AI Machine Guard:
Advanced Machine Health Solution
To Lower Cost, Increase Resilience and Digitalize Safety Management
For further details, please email: johnphuoc@[Link]
From Talent Team From Tsinghua
AI Company established a cooperation base and joint laboratory with the Department of Mechanical Engineering of Tsinghua University, which has
30+ years of academic research in the field of vibration diagnosis and 500+ academic works in the field of vibration diagnosis and has won the
second prize of the National Natural Science Award, the first prize of the Natural Science Award of the Ministry of Education, etc.
Zhang Feibin
Meng Li Wang Fabio
Postdoctoral Fellow
Chairman CMO
CEO CTO
Tsinghua University
Ph.D. Tsinghua University-MIT Master of Tsinghua University
Expert in mechanical dynamics, fault diagnosis
Engaged in MEMS, sensor, integrated circuit Project experience of GE, CGN,
mechanism and model algorithm, signal
and other professional technology research for COMAC, AECC, Zijin Mining and other
extraction and analysis algorithm. Published
many years. Published more than 20 domestic large enterprises.
more than 14 patents, 12 international top-level
Professor Chu Fulei and foreign patents and more than 10 articles, 1 monograph, and 4
Director of Fault Diagnosis Committee of international top-level articles. software copyrights.
Chinese Society of Vibration Engineering
Member of International Federation for
the Promotion of Mechanism and
Machine Science (IFToMM) Liu Zhi Lou Jiamiao Yang Kangding
Tenured Professor, Department of PMO
PMO CPO
VP Tsinghua Master COO
Machinery, Tsinghua University Master of Tsinghua University Master of Tsinghua University Supervisor experience in strategic planning and investment
Recipient of the National Outstanding Design and project management experience in
management in top 500 state-owned enterprise
Youth Fund Work experience in SHELL, an international Minmetals-MCC Group. Market and strategic research,
the world's top 500 state-owned enterprises.
top manufacturing company, and served as project investment and operation, investment and financing
Vice President of Chinese Society of Registered Public Equipment Engineer, rich
management of listed companies in metallurgy, mining,
Vibration Engineering experience in process engineering design.
Digital Product Director of SHELL.
environmental protection, aviation fields.
2
From mechanical engineering to intelligent system
development
GE global supplier，
Team built with Piston pump R&D, Aircraft engine support
background from US manufacturing in equipment R&D,
and Germany Jiangsu Changzhou manufacturing
2012- 2013 2014 - 2016 2017 - 2018 2019-2020 2021-2023 2024-2025
Intelligent hardware
production line
The predictive
landing.
maintenance business
Preparing for IPO.
covers multiple
industries and grow at
Built for predictive a high speed
maintenance service Chemical industry,
Set up a joint laboratory metallurgy, electric
with Tsinghua University power, environmental
protection and other
industries to establish
benchmarking cases,
business sustained rapid
growth
What is AI Machine Guard
Real-time machine heath care,
know where the fault is, and
guide how to fix it
The system continued to run for
h
The accumulated intelligent
analysis data was 111,550 groups
• Real-time health care: 7*24h real-time
machine health care to keep your
production lines running at their very best
• Accurate AI diagnosis: accurate diagnosis
into specific fault types and components,
to help guide maintenance
• Early Warning: predict very early potential
failure, normally 2 months in advance
Rotor imbalance
Poorly
Rubbing lubricated
Motor
Loose failure
ning
Misalignment Bearing damage
4
Advanced technology: AI & mechanism models
The raw vibration
signal relies on expert
Mechanism
& manual labor
Models
Fault characteristic
frequency
Fundamental
frequency and its
multiplication
Knowledge
Require less
failure samples
For example, a
AI Technology
device with 10
failure modes
Data Algorithm would require tag
numbers of ：
0 1 Adaptive Algorithm Time-
𝐶10 + 𝐶10 +⋯+ Frequency Analysis
10
𝐶10 =1024 Transition recognition without
Computing
With a multi-label hardware support
power model, only 10 • Adapt to time-varying conditions
• No speed sensor required
labels are required.
• Suitable for fast changing
conditions such as aircraft
5
Technology Roadmap of Preventive Maintenance
We
We
Industry level
Industry level
Solution on quayside container crane
Vibration Onsite
Sensor gateway
Onsite
server
Machine served:
Front girder pitching mechanism
Trolley traction mechanism
Main lifting mechanism
Motor fault:
Rotor unbalance, rotor misalignment, mechanical loosening, component friction, abnormal rolling bearing, bearing poor lubrication,
electromagnetic vibration
Gear box fault:
Rotor unbalance, rotor misalignment, mechanical loosening, component friction, abnormal rolling bearing, bearing lubrication poor, gear meshing
poor
Hardware solution
SN Name Type Qty.
1 Vibration sensor FXS010 23
2 Online data acquisition module FXM800 8-ch 1
3 Online data acquisition module FXM800 16-ch 1
4 Onsite server 1
5 Cabinet 42U*G36042 1
6 Switch TL-SF1005 1
Product Wired Vibration Sensor
Signals Vibration
Sensor technology Piezoelectric
Frequency response/sampling 10k / 50k
frequency (Hz)
Edge calculation RMS
Communication Ethernet，Wi-Fi
Battery life N/A
Signal collector Wired capture card, 8/16/24 channels
Sensor location
Point 7
Point 3 Point 4 Point 6
Point 2 Point 5
Point 1
Sensor location
Point 3
Point 2
Point 6
Point 5 Point 4
Point 1
Sensor location
Point 9
Point 2
Point 4 Point 8
Point 5
Point 6 Point 7
Point 3
Point 10
Point 1
Functional Module of AI Machine Guide
Real-time monitoring AI diagnosis Data Dashboard
12
Analysis toolkit Statistical report Mobile app
Project Case: SIPG Mingdong Quay
➢ Phase I monitored equipment includes 3 cranes, with a total of
Enable lean
69 measuring points. management
Real OEE monitoring &
forecasting, scientific
production scheduling, lean
management
Eliminate unplanned
downtime
Advance maintenance to
eliminate potential safety
hazards and avoid unplanned
shutdowns
Reduce
maintenance costs
Prepare spare parts in advance to
formulate a maintenance plan,
reduce maintenance time and
maintenance costs, and confirm
the maintenance effect after the
maintenance is completed.
Machine Health Management Platform Dashboard Diagnosis Report Tools Setting SIPG
The System Run For Monday, January 1, 2022 12:00:00
hours
Asset1
Machine Health Index
Asset1
Asset1
Asset1
Asset1
Machine Health Status Asset1
More Information
Asset1 Asset1
Asset1 Asset1 Alert Running
Asset1
Asset1 Asset1
Health Index
11 Asset1
assets
Summary of Equipment Event Occurrence of Failure Mode (This Year) Plant Health Status Trend (Weekly)
Intelligent diagnosis: Alert
Pretreatment workshop: UNITED No. 2 centrifuge
Intelligent diagnosis: Acceptable Rotor imbalance
Pretreatment workshop: UNITED No. 2 centrifuge
Pretreatment workshop: UNITED No. 3 centrifuge Rubbing
Pretreatment workshop: UNITED No. 4 centrifuge Poorly lubricated
Pretreatment workshop: UNITED No. 5 centrifuge
Pretreatment workshop: UNITED No. 6 centrifuge Motor failure
Pretreatment workshop: UNITED No. 7 centrifuge Misalignment
New inspection/repair this week Bearing damage
On-site On-site
inspection maintenance
Healthy Acceptable Alert Fault