Approaches to the Design of Internal EHV and HV Networks for Large Data Centres – Three Case Studies

For most data centre developers owning EHV infrastructure is new, as are transmission connections at EHV with the additional complexity and compliance requirements these bring.

Across the large number of data centres we develop grid connections and internal EHV and HV networks for, we have observed how different design philosophies can lead to different outcomes. The design philosophies are driven by redundancy requirements, particularly that some collocated data centres mimic regulated network principles while others use distributed redundancy schemes. Here we focus on three of the larger data centres, collocated and enterprise types, in three countries, owned by three different organisations, and two design approaches.

Top-down approaches design the infrastructure at each level as it would be for regulated networks. Bottom-up approaches, combine a number of standardised data centre blocks and the

HV and EHV network is then developed above that.

The top-down approach is better suited to large data centres with significant self-generation as it allows the central infrastructure to be created in a coordinated manner, while managing equipment thermal and short-circuit ratings. The bottom-up approach is better suited to smaller data centres that do not require multiple internal voltage levels above the module level and the risk of high fault levels in these designs must be recognised from the start. It is also becoming increasingly common for existing generation sites to be reconfigured to directly supply data centres. This augmentation largely follows the top-down approach as the centralised infrastructure already exists as well as secure utility connections.

We have observed challenges with both approaches and have key considerations for designing these large connections, while retaining the modular approach of small data centres as far as possible. Key considerations should include:

•

•

•

• Grid-code compliance: establish the full code requirements for the facility and roll down the requirements to end-users.

End-user agreements: agree availability and redundancy arrangements at the outset and roll down of the grid code requirements.

Equipment ratings: establish if pre-approved equipment is to be used and what range of ratings are available.

Early execution of studies: the thermal and short-circuit rating requirements should be considered for early concepts, together with key grid code requirements.