Skip to content

grizonic/model_based_optimization_power_network

Repository files navigation

The purpose of this repo is mainly to host this code which is part of my graduation report. Therefore, it will probably be quite incomplete for those trying to understand the point behind it. If you're interested in oil-impregnated insulation, power transformers, decentralized control... feel free to contact me.

Framework of Model-Based Optimization applied to the Electrical Power Network

The advanced ageing of the electrical power network combined with the inevitable increase of renewable energy utilization and the constant increase of power consumption over the last decades call for a change in the current power network paradigm. Electrical world stakeholders have to embrace the situation and come up with methods of better utilizing the existing, but also new, network resources. The ongoing financial crisis also behaves as an eye opener for new concepts and smart opportunities.

This project’s approach is to create a modern and automated layer on top of the existing power network physical layer which makes use of several physical models to interpret the current and future expected health status of the network’s components. With this automated layer, the network operator can configure the network to be a self-sustainable, or at least more independent, and a more insightful system. Real-time and projections of the network’s health status can be used as power flow optimization factors.

Simulations based on real power flow data have shown that distributing thermal loading throughout neighbouring power components results in a more efficient resource utilization as it lowers the overall network accelerated ageing factor, while keeping the power flow characteristics within the legislated integrity limits.

It has been shown that, by applying this health prediction framework to the IEEE-14 bus network and allowing the layer to act in a decentralized fashion independently of the network operator, the overall network lifetime utilization can be reduced by 80% by rerouteing only 13% of the overall power flow. This rerouteing of power flow does not compromise the load requirements, nor it uses more power components than the ones present in the IEEE-14 bus network configuration.

Although it is difficult to implement this framework layer on top of the electrical power net- work, due to its well-known structural inertia, it has been shown that this system can effectively be deployed in different steps in time. Furthermore, the level of trust the network operator has in the system can also be incrementally upgraded, as this layer can be used at first simply as a more insightful source of data and only when desired it can be used as a more active factor in the electrical power network management.

Contents

Tettex 2840 Script:

A simple python script that translates the raw xml outputted by the Tettex 2840 and produces quite costomizable csv files.

Single Agent Optimization:

With an ugly Matlab GUI created in GUIDE and using real power flow data, you can variate several inputs to analyse the behaviour of a single agent built on top of a power transformer;

IEEE 14 network simulation:

Agents are placed on all three transformers of the network. The network is transformed into a self-sustainable mesh that tries to distribute thermal loading throughout the less loaded nodes in order to avoid the steep lifetime consumption of highly thermally loaded transformers.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published