Home Features Health, Safety and Environmental Management: Applying Data Science and AI

Health, Safety and Environmental Management: Applying Data Science and AI

by Brian Sims

Effective health, safety and environmental (HSE) management is more crucial than ever, observes Ran Merkazy, and especially so in the high-risk, capital-intensive sectors including energy and oil/gas through to transportation. As legislation becomes more stringent, and public expectations heighten, we’ve seen that a single HSE incident can disrupt business operations and even damage corporate reputation.

The issue is less about the lack of safety data, as this is ever-growing in a digital age. Rather, the focus is on how to best use the plethora of information that’s frequently gathered. Applying data science and Artificial Intelligence (AI) points the way forward, helping HSE managers to see beyond current practices, gain greater insights and move towards adopting leading-edge systems.

HSE incidents were traditionally logged on paper, meaning that data existed, but not really in a useable format. As digital tools for managing HSE issues have evolved, the profession has moved towards a ‘transitional maturity’ stage. In the best case scenarios, digitised capture systems and supporting HSE procedures are applied in a structured rather than an ad hoc way, giving organisations the assurance that critical risks are being effectively recorded and better managed.

These new HSE engagement programmes are generating vast amounts of disparate HSE data, all of which needs to be reviewed and analysed. What’s more, the level of data is set to grow. The increasing role of employees in shaping approaches to risk assessment and risk management is a future pattern. To meet HSE goals, as well as regulatory compliance requirements and wider business improvement aims, organisations are adopting a human-centered approach at a corporate and strategic level. This means there will be greater input and feedback from personnel to contribute to HSE analytics and inform risk trade-off decisions. In short, there’ll be even more data.

The evolution of HSE management

The evolution of HSE management

Handling all of this information is proving problematic, but there are solutions ahead. Advanced technology is already emerging to help in innovative new ways, unlocking insights and creating value for businesses. This is an ongoing journey. Before we look forward, though, it would be worth highlighting the missed opportunities to reduce HSE risks and the common challenges faced by professionals operating in the field today.

Missed opportunities

While most companies capture HSE incidents, many struggle to absorb and understand the often unorganised, but valuable data in their possession. Emerging risks and safety issues are missed, in turn realising a lack of understanding of the hazards personnel face at work on a daily basis.

Hard, data-based and actionable insights remain hidden. The opportunity to put in place HSE strategies, prevention plans and mitigation activities in order to prevent common issues from being repeated is lost. This could be described as operating with a ‘rear view mirror’ mindset instead of looking ahead to identify upcoming risks and evolve existing HSE systems to address those challenges.

Currently, most organisations have too much HSE data and too little time to investigate it. The wide range of data sets generated by workforce safety reports, near miss reports, permits to work, job hazard analysis, audits, inspections and observations are rarely integrated. There’s no single repository to mine for HSE insights. In fact, often the most useful information stays locked away in the free-text description of the incident where the clues lie as to the direct and root causes of the incident.

Without an integrated HSE source, analysing safety data is labour intensive and highly inefficient. For these reasons, we estimate that, even when available, over 75% of company data cannot be easily analysed. Who has the necessary time and resources in the busy modern world of business?

The process of extracting insight from data

The process of extracting insight from data

This challenge is exacerbated by three further factors. Companies operating in high-hazard industries have HSE procedures in place, but they tend to be individual ‘siloed’ procedures isolated from one another. This makes it extremely difficult to see the bigger picture across the entire organisation, whether behaviourally, geographically or historically speaking.

Some organisation can also experience difficulties in standardising, collecting and digitising safety data.

Third, data gathering is often not targeted correctly to detect and monitor significant risks and draw out game-changing insights. Measuring low-risk activities, for example, will add to the data pile, but it’s unlikely to be useful in preventing the next incident.

Moving to a leading-edge HSE system

Organisations can now use advanced technology solutions to leverage the power of their data, creating a leading-edge and predictive approach towards HSE management. Rapid progress in data science, powerful computing and enhanced algorithms – including machine learning and AI – are all playing their part in this evolution. Data-driven insights are delivering a new HSE era.

Through digitisation, organisations can seamlessly integrate separate data sets across all sorts of formats into a ‘data lake’: a storage repository of vast raw information, retained in its native format until needed. With integration, organisations can also draw additional insights from their existing organisational data.

Richer context, beyond traditional reporting techniques, enhances risk prediction. Sensor technology is introducing informative new images to HSE management and enabling body-condition monitoring of people, equipment and assets. These developments are helping to provide actionable and often real-time insights into HSE risks and incident management.

For a while now, Lloyd’s Register has been testing a variety of new technologies through its innovation teams and SafetyTech Accelerator. This has produced a number of new partnerships between Lloyd’s Register and cutting-edge providers of technology in the area of video hazard analytics, PPE monitoring, locations and vital signs monitoring.

Advanced data and digital tools enable organisations to identify and analyse the direct causes of incidents at scale, presenting a clearer picture than ever before of where to focus HSE efforts. Tools include visual recording and sensor systems, as mentioned above. Fresh insights gathered are helping to identify issues that were previously unknown and prevent safety risk incidents from happening.

Accidents due to human error

Using AI for HSE management

Using AI for HSE management

Up to 80% of accidents result from human error. However, human behaviour has been hard to measure to date. Now, AI is being used for advanced text analytics. Lloyd’s Register has recently developed LR SafetyScanner, a Natural Language Processing (NLP) technology system which automatically generates categories of hazards using this approach.

A sub-field of AI, NLP enables computers to understand and process human language instantly. Our Hazard AI engine has been trained on 24 hazard categories, instantly scanning all text data to generate a new and bigger risk picture for organisations. The results reveal previously hidden insights.

Along with text analytics, vision and vital signs analytics are also being integrated. Human interactions and negative physical conditions, such as fatigue and heat exhaustion, can now be understood in the context of risks and incidents. A previously unobtainable detailed level of insights is helping to enhance HSE strategies and processes from a very human perspective.

New dashboard technologies are making it easier to identify, investigate and communicate key insights, informing better decision-making and building greater awareness of critical HSE issues. Introducing advanced analytics to the mix gives HSE professionals a powerful tool at their fingertips to spot emerging risks and patterns.

Transformational path

Ran Merkazy

Ran Merkazy

For many organisations and sectors, HSE management is already on a transformational path. While digital reporting and processing of safety data is increasing, activities remain unconnected from one another. By adding advanced data science and AI to HSE management, organisations can bring all of the relevant information they need together in one place. With a single view of HSE matters, priorities can be identified and efforts focused in the right areas, reducing incident rates and providing assurance that critical HSE risks are effectively recorded and well-managed.

The adoption of technology will be critical in creating a safer workplace, but we should not neglect the importance humans also play here. Internal training and education programmes remain equally fundamental.

When reporting an incident, for instance, we ask employees to categorise the type of incident they encountered, helping our specialists to analyse the event and run checks for trends. Around 20%-40% of reports are tagged under ‘other’ as its hazard category, making it difficult to understand what’s actually being reported. Even AI will struggle to make sense of ambiguous data.

Ran Merkazy is Vice-President of Products and Services Innovation at Lloyd’s Register

You may also like