simple-beauty.css

The Dyson blog: Artificial and Human Intelligence

2025-10-08 18:03:27

‘The radiology is very close by.

This building contains hundreds of thousands of samples, stored in vats of liquid nitrogen at -278o centigrade.Currently, accessing the samples is a manual process which exposes the users to risk of asphyxiation due to their close proximity to the liquid Nitrogen.

The Dyson blog: Artificial and Human Intelligence

As the Nitrogen evaporates, a blanket of cold Nitrogen forms on the floor and rises upwards, thereby excluding the oxygen necessary for human life.. For the new facility, they wanted to make sure no-one had to go near the storage vessels.So they commissioned a US company called Brooks to design a robot that could collect samples and deliver them to the scientist, in a way that’s a bit like a vending machine.Our challenge was to design and build the building around the robot system which was a world first and still in development.. Additionally, the way Bryden Wood have designed the extraction systems is a UK - and probably world first.

The Dyson blog: Artificial and Human Intelligence

Because these rooms are highly hazardous, the air change rate must be high in volume, but it must also be at low level physically.In every other facility I know of, extracting at low level is done with ductwork.

The Dyson blog: Artificial and Human Intelligence

It's inefficient because you can only put ductwork in certain areas, as people cannot be required to step over it as they move around the space.

Worse, because of the way ductwork is made, the extraction actually takes place about 15cm from the floor, which would allow the cold nitrogen to completely miss the extraction and bleed out into the room..In one recent project, we worked on the quality assurance (QA) elements of our client’s operation.

QA covered four distinct areas of manufacturing, both primary and secondary, and small and large module, split across two campuses and 13 separate testing laboratories, each conducting variety of tests and other QA processes It is not hard to imagine the complexity that engendered.And while many people had an understanding of parts of the process, no-one had a complete understanding of the whole picture.. We gathered, consolidated and agreed enormous amounts of site and process data with the client and then assessed, in a variety of ways, each of their laboratories.

We gave each laboratory a consolidated, weighted score based on their effectiveness and considering any known issues (always using visualisations and the agreed common language).. We produced visual analytics of the entire web of processes on site in a way that was clear and, as a result, very powerful: it gave the client the tools to be able not just to see and understand their complex processes in their relative context, but also to discuss them with each other (regardless of specialism and teams) and senior management.. We could then map this analysis against a range of desired objectives and value drivers, to describe dependencies, adjacencies and requirements, and how to be able to measure outputs.We presented a wide range of variables, for example: density of operations in laboratories by m2; activity in terms of people per m2; test time by laboratory and category of test; laboratory capacity by time taken per test, and by number of tests carried out per year; and so on..