Modelling and results

The thesis work has three main objectives. From the distribution system operators perspective, the main objective is to investigate how the technical and non technical features differentiate among DSO networks in Europe. From the modelling perspec-tive, the second main objective is firstly to define a method which incorporates the previous findings to properly design a tool able to reproduce representative urban networks, and secondly to validate them through a statistical methodology. From the electric vehicle’s infrastructure perspective, the third main objective is firstly to understand the electric vehicles demand behaviour and develop models capable of reproducing them, and secondly to assess, through a methodology, the EVs charg-ing infrastructure features and performances.

The main findings regarding the DSO level analysis are summarised in the fol-lowing list.

• The technical aspects concerning DSOs derived from the survey were analysed.

– A clear picture of distribution grids is presented by gathering and cluster-ing network structure, network design, reliability indexes and distributed generation indicators. Large ranges have been observed for several indica-tors that highlight the differences among the existing DSOs in designing and operating their grids.

– Combined with the literature review, the DSO Observatory project find-ings confirms the importance of providing an open data platform due to the limited existence in the number, size and geo-referenced network grids.

• The smart grid dimension describes the new challenges threatening the DSOs.

– The remuneration tariffs of DSOs in most of the cases are based either on energy consumption (volumetric) or on contracted power (capacity based). Those ranging from purely volumetric to purely capacity based ones. Indeed, in the new emerging scenario, the different approach might play a relevant role in facilitating the distributed generation integration.

– The data exchange interface between DSOs and transmission system oper-ators (TSOs) highlights once again complete different approaches in which communication and data sharing occur either frequently (such as real-time) or only in case of urgent situation or when mandatory. This reluc-tant behaviour can represent a barrier towards an overall improvement of the electricity market. Indeed, a more transparent and frequent share of data can potentially offer new opportunities for final consumers.

– The development of demand response and demand side management pro-grams are implemented only by few DSOs. At the moment, only few consumers have financial benefits in accepting a shifting request from a DSO.

– The smart metering roll-out situation has shown unexpected outcome.

Indeed, in certain Member States where the Cost-Benefit-Analysis proved negative, some DSOs decided to undertake the smart meters roll-out. This occurs because they consider it as a breakthrough technology to improve consumer’s and grid infrastructure monitoring.

The main findings regarding the modelling level analysis are summarised in the following list.

• The development of a Distribution Network Modelling platform aim at facing the lack of available synthetic distribution networks.

– DiNeMo is the new platform that aims to provide stakeholders in the elec-tricity sector with a solid tool based on real data and capable to reproduce the representative distribution grid of a given area of interest.

– The construction of the distribution network grids is based on several net-work indicators built upon the DSO Observatory survey. The layout is geo-referenced and takes into account the geographical location of build-ings, street topology and environment aspects.

– DiNeMo output are diverse, going from consumers and substations shape-files to MATPOWER shape-files. All these information can be utilized to per-form different scenarios, such as testing the maximum installation of pho-tovoltaic panels within the observed urban area.

• A statistical methodology to validate representative distribution networks grid has been developed.

– The reliability of the network developed by DiNeMo is verified through a statistical methodology developed in collaboration with European DSOs.

It is based on 10 selected indicators calculated from DiNeMo, which are compared with the real ones provided by the DSOs.

– The results show that DiNeMo is capable of designing with good accuracy both urban and semi-urban areas of interest below a HV/MV substations.

This is possible with few inputs provided by the user and the DSO’s data collected from the DSO survey.

The main findings regarding the electric vehicle’s infrastructure level analysis are summarised in the following list.

• The study conducted on the a large dataset of electric vehicles charging columns (2900) show interesting outcomes including energy demand, in and plug-out behaviour.

– Among the 30.000 EV users utilizing the Netherlands charging columns infrastructure, 25% of the energy demand is supplied in the weekend.

Moreover, daily plug-in and plug-out distribution profiles highlighted re-markable differences among weekdays and weekends.

– Multi-modal probability distributions were identified fora number of rele-vant variables, and were handled through a Beta Mixture Model approach.

A statistical analysis of connected, idle and charge times provided the fol-lowing results: 50% of the recharges last for less than 4 hours; the idle time depends on the geographical location of the charging station, and on average it lasts also for 4 hours.

• A second analysis on the Dutch dataset has been used to define a methodology, composed of eight indicators, allowing a comparison among EV public charging infrastructures.

– The analysed database reveals a low energy use ratio and high availabil-ity of the infrastructure, which consequently indicates a low energy flows through the network. Thus causing a low allocation of the carbon in-tensity, which reflects into a higher value compared to the one found in literature.

– The correlation between fuel station, parking lots, and population density indicates a moderate to strong parameter, which indicates that the net-work distribution is appropriate. The idle time revealed that the netnet-work could be over dimensioned, which may sound reasonable at the early stage of EV adoption. There is a moderate charger and geographic concentra-tion in terms of energy demand.

On the whole, the results from this thesis show that the increasing attention toward the distribution sector should not be underestimated by the main actors, which appears to have complete different approaches in terms of smart grid projects. It

is urgent for policy makers and stakeholders involved to align DSOs to a common strategy to tackle the introduction in the distribution network grids of new players.

The DiNeMo tool may be used to perform preliminary research studies concerning the installation of new charging infrastructure, renewable energy installation or network reinforcement analysis.