The Power of Data

How Tsaro AI is powering digital twin for elevators and escalators

by Anas Siddiqui
images courtesy of Tsaro

AI has become a household term since the release of ChatGPT in November 2022, demonstrating AI’s ability to perform human-like tasks such as learning, planning and problem-solving. While AI is not a replacement for human intelligence, it serves as a powerful tool for processing and analyzing vast amounts of data at speeds far beyond human capability. By presenting actionable insights, AI helps users anticipate outcomes and make informed business decisions, streamlining operations across industries.

Even before the emergence of ChatGPT, AI had been digitally transforming businesses in various ways. The vertical-transportation (VT) industry faces its own unique challenges and opportunities where AI and modern technologies like big data can make a significant impact. This industry comprises numerous manufacturers and models, each with proprietary data formats, and relies on a wide array of software tools to manage maintenance, safety and compliance records.

This fragmented ecosystem creates a challenge: consolidating diverse data sources into a single platform to unlock the potential of advanced AI technologies. For example, AI can provide predictive alerts such as identifying when an elevator motor is likely to fail within the next two hours, and automate the workflow by creating an emergency work order, ensuring a technician is dispatched and on-site before an outage occurs. 

Additionally, AI can analyze equipment data, maintenance logs and financial records to streamline capital planning, predicting the optimal time for modernization and reducing the total cost of ownership (TCO).

At Tsaro, we have been focusing on the following key areas:

1. Developing an integrative digital twin platform for all types of VT equipment including elevators, escalators, Cartveyors^®, cart lifts and moving walks
2. Establishing robust data ingestion capabilities by integrating diverse data sources such as:
   • MAD Elevator’s Video Management System (VMS) connected to elevator controllers
   • Robustel’s EV8100 elevator voice gateway for capturing discrete fault signals
   • Third-party wireless temperature and vibration sensors.
3. Building data connectors to incorporate additional sources such as manual maintenance records, proprietary software solutions and third-party vendor management systems
4. Training AI models to detect anomalies that indicate potential equipment breakdowns, enabling predictive alerts and transitioning toward “fix-before-break” maintenance strategies to achieve zero downtime

In the second half of 2024, we completed an Internet of Things (IoT) project with one of the largest multinational big-box retailers at one of its Washington, D.C., stores. This project involved enabling all VT equipment in the store — including elevators, escalators, cart escalators and cart lifts — to capture various data points including:

• Elevator controller fault signals, movement data (Figure 3)
• Elevator machine room ambient temperature and pump unit temperature
• Elevator roller and guide vibration (Figure 4)
• Escalator motor temperature and vibration, and handrail temperature
• Cartveyor and cart lift motor temperature and vibration

Using Tsaro’s AI platform, we provided real-time alerts for breakdowns while enhancing alert accuracy by comparing equipment data with actual work orders and service calls. These insights reflected various scenarios, such as:

• Planned equipment shutdowns for preventive maintenance
• Emergency and non-emergency breakdowns requiring work orders

The platform not only automates work order creation, but also enables predictive maintenance, analyzing anomalies to address potential issues before actual breakdowns occur.

Figure 5 illustrates aggregated elevator pump unit temperature data over a month. The AI system first establishes a temperature envelope by recognizing normal temperature ranges and flagging any deviations from these norms. By training the model on this data, AI-powered anomaly detection can identify unusual patterns or outliers in the dataset that deviate from the expected behavior, enabling early detection of potential issues.

Another significant application of AI at Tsaro involves helping customers digitize manual data and processes within the VT industry. This includes ingesting data from customers’ on-premises or cloud-based databases and integrating with other software platforms like ServiceChannel and ServiceNow via APIs (application programming interfaces). 

Tsaro’s cloud-based backend architecture is designed for flexibility, enabling seamless data ingestion from any source. This architecture also supporting the training of AI models to empower autonomous agents capable of analyzing data, setting goals and taking actions with minimal human supervision. 

By leveraging advanced AI techniques, Tsaro is transitioning from rule-based alerts to non-rule-based decision-making processes, paving the way to revolutionize how technology is utilized in the VT industry. The Tsaro AI platform envisions becoming a trusted partner for customers — solving problems, taking actions and making informed decisions collaboratively and efficiently.