Process Safety Management

&
Use of AI in PSM

S.Datta Choudhury
Dy. Director (Safety)
CLI, Mumbai
CLI Mumbai
 Employers must follow PSM
guidelines
• (1) Develop and maintain written safety information
identifying workplace chemical and process
hazards, equipment used in the processes, and
technology used in the processes;
• (2) Perform a workplace hazard assessment,
including, as appropriate, identification of potential
sources of accidental releases, identification of any
previous release within the facility that had a
potential for catastrophic consequences in the
workplace, estimation of workplace effects of a
range of releases, and estimation of the health and
safety effects of such a range on employees;
Employers must follow PSM guidelines
• (3) Consult with employees and their representatives
on the development and conduct of hazard
assessments and the development of chemical
accident prevention plans and provide access to
these and other records required under the standard;
• (4) Establish a system to respond to the workplace
hazard assessment findings, which shall address
prevention, mitigation, and emergency responses;
• (5) Review periodically the workplace hazard
assessment and response system;
Employers must follow PSM guidelines
• (6) Develop and implement written operating
procedures for the chemical processes, including
procedures for each operating phase, operating
limitations, and safety and health considerations;
• (7) Provide written safety and operating information
for employees and employee training in operating
procedures, by emphasizing hazards and safe
practices that must be developed and made available;
• (8) Ensure contractors and contract employees are
provided with appropriate information and training;
Employers must follow PSM guidelines
• (9) Train and educate employees and
contractors in emergency response
procedures in a manner as comprehensive
and effective as that required by the
regulation.
• (10) Establish a quality assurance program to
ensure that initial process-related equipment,
maintenance materials, and spare parts are
fabricated and installed consistent with design
specifications;
Employers must follow PSM guidelines
• (11) Establish maintenance systems for critical
process-related equipment, including written
procedures, employee training, appropriate
inspections, and testing of such equipment to
ensure ongoing mechanical integrity;
• (12) Conduct pre-startup safety reviews of all
newly installed or modified equipment;
Employers must follow PSM guidelines
• (13) Establish and implement written
procedures managing change to process
chemicals, technology, equipment and
facilities; and
• (14) Investigate every incident that results in
or could have resulted in a major accident in
the workplace, with any findings to be
reviewed by operating personnel and
modifications made, if appropriate.
 Importance of the AI in PSM

• Every 7 seconds one worker is injured

• Manual inspection is done 4 hours in a day.
• Rest 20 hours, no inspection causing huge loss
of 161 Billion Dollar in 2019.
• Manual inspection catches – 10 % violations
• AI based inspection can catch 90% violations
• Drag and drop safety rules possible through AI
• Real Time Notifications through SMS, Email
• Existing camera can be used for AI
• Easily implementable without IoT sensors
Industrial Revolution
 New digital equipment
and technologies

• Wearables for personnel - Can monitor personal health and exposure to

hazardous materials and environments, and track wearers.

• Smart personal protective equipment (PPE) - Electronic sensors in hard

hats, eyewear, footwear, gloves, etc. can continuously collect and
monitor data on the work environment, the user, and/or its own
functioning.

• Voice recognition - Allows hands-free recording of information and

issuing commands by people using tablets and other devices.

• Drones - Can act as data collection devices, for example, to carry out
inspections. Such use is particularly beneficial when hazardous or
difficult-to-reach areas are involved.
THEMES OF INDUSTRY 4.0(Contd.)
• Proximity sensors for vehicles - Sense when a vehicle gets close to
an object to avoid collisions or keep the vehicle on track.

• Machine vision - Uses image processing to enable automated

decision-making, for example, in automated inspections of
equipment.

• Intelligent autonomous robots - Collaborative robots (cobots) can

be employed to help humans or other robots perform their tasks.

• Fundamentally, Industry 4.0 is about data - its generation,

collection, storage, transfer, processing, and use. These data
operations are facilitated by new, updated, and existing technologies
that are adapted to human use.
 Artificial Intelligence
• It refers to the simulation of human intelligence
processes by machines, especially computer
systems.
• AI encompasses various technologies and
techniques that enable computers to perform
tasks that normally require human intelligence,
such as learning from experience, understanding
natural language, recognizing patterns, solving
problems, and making decisions.
 AI enables many OSH improvements

