Introduction
UK Power Network maintains the 132kV voltage level network
and below. An important part of the distribution network is the stepping down
of voltage as it is moved towards the household; this is achieved using
transformers. Transformers have a maximum rating for the utilisation of these
assets based upon protection, overcurrent, switch gear, etc. This dataset
contains the Grid Substation Transformers, also known as Bulk Supply
Points, that typically step-down voltage from 132kV to 33kV (occasionally down
to 66 or more rarely 20-25). These transformers can be viewed on the single
line diagrams in our Long-Term
Development Statements (LTDS) and the underlying data is then found in the LTDS
tables.
This dataset provides half-hourly current and power
flow data across these named transformers from 2021 through to the previous
month across our license areas. The data are aligned with the same naming
convention as the LTDS for improved interoperability.
Care is taken to protect the private affairs of companies
connected to the 33kV network, resulting in the redaction of certain transformers.
Where redacted, we provide monthly
statistics to continue to add value where possible. Where monthly
statistics exist but half-hourly is absent, this data has been redacted.
To find which transformer you are looking for, use the ‘tx_id’
that can be cross referenced in the Grid Transformers Monthly Dataset, which describes by month what
transformers were triaged, if they could be made public, and what the monthly
statistics are of that site.
If you want to download all this data, it is perhaps more
convenient from our public sharepoint: Open Data Portal Library - Grid Transformers - All Documents (sharepoint.com)
This dataset is part of a larger endeavour to share more
operational data on UK Power Networks assets. Please visit our Network Operational
Data Dashboard for more operational datasets.
Methodological Approach
The dataset is not derived, it is the measurements from our
network stored in our historian.
The measurement devices are taken from current transformers
attached to the cable at the circuit breaker, and power is derived combining
this with the data from voltage transformers physically attached to the busbar.
The historian stores datasets based on a report-by-exception process, such that
a certain deviation from the present value must be reached before logging a
point measurement to the historian. We extract the data following a 30-min time
weighted averaging method to get half-hourly values. Where there are no
measurements logged in the period, the data provided is blank; due to the
report-by-exception process, it may be appropriate to forward fill this data
for shorter gaps.
We developed a data redactions process to protect the
privacy or companies according to the Utilities Act
2000 section 105.1.b, which requires UK Power Networks to not disclose
information relating to the affairs of a business. For this reason, where the
demand of a private customer is derivable from our data and that data is not
already public information (e.g., data provided via Elexon on the Balancing
Mechanism), we redact the half-hourly time series, and provide only the monthly
averages. This redaction process considers the correlation of all the data, of
only corresponding periods where the customer is active, the first order
difference of all the data, and the first order difference of only
corresponding periods where the customer is active. Should any of these four
tests have a high linear correlation, the data is deemed redacted. This process
is not simply applied to only the circuit of the customer, but of the
surrounding circuits that would also reveal the signal of that customer.
The directionality of the data is not consistent within this
dataset. Where directionality was ascertainable, we arrange the power data in
the direction of the LTDS "from node" to the LTDS "to
node". Measurements of current do not indicate directionality and are
instead positive regardless of direction. In some circumstances, the polarity
can be negative, and depends on the data commissioner's decision on what the
operators in the control room might find most helpful in ensuring reliable and
secure network operation.
Quality Control Statement
The data is provided "as is".
In the design and delivery process adopted by the DSO,
customer feedback and guidance is considered at each phase of the project. One
of the earliest steers was that raw data was preferable. This means that we do
not perform prior quality control screening to our raw network data. The result
of this decision is that network rearrangements and other periods of non-intact
running of the network are present throughout the dataset, which has the
potential to misconstrue the true utilisation of the network, which is
determined regulatorily by considering only by in-tact running arrangements.
Therefore, taking the maximum or minimum of these transformers are not a
reliable method of correctly ascertaining the true utilisation. This does have
the intended added benefit of giving a realistic view of how the network was
operated. The critical feedback was that our customers have a desire to
understand what would have been the impact to them under real operational
conditions. As such, this dataset offers unique insight into that.
Assurance Statement
Creating this dataset involved a lot of human data
imputation. At UK Power Networks, we have differing software to run the network
operationally (ADMS) and to plan and study the network (PowerFactory). The
measurement devices are intended to primarily inform the network operators of
the real time condition of the network, and importantly, the network drawings
visible in the LTDS are a planning approach, which differs to the operational.
To compile this dataset, we made the union between the two modes of operating
manually. A team of data scientists, data engineers, and power system engineers
manually identified the LTDS transformer from the single line diagram,
identified the line name from LTDS Table 2a/b, then identified the same transformer
in ADMS to identify the measurement data tags. This was then manually inputted
to a spreadsheet. Any influential customers to that circuit were noted using
ADMS and the single line diagrams. From there, a python code is used to perform
the triage and compilation of the datasets.
There is potential for human error during the manual data
processing. These issues can include missing transformers, incorrectly labelled
transformers, incorrectly identified measurement data tags, incorrectly
interpreted directionality. Whilst care has been taken to minimise the risk of
these issues, they may persist in the provided dataset. Any uncertain behaviour
observed by using this data should be reported to allow us to correct as fast
as possible.
Additional information
Definitions of key terms related to this dataset can be
found in the Open
Data Portal Glossary.
Download dataset information: Download dataset information: Metadata (JSON)
We would be grateful if you find this dataset useful to
submit a “reuse” case study to tell us what you did and how you used it. This
enables us to drive our direction and gain better understanding for how we
improve our data offering in the future. Click here for more information: Open Data Portal Reuses — UK Power Networks