Network Event and Alarm Transparency (NEAT)

Objective(s)

The objective of the project are to;

• understand the data that can be used to provide context for the alarm and event analysis
• assess the quality of this data and where possible correct quality issues
• derive preliminary insights from the data to inform the selection of models
• carry out advanced analytics to understand the relationships between alarms and events in different systems and the external datasets
• create a prototype dashboard to allow the analysis to be run regularly and present the user with results
• trial the use of the dashboard using real data over a period of time
• consolidate and share the learning from the project

Problem(s)

New DSO related control systems, such as Active Network Management (ANM) and System Voltage Optimisation (SVO), are increasing the number of alarms and events that need to be managed in the control room. This is expected to increase further as new systems are developed to support new DSO functions. Therefore we expect both the types of events and number of events to increase as systems make greater use of new data in-feeds as well as existing SCADA inputs. The underlying cause of events is sometimes difficult to ascertain and where systems interact with each other, understanding the relationships between their events and alarms will reduce the time spent supporting new DSO related systems and ensure that they can operate optimally.

Method(s)

The project will develop a tool to analyse the relationship between alarms and events within and across the control room system and those for ANM and SVO. This will be sufficiently generic that future systems can be included and analysed in the same way. It will have functions to highlight relationships helping identify common causes that can then be used to put remedial actions in place, which will include automatic grouping of alarms to simplify their management.

• Scope

  The investigation will consider how the alarms in the new ANM and SVO systems relate to the alarms within PowerOn or to each other. However, the approach taken and prototypes developed will be generalised as far as possible so that it could be applied to future systems yet to be developed, or systems in use by other DNOs.

  The analysis will result in rules or information that can be used to understand root causes for alarms and events allowing the system support staff to take steps to reduce their frequency. For the avoidance of doubt, the existing alarm and event management facilities within PowerOn already provide a number of features to simplify and prioritise the information provided to the control engineers and the project does not intend to duplicate these.

  The project approach will contain six distinct work packages:

  WP1. Specification – Gather detailed business domain knowledge to ensure a solid understanding of the National Grid Electricity Distribution (NGED), formerly WPD, network region, business processes and availability of data.

  WP2. Design – This work package will look into the design of the NEAT platform through a data quality assessment, data analysis and identification of advanced analytical techniques. This will cover checking the data is complete and whole, identification of outliers or other logical inconsistencies and ensuring that the joining of datasets does not lose information or incur further data issues. An initial phase of data analysis will be used to derive preliminary insights and observations. This may involve the use of statistical models but primarily aims to identify characteristics of the data set and inform the formal modelling task.

  WP3. Build – During the build phase the advanced analytical techniques and dashboard development will be completed. Depending on the insights identified during the initial data analysis during the design phase these may include; clustering analysis, network association analysis, spatial analysis where geographical information is present, time series analysis, and predictive forecasts. The analytical techniques will be integrated into the NEAT dashboard and will also include the development of stage that load data from the databases into the analytic tool to be formalised and available to run on a regular basis.

  WP4. Deploy and Test –Selection and integration of the project into the NEAT dashboard. This stage also requires the scripts that load data from the databases into the analytic tool to be formalised and available to run on a regular basis as well as user acceptance testing.

  WP5. System Trial & Analysis The prototype dashboard will undergo further development during a live system trial to extract insights from the alarm and event analysis. This will include reviewing trends during the trial period to look for results improvement and learnings.

  WP6. Dissemination and Closedown Report – This will cover the findings in terms of alarm relationships as well as trends from the ANM and SVO alarm systems. This will highlight the benefits and improvements possible through better understanding of the interactions between different events and improving the analysis from control room operators.

• Success criteria

  The success of the project would be indicated by the following outcomes.

  • The project has gained an understanding of the new types of alarms and events associated with new systems supporting DSO functions, how these differ from “traditional” control system alarms and how the needs for their management differ to traditional alarms and events.
  • The data sources that are available to contextualise these alarms have been explored and relationships between data items and the alarms and events have been found.
  • The learning gained from analysing relationships between alarms and events and the contextualising data has been incorporated into a dashboard which is suitably generic in design to facilitate future systems to support DSO functions that are not yet known.
  • A dashboard to assist with the management of these alarms and events has been developed and tested. Learning from the trial has been used to recommend changes to datasets, processes, systems etc. to reduce alarms and events in the future and/or the way in which alarms and events are managed has been improved to reduce the time spent on their management.
• Potential for new learning

  This project has potential for new learning as the processes to correlate alarms and events with new systems to support DSO functions have not previously been developed. The systems that are raising these alarms are new and resulting in more complex and unfamiliar alarms for the control engineers. They are also not trivial, as an in-depth investigation of the system is required to identify the events that led to alarms before they can be resolved.

  The solution proposed as part of this project will aid the processing of these alarms in two innovative areas:

  • Use of analytical modelling methods, such as Bayesian Analysis, Machine Learning and other statistical techniques to develop clustering, predictive and inferential models. These methods will seek to describe the correlations and causation between events, data inputs, and resulting alarms, so they can be understood and resolved.
  • State of the art dashboarding visualisation for the analysis, processing, and presentation of alarms to the control engineers and more detailed investigation teams.
• Documents and links

  Registration and PEA Document 

  Annual Progress Report 2022

  Design 

  Trials and Analysis Report

  Dissemination Webinar Slides

  Project Closedown Report 

  ENA Project Closedown Report

• Monthly updates

  • {{update.title}}

  Show more