Identify Regulations and Standards by the type of Equipment, using AI

By ,

Over the years we have seen organizations struggle to identify applicable regulations and standards for their operations. Most consulting companies and data providers organize regulatory information by subject areas such as Air, Waste, and Water. Documents may also be organized based on actions such as Reporting, Monitoring, and Labelling. Yet, in most cases companies do not know which subject areas they need to pay attention to or which activities they engage in. One of the few things that organizations have a strong grasp of are the physical assets they have onsite. Because you buy, lease, and track the equipment you use at your operations, you typically have an accurate list of the equipment. Why not identify your regulatory obligations based on the equipment you have on site?

As far as we have seen, there are a number of challenges with tagging documents with equipment. The first challenge is that the same equipment may be named differently in various parts of the world, and by agencies and regulators. This is why Nimonik has developed an equivalency table that allows us to detect and map equipment between jurisdictions and languages and dialects. Our work on this continues and is evolving as we detect more and more variations of terms around the world.

Another challenge is that equipment is often made up of various subcomponents and sub-subcomponents. The level at which a piece of equipment is regulated varies from jurisdiction to jurisdiction and from equipment to equipment. Air Conditioning units have subcomponents that may use halocarbons, it may be that a regulation references the subcomponent and not the actual air conditioner. Frankly, there is no easy solution to this. So far we have attempted to build a list at the level of detail that is commonly found in regulatory documents. We will inevitably miss some items and we ask users of NimonikApp to contact us with comments and ideas on our current list of equipment.

As with any terminology, we often have to deal with variations on a term – plural versions of equipment, equipment in a verb or adverb format, and other twists and turns. To achieve this, Nimonik employs some standard Natural Language Processing and Artificial Intelligence technology to dig down to the root of an equipment and search through the documents for that. This has been fairly effective, but still requires improvement.

The last challenge I would like to mention is the use of terms that are not usually used as equipment. Examples are things like, ‘face’, which in the mining industry means “the end of a drift, crosscut or stope in which work is taking place.” One could argue that even that is not really equipment, but rather a place or activity. Of course ‘face’ is also used in a lot of other contexts; from a human face to a face shield to the face of a building. Nimonik has built confidence levels for a variety of equipment terminology and has a robust quality control process to progressively improve the accuracy of our tagging system. Eventually we hope our AI will allow us to detect which ‘face’ a regulation is talking about and deliver that to the right customer, at the right time, and in the right place!

To see a full list of the Topics we have tagged Regulations and Standards with so far, please visit NimonikApp EHS Topics and take a look at Equipment.