Distribution grids play an important role in delivering electricityto end users. Electricity customers would like to have a continuouselectricity supply without any disturbance. For customerssuch as airports and hospitals electricity interruption may havedevastating consequences. Therefore, many electricity distributioncompanies are looking for ways to prevent power outages.Sometimes the power outages are caused from the grid sidesuch as failure in transformers or a break down in power cablesbecause of wind. And sometimes the outages are caused bythe customers such as overload. In fact, a very high peak inelectricity consumption and irregular load profile may causethese kinds of failures.In this thesis, we used an approach consisting of two mainsteps for detecting customers with irregular load profile. In thefirst step, we create a dictionary based on all common load profileshapes using daily electricity consumption for one-monthperiod. In the second step, the load profile shapes of customersfor a specific week are compared with the load patterns in thedictionary. If the electricity consumption for any customer duringthat week is not similar to any of the load patterns in thedictionary, it will be grouped as an anomaly. In this case, loadprofile data are transformed to symbols using Symbolic AggregateapproXimation (SAX) and then clustered using hierarchicalclustering.The approach is used to detect anomaly in weekly load profileof a data set provided by HEM Nät, a power distributioncompany located in the south of Sweden.

The energy transition is accompanied by massive electrification of uses and sectors such as transport. As a result, the pressure on the electricity grid is increasing, and the time to connect to the power system is lengthening. Deploying new infrastructure is a laborious and expensive process but there are alternatives to exploit the flexibility of the power grid. The deployment of smart meters opens the door to many applications related to flexibility on the consumer side, to reduce peak loads that threaten grid capacity. Targeting the right consumers for Demand-Side Management (DSM) is a prerequisite to maximizing the chances of success of such programs. This degree project replicates and adapts the method developed in [14] to segment residential customers. It consists of encoding Daily Load Curves (DLC) using a dictionary of Typical Load Profiles (TLP) and grouping consumers according to the distribution of their TLP. A temporal analysis of the main TLP reveals different consumption behaviors. Customers are segmented into groups that reflect the degree of volatility of their consumption. This enables a classification based on the potential for Energy Efficiency (EE) or Demand Response (D/R) programs. We address the issue of attribute detection using the distribution of TLP of customers. In particular, several classification algorithms are compared to detect TLP characteristic of Electric Vehicle (EV). The obtained load shapes show consumption peaks at night, which may correspond to the charging time of EV. The method is discussed, especially the choice of the number of load profiles to be included in the dictionary of TLP. It proves to be useful to group consumers with similar consumption profiles and opens the door to applications such as individual household consumption forecasting. / Energiomställningen kräver en massiv elektrifiering av användningsområden och sektorer som t.ex. transportsektorn. Detta leder till att trycket på elnätet ökar och att tiden för att ansluta sig till elnätet blir allt längre. Att bygga ut ny infrastruktur är en mödosam och dyr process, men det finns alternativ för att utnyttja elnätets flexibilitet. Utplaceringen av smarta mätare öppnar dörren för många tillämpningar som rör flexibilitet på konsumentsidan, för att minska toppbelastningar som hotar nätkapaciteten. Att rikta in sig på rätt konsumenter för DSM är en förutsättning för att maximera chanserna att lyckas med sådana program. I detta examensarbete replikeras och anpassas den metod som utvecklats i [14] för att segmentera hushållskunder. Den består av att koda DLC med hjälp av ett lexikon av TLP och gruppera konsumenter enligt fördelningen av deras TLP. En tidsmässig analys av de viktigaste TLP avslöjar olika konsumtionsbeteenden. Kunderna delas in i grupper som återspeglar graden av volatilitet i deras konsumtion. Detta möjliggