Digitally Enabled Process Safety Management for Chemical Plants
Manoviram Rath & Rajesh Thilak
L & T Technology Services Limited,
Vadodara India

The author, in this article, addresses how to digitally enable the whole Process Safety Management and Risk-based Inspection Process throughout the life-cycle of the plant to help in learned decision making using all the past data.

The importance of a safe process design cannot be emphasized enough in process plants across the world. Thanks to decades of learning from chemical plant disasters, the design of process plants has become safer and more reliable to operate and maintain.

Relevant international standards have been frequently updated to incorporate such learning and to reflect the need of the hour. Every new process plant design incorporates such guidelines through the Operational Risk Assessment (ORA) process, the Hazards and Operability Study (HAZOPs), and the associated risk mitigation design plans as a part of an elaborate Functional Safety Risk Management plan. Thus, a safe plant gets designed, erected, and commissioned.

However, after years of regular operation of the plant, it becomes a colossal task to retrieve all maintenance and inspection reports, design documents, and the decisions that were made based on multiple intermediate studies during the intervening period. The common practice is that the learning during the period applicable to all critical systems and chemical processes are lost in most instances. Every new plant-based modification or any critical equipment inspection plan needs to undergo the process without incorporating any of the previous learning or incorporate some of the available learning.

Elements of Process Safety Management:
Process Safety Management is a proactive and systematic approach to identify, evaluate, mitigate, as well as to prevent potential hazards resulting in fires, explosions, chemical reactions, or chemical releases that occur because of failures in process, procedure, or equipment. This needs to be operational throughout the plant's life-cycle. The following minimum needs are to be established for an effective Process Safety Management.
  • Collection of Process safety/Process critical information
  • Process Hazard Analysis
  • Mechanical Integrity Evaluation
  • Design and Selection of Protective Devices and Controls
  • Formalizing Safe Work Practices
  • Operating Procedures
  • Emergency Response Procedures
  • Trainings for employees and contractors
  • Pre-Start up Safety Review
  • Accident / Incident Investigation
  • Regular Audits
  • Management of Change
  • Maintaining Records at Each Stage
  • Carrying out Process Safety Restudies at Regular Interval
Process Safety Information Governance:
Organizational procedures still follow traditional Process Safety Governance. The method relies entirely or partially on the aforementioned elements. Data Management Systems are critical to information governance, while the enforcement and monitoring is done manually.

Risk Based Process Safety (RBPS) Design and Improvement Criteria : Organizations require robust process safety management systems, and to build & operate them, they must take the RBPS approach. In order to make best use of this approach, organizations must follow the guidelines of designing and improving the various process safety activities to ensure effort appropriation. The approach depends on detailed process safety prediction on the basis of three crucial factor:
  • Requisite understanding of associated risk on which the process safety practices are focused.
  • The correlation between the demand of process safety work and available resources.
  • The advantages and limitations of the process safety culture within the organization.
This RBPS approach has various crucial advantages. Some of them are:
  • Negating the gaps
  • Preventing inconsistencies
  • Preventing excess work, insufficient work, and rework

It is of paramount importance for the organizations to integrate their RBPS activities with the existing management systems, in order to optimize process safety. Some of these management systems cater to security, personal health and safety, product quality, equipment reliability, and environmental protection.

Risk Based Inspection
This method of assessment and inspection relies on the possible risks. Equipment/areas to be inspected are prioritized as per the risk involved. For instance, only a small number of equipment in a plant may run with considerable risks. This form of inspection focuses on higher level of coverage for high-risk items, while - for a low-risk item - dedicating only as much effort and resources as it demands. Risk-based inspection is in contrast with time-based inspection, which follows the doctrine of assessing all equipment and systems, irrespective of the risk each one of them carries , thus resulting in higher cost and effort.
However, risk-based inspection can work effectively only on the basis of large amount of historical data. In other words, to assess the risk and accordingly to run inspections & diagnosis, the availability of all relevant records - such as process safety information, failure analysis reports, inspection history, corrosion data, and incident analysis - is a pre -requisite. Furthermore, a quantitative approach towards the risk makes this inspection method more effective. Rating equipment and systems, on the basis of health status and possible failures, determines the risk categories that they fall in and are accordingly prioritized. Thorough analysis of every aspect of risk results in clearer assessment and inspection and, in turn, faster detection and addressing of the problems.

Digital Enablement
It is important for organizations to realize and move toward creating definitive business value by putting the process safety measures in place and focusing on eradicating any possibilities of injury, loss, and interruption. In the digitalized landscape of Industry 4.0, organizations can ensure optimum productivity, quality, flexibility, and scalability only by showing due respect to implementation of impermeable process safety measures and follow industry best practices. Digitalized processes demand cutting-edge technological innovations, integrated engineering, and seamless operations. Big Data has a critical role to play for company leaderships to make informed decisions.

With the help of modern technologies, due efforts are being made to overcome issues plaguing prompt decision making. Some of these issues are:
  • Lost learning
  • Misplaced records
  • Duplication of data
  • Multiple histories of records
It is important that the Operation team/HSE Manager is empowered to make informed and analytical decisions by using the consolidated history of all safety records. A digitalized approach is critical to achieve that status. This digital enablement can happen by executing the following actions:
  • Defining and designing a robust information tracking process. For that to happen, the Document Management System has to be dynamic and interactive, while also enabling User Interface customization.
  • Creating an easily accessible database that stores all safety and non-safety Records.
  • Making best use of Optical Character Recognition technology in order to capture all the relevant information and attributes.
  • Using Natural Language Processing to sort and to categorize the unstructured data.
  • Analyzing records of design data, safety incident reports, and safety records with the help of artificial intelligence(AI)/machine learning (ML).

Broad Framework
A large number of safety related documents/records are first categorized, and all the information are then further sub-categorized based on required keywords. The intelligent database of every safety device, safety instrumented function, scenario, cause and effect matrix, risk scenarios, etc are defined through various algorithms. A suitable User Interface is thus being made by enabling the Operation Manager/Safety Manager to input required data for various safety scenarios. The algorithm would respond to such inputs and provide an array of suggestions.

The AI/ML based system analyses the historical data & records, and helps in defining the risk values (in case of risk based inspection) and process safety decisions, which would very accurately map to the existing condition of the plant. While the rigors of the engineering analysis cannot be done away with, the use of AI/ML techniques would help remove a large amount of subjectivity associated in such decision-making process.

The most important benefit coming out of this governance mechanism is because of reliable and readily available safety information including historical data right from the conceptual design onwards. The management is no longer dependent on the historical knowledge of the individual operator /manager for the particular area of the plant. The amount of subjectivity while defining the risk matrix is now replaced with ML based suggestions.

Framework for Digital Enablement
A broad level framework for digital enablement for process safety management and risk based inspection plan can be defined as below.

The first step would be to identify the gaps in the as-is scenario. A team of domain experts and digital technologists would map the available information into various buckets based on whether the data are available on paper, scanned document, hosted in portal, etc. The existing process safety management process of the plant also needs to be re-mapped as to incorporate the digital enablement. The sub-processes, for which digital framework incorporation is required, are to be redefined. Making all the historical records available is the key in effective use of digital solutions. Otherwise, it would simply result into another document management portal hosting all the information in digital form.

Digital enablement in process and plant safety area is still in evolving stage in most of the chemical plants, primarily because of the safety managers' general reliance on traditional approach and also their wariness about the unpredictability of digital systems. However, the increasing expectation of management to make quick and clear decisions is slowly changing the mindset. In the coming years, we would see more and more occurrence of such systems as rule than as exceptions.