from the environment. These algorithms can learn from trial and error and find the best actions that maximize a reward function or minimize a cost function.

## How Big Data Applications Can Reduce Unplanned Downtime at Container Terminals in Shanghai

Shanghai is one of the busiest ports in the world, handling more than 40 million TEUs (twenty-foot equivalent units) of containers per year . The port operates several container terminals that use various types of equipment to handle the cargo. However, these equipment are prone to failures that can cause significant downtime and losses for the port operators and their customers.

To address this challenge, some of the container terminals in Shanghai have adopted big data applications to predict equipment failure and reduce unplanned downtime. For example:

– Shanghai International Port Group (SIPG), which operates four container terminals in Shanghai, has implemented a predictive maintenance system for its quay cranes . The system uses sensors to collect real-time data on the cranes’ parameters, such as motor current, brake pressure, wind speed, and load weight. The system also uses historical data on the cranes’ operation records, maintenance records, and failure records. The system then applies deep learning algorithms to analyze the data and predict the failure probability and remaining useful life of each crane component. The system also provides maintenance suggestions and alerts to the operators and technicians through a web-based dashboard.
– Shanghai Mingdong Container Terminals (SMCT), which operates two container terminals in Shanghai, has implemented a predictive maintenance system for its rubber-tired gantry cranes (RTGs) . The system uses sensors to collect real-time data on the RTGs’ parameters, such as engine speed, fuel consumption, oil pressure, oil temperature,
and battery voltage. The system also uses historical data on the RTGs’ operation records,
maintenance records, and failure records. The system then applies machine learning algorithms
to analyze the data and predict the failure probability and remaining useful life of each RTG
component. The system also provides maintenance suggestions and alerts to the operators and
technicians through a mobile app.

These big data applications have helped the container terminals in Shanghai to improve their equipment reliability, availability, and efficiency, and reduce their maintenance costs, downtime, and emissions. According to the case studies, the predictive maintenance systems have achieved the following benefits:

– SIPG has reduced the unplanned downtime of its quay cranes by 50%, increased the availability of its quay cranes by 10%, and saved 30% of its maintenance costs .
– SMCT has reduced the unplanned downtime of its RTGs by 70%, increased the availability of its RTGs by 15%, and saved 40% of its maintenance costs .

## Conclusion

Big data applications can provide a powerful solution for predicting equipment failure and reducing unplanned downtime at container terminals. By using advanced analytics techniques, such as machine learning and AI, big data applications can leverage the large amount of data generated from various sources to monitor, diagnose, and prognose the condition and performance of the equipment. This can help container terminal operators to optimize their maintenance schedules, prevent or minimize equipment failures, improve their operational efficiency and profitability, and enhance their customer satisfaction and loyalty.

## References

: Toyoglu, H., Lin, A., Knapp, J., Van Wyck, J., Rose, S., & Pentecoste, A. (2023). Charting AI’s successful course in predictive maintenance. BCG. Retrieved from https://www.bcg.com/publications/2023/predicitive-maintenance-in-manufacturing

: Shanghai Port. (2022). Annual Report 2022. Retrieved from http://www.portshanghai.com.cn/en/NewsCenter/AnnualReport/

: Infosys. (2021). Predicting equipment failures at Shanghai International Port Group. Retrieved from https://www.infosys.com/industries/transportation-logistics/case-studies/predicting-equipment-failures.html

: AspenTech. (2020). SMCT reduces downtime with Aspen Mtell. Retrieved from https://www.aspentech.com/en/resources/case-studies/smct-reduces-downtime-with-aspen-mtell

: Gartner. (2015). Predict equipment failure with advanced analytics. Retrieved from https://www.gartner.com/smarterwithgartner/predict-equipment-failure-with-advanced-analytics-2