• Automation of many manual activities removes

opportunities for human error while providing data that can
be used to generate an immediate response to events when
needed.
• Continuous process monitoring provides the means to
detect problems before any failures occur avoiding
hazardous situations and production downtime.
• Movement up the hierarchy of controls results from
automation placing less reliance on administrative controls
and helps to eliminate potential employee exposure to
hazards.
• Provision of assistance to people performing difficult or
hazardous tasks.
• Intelligent wearables allow facility personnel to more safely,
reliably and efficiently accomplish their tasks.
 AI BASED OSH APPLICATIONS
• Digital devices can communicate with smart phones designed
for industrial use. Thus, facility personnel can automatically
connect their industrial smart phone to, for example, a
portable or fixed gas detector to provide immediate alerts in
the event of a release.
• Industrial smart phones and notepads can provide workers
with on-demand instructions and visuals for performing tasks
including lists of required tools and information on their use.
• Worker competence can be improved through training that
leverages virtual reality so that workers can experience tasks
before performing them on live plant.
• PPE can be embedded with sensors or radio-frequency
identification (RFID) technology in order to collect and
transmit data. Managers and workers can be automatically
notified when workers are not wearing required PPE or not
 AI enables many OSH improvements (Contd.)

• Digital permits automate previous paper-based processes.

• Digital inspections replace manual ones.
• Big data and analytics identify important trends in
operational data for safety improvements.
• Process hazard (PHA) analytics and big data enable PHA
studies to be mined for insights into accident prevention
and also provide valuable information for other process
safety elements, such as mechanical integrity and
operating procedures.
• Safety records, such as past accident data can be analyzed
to provide insights into how the severity and likelihood of
incidents can be reduced.
 AI BASED OSH APPLICATIONS (Contd.)

• Safety managers can view data transmitted from PPE and wearables
on demand and in real time on their notepad, laptop, or smart phone.
They can immediately determine which workers are using particular
equipment or devices, whether the workers are qualified in its use,
whether the equipment or device is fit for service, and they can view
device readings.
• Electronic confined space monitoring in real time can be used to
monitor worker biometric values (heart rate, body temperature,
breathing rate) as well as their exposure levels to toxics to alert
workers and managers to a potentially dangerous situation, and to
guide emergency rescue operations, if necessary.
• Safety managers can run reports for the various data streams they
receive from field devices and facility personnel. Trends in the data
can alert safety managers to intervene in order to prevent the
development of dangerous situations.
 AI BASED OSH APPLICATIONS (Contd.)
• For the mechanical or asset integrity element of process safety,
a particular smart device in a smart manufacturing facility, for
example, a safety valve, can notify its failure or imminent
failure to the smart integrated manufacturing management
(SIMMS), which can then either take independent action, such
as shutting down a unit or taking other corrective action,
depending on the circumstances, or notify a responsible person
to take action.
• The SIMMS can also initiate inspections or replacements of
similar safety valves, including ordering needed replacements.
Process safety incidents and downtime are reduced or avoided.
• A smart integrated manufacturing management system for may
also applied to process safety management (Smart PSM TM ) to
each element of process safety and how they can be managed
with a SIMMS.
Use the power of technology to monitor 24/7 at all sites

• As a manager, you simply cannot monitor

workers on proper execution of tasks or use of
resources personally 24/7.
• With technology, you can.
• AI remote-monitoring solution can be adapted
for a multitude of industries and
environments, using a variety of IoT-enabled
devices.
 Increase safety with almost any IoT device
• AI workplace safety system was built with IoT-enabled
cameras in mind, but that’s really just the beginning. The
system could be configured to leverage additional assets!
• Facial, image, and speech recognition applications
• Wearable
• Temperature, wind, pressure, and fluid gauges
• Odometers, speedometers, and scales
• Moisture and fluid levels
• GPS, proximity, and boundary detection
• Radio-frequency identification (RFID)
• Lighting and sound meters
• Air quality, particulate levels
 Pricing?
• Pricing is based on two factors:
• The number of cameras (i.e., connected
devices)
• And
• The number of images being processed every
minute.
 AI Sample Architecture

• 5 main components as below:

• Client Software – for camera integrations
• AI Stack – for real-time video processing
• Backend – for transmitting the AI output, managing
Application Programming Interface (APIs) and
databases
• Static Media Storage – for storing the alert
images/videos and AI training images
• Application Dashboard – for users to manage and
track overall OHS performance
Simple Object Access Protocol
Representational State Transfer

Software

Data cloud
 AI Models and Algorithms
• 10x Object detection
• 2D and 3D Body Pose Estimation
• Person Reidentification AI
• Face Anonymization
• Fire and smoke detection
• House keeping models like spill, leakage etc
• Tracker : Receive camera stream and analyze
each use-case with prediction from AI models
and trigger alerts
 Predictive Risk Analysis
• AI algorithms have the capability to process vast amounts of
data and identify patterns in a fraction of the time it would
usually take a human— or even a team of people— using
outdated data analysis methods.
• In OHS, this capability can be leveraged to predict, mitigate,
and outright eliminate potential risks.
• Drawing on a vast data ocean of publicly available historical
incident records, AI can identify patterns and correlations,
proactively pinpointing potential hazards and helping
organizations implement preventive measures and inform
safety procedures.
• This approach allows for a more targeted allocation of
resources and reduce workplace incidents.
 Real-time Monitoring and Early Warning Systems

• AI-powered sensors and monitoring systems can

provide real-time data on various environmental
factors such as air quality, noise levels, and
hazardous substance concentrations.
• These systems can detect potential problems and
issue alerts when critical thresholds are reached,
empowering leaders to quickly make data-driven
decisions and enact swift action.
• Such early warning systems not only enhance
worker safety but also facilitate proactive
environmental management and regulatory
compliance.
 Compliance Management
• OHS compliance involves navigating a complex
web of regulations and standards.
• AI simplifies this process by automating
compliance management through the algorithms’
analysis of regulations, identification of
applicable requirements, evaluation of compliance
gaps, and suggested corrective actions.
• This streamlines the compliance workflow,
ensures adherence to legal obligations, and
reduces the risk of penalties or reputational
damage.
 Incident Response and Investigation

• In the event of an incident, AI can support an

organization’s incident response and investigation
efforts.
• By analyzing incident reports, photographs, and
other relevant data, AI algorithms can help identify
root causes, contributing factors, and systemic
issues.
• This analysis provides valuable insights for
developing corrective and preventive actions,
leading to improved safety practices and a
reduction in incidents over time.
Virtual Training and Simulations
• Training plays a crucial role in OHS management by
ensuring employees are equipped with the knowledge
and skills the need to safely handle work-related
situations.
• AI can enhance training programs by providing
immersive virtual simulations.
• These simulations can replicate hazardous scenarios
and allow employees to practice responding to them
in a controlled environment.
• Such training not only increases preparedness but also
minimizes risks associated with real-life training
exercises.
 What is OSH ?
• OSH means workplace safety and health, is a
multidisciplinary field focused on ensuring the
well-being, safety, and health of workers in
their work environments.
• It encompasses a range of practices, policies,
and regulations aimed at preventing work-
related injuries, illnesses, and accidents while
promoting overall employee well-being.
 Key aspects of OSH
• Workplace Safety : This involves identifying and mitigating
potential hazards in the workplace that could lead to accidents
or injuries. Examples of workplace safety measures include
proper equipment maintenance, the use of personal protective
equipment (PPE), ergonomic design, and fire safety protocols.
• Health Protection: This aspect of OSH focuses on preventing
occupational illnesses caused by exposure to harmful
substances, pollutants, or conditions. It includes monitoring and
controlling factors such as toxic chemicals, noise levels, air
quality, and radiation etc.
• Risk Assessment : Organizations conduct risk assessments to
identify potential hazards and evaluate the risks associated with
specific tasks, equipment, or work environments. This helps in
implementing appropriate preventive measures and safeguards.
 Key aspects of OSH (Contd.)
• Training and Education :Proper training and education for
employees are crucial to ensure that they understand safety
protocols, know how to use safety equipment, and are aware of
potential hazards. Regular safety training sessions help
employees stay informed and prepared.
• Regulations and Compliance : Governments and regulatory
bodies establish occupational safety and health regulations to
set minimum standards for workplace safety. Organizations are
required to comply with these regulations to ensure the safety
of their employees.
• Emergency Preparedness :Having well-defined plans for
emergencies such as fires, chemical spills, or natural disasters is
essential. Employees should be trained on emergency
procedures and evacuation protocols.
 Key aspects of OSH (Contd.)
• Health and Wellness Programs: OSH also encompasses
promoting the overall well-being of employees through
initiatives such as wellness programs, stress management, and
mental health support. A healthy workforce is less likely to
encounter accidents or illnesses.
• Ergonomics: Ergonomics involves designing workspaces,
equipment, and tasks to fit the capabilities and limitations of
workers. Proper ergonomic design can prevent musculoskeletal
disorders and improve overall comfort and productivity.
• Injury Reporting and Investigation: Establishing mechanisms
for reporting and investigating workplace incidents and near-
misses is crucial for understanding the root causes of accidents
and implementing corrective actions to prevent their recurrence.
 Key aspects of OSH (Contd.)
• Continuous Improvement : OSH is an
ongoing process that involves regular
assessments, updates to safety protocols, and
improvements based on feedback and incident
analysis.
• Use of PPE: Use of appropriate PPE is must to
protect workforce from injury and illness as it
is last level of defence.
 Ergonomic Solutions
• The posture of the human body is detected
with body pose estimation model, which in
turn is used to decide whether an action is
performed in ergonomically correct way.
• Object Detection : It is responsible for
detecting objects for predictions to analyze
each case.
 Area Controls

• Establish worker minimum and maximums,

prevent workers from entering restricted areas
• Minimize occupational exposures through set
time limits. (Like STEL)
 Alert Analytics

• Unsafe act/condition image or video footage

• Real-time and date filtered alert analytics
• Analytics trended by shift
• Categorized events into low, medium, and high
risks
 Compliance Tracking

• Track and recognize safe behaviors

• Safe behaviors are integrated to determine the
overall compliance score
• The facility, shift, and department-based
compliance tracking
 Key Point Indicators(KPI) Tracking

• Preset health and safety KPI dashboard

• Filter KPIs by facility, time period, or shift
• Set, measure, and monitor progress on
monthly goals
 PPE Detection
• As the last line of defense in worker
protection, the criticality of the right personal
protective equipment cannot be overstressed.
• Instituting timely worker notification and
intervention using AI can greatly improve
current safety practices and encourage a
productive safety culture.
What type of hazards can be prevented?
• A variety of OHS scenarios on software platform for you to
ensure worker safety in your facilities and reduce the injuries
experienced to zero.
• These use cases can be set up as a rule to send notifications to
the OHS teams in the cases of slips, trips & falls, spill &
leakage, working at heights, electrical hazards, worker machine
interaction, ergonomics manual lifting, human machine
interaction, chemical handling, being hit by falling object,
being hit by a moving object, being trapped between stationary
and moving objects, vehicle collision and many more.
• Any kinds of OHS scenarios can be set up as a rule on your
platform.
 OHS scenarios – May be rule on your OSH platform
• A variety of Core AI features can be activated to detect unsafe acts
and conditions in your workplace.
• The PPE detection model can be used to ensure safe working
environment by detecting the use of Helmet, Glove, Apron, Sleeve,
Respiratory Protective Equipment, Reflective Vest and Glasses.
• The Area Controls model can be used to ensure working areas in
your facility are always safe with various EHS scenarios such as
limiting the number of people in an area, identifying the cases of
working under suspended load, or setting exclusion zones and time
limits.
• The Housekeeping model can be used to reduce the risks of slips,
trip, falls and collisions by ensuring vehicle paths and pedestrian
walkways in your facility are always clear and safe.
OHS scenarios – May be rule on your OSH platform

• The Vehicle Controls model can be used to

manage your workplace traffic safety by
ensuring safe operation of vehicles at your
workplace.
• The Behavioral Safety model can be used to
avoid line of fire injuries such as pedestrians in
vehicle paths, nonuse of handrails, and the
violations of working at heights.
 OHS scenarios – May be rule on your OSH platform
• The Pandemic Control Measures model can be used to prevent
contagious diseases by ensuring social distancing as well as proper
mask use.
• The Emergency Alerts can be used to get immediate notifications
in situations such as when a worker falls on the ground or there is a
fire-related emergency.
• What’s more, with Ergonomics AI, system detects risky actions
for the body and prompts the managers to take the situation under
control before any hazardous conditions arise.
• A variety of use cases can be set up as a rule under Ergonomics AI
feature to detect unsafe body movements, which are Trunk
Position, Trunk Twisted Position, Trunk Side Bending Position,
Lower Arm Position, Upper Arm Position, Neck Position, Neck
Bending Position, and Arm Working Position.
 AI – OSH Report Writing
• Description of incident will be analysed based on
input data and action taken report will show the
quality of report.
• Example – A employee has had a fall.
• AI may ask data – 1) How far in m or ft ?
2) Injury type ?
• Action taken – 1) First-aid was administered
2) He was shifted to Occupational Health Centre
3) Post basic treatment was shifted to hospital for
investigation.
• AI Response – 1) Very Weak 2) Good 3) Excellent
 Thank you

'Our Aim – Zero Harm’

[email protected]
7003008407