Moxy demonstrates how an open, collaborative approach to modelling can reveal insights into the future of the NSW Electricity Network.

Sydney Uni, with support from Delta Electricity have embarked on the Energy 2050 Programme to provide insight into how Australia’s Electricity Network will evolve over the next 40 years in response to changes in demand patterns, introductions of new technologies and the impacts of carbon prices. 

The Programme seeks build an expert system that can help industry make long term investment decision despite the uncertainties associated with market price, changes in regulatory environment and the impact of emerging technologies. 

The Programme differs from traditional modelling endeavours that are by enlarge economic in nature and so divorced from the physical reality of the existing energy network. Such traditional models take no account of geography, climate or security of the network. These models are typically closed, where only the findings are made public and as a result they do little to promote consensus or understanding. Existing models are essentially static, that cannot be explored or expanded by users to test future scenarios in a practical way. 
 

• Will the Mandatory Renewable Energy Targets (MRET) provide affordable, competitive and reliable 24/7 electricity supply in 2020? Will it deliver the expected environmental outcomes? 
• What is the impact on state revenues of a shift towards renewable energy sources?
• What would be the impact on the grid if non-competitive energy intensive industries were to be shut down?

As the first step of the Programme, Sydney Uni commissioned Moxy to demonstrates how our Smart Wiki platform would allow such a complex model of the industry to be created in an organic, bottom up manner. Given the broad scope of the Programme, the modest initial funding and the dynamic pace at which the industry is changing, a traditional top down - requirements, design, build and verify  approach to model building was not going to produce useful insights. Instead the team elected to pursue a more open source approach to development where different pieces of analysis could be conducted in isolation and combined via the Smart Wiki. By integrating these different pieces of work, the team would progressively be able to answer question of higher and higher complexity.

This bottom up approach to modelling complex systems allows the team to focus on analytical problems that yield immediate benefits. Individual projects can be executed in order to meet the short term objectives of different clients. By using the Smart Wiki that effort is captured in a growing semantic knowledge base which allows the team to respond to bigger analytical challenges over time.

Proof of Concept

The initial proof of concept was to use the Smart Wiki to perform load flow analysis on the NSW electricity network. Such numerical analysis establishes a physical basis for how power will flow around the system and can be used to determine if the network is stable given particular supply and demand for electricity. The model is then be used to evaluate different scenarios, such as line outages or power station shutdowns and can calculate which if any power lines will be overloaded or which substations will operate outside their safe limits. For example one scenario shown below, envisaged was of a sudden outage of the Liddell Power Station, located near Mussellbrook to the north east of Newcastle and as can seen from the graphic, it results in a drop in bus voltages far to the south, in Kangaroo Valley.
 

Load flow analysis of NSW Electricity Network with an outage at Liddell Powerstation

The computations are performed using Matpower – which is a software extension for Octave, a Matlab clone. The Octave scripting language is integrated into the SmartWiki, which allows analysts to create Wiki pages that contain Octave script. The Wiki make these script available over the web and provides version history. When these pages are saved, the computer evaluates the script and immediately publishes the results to the web page which means that any user can display the results via a web browser. The code that generates these results is also shared via the website and anybody can take the script, modify it  and evaluating it in Matpower.


The Smart Wiki not only holds the mathematical scripts, but also the data on which the scripts operate. Information on each page of the wiki is semantically tagged which means that a network of pages and properties is formed. This network can be expanded by users as required by the needs of future analysis and can be queried dynamically by the mathematical scripts using a query language call SPARQL. SPARQL is conceptually similar to SQL which is commonly used to access information contained within relational databases. Whereas SQL is convenient for querying tabular data, SPARQL is used to query data that is arranged into a network, like the pages and properties that exist in the SmartWiki.

Further Analysis
Having demonstrated that the SmartWiki can be used to represent the knowledge necessary to perform load flow analysis on the NSW Electricity Network and that the computations can be performed in the SmartWiki environment, the next step is to expand the number of concepts that the SmartWiki understands. Possible avenues of exploration are:

• Evaluate different future configurations of the network, particularly around the deployment of large scale renewable energy projects and the impact that such intermittent power generation will have on network reliability.
• Expand the understanding of how electricity supply and demand varies throughout the year and in response to changes in weather. How does the weather impact on network stability as the proportion of renewable energy increases?
• Incorporate economic modelling of electricity prices and tariffs.
• Expand network to national level and incorporate gas and petroleum flows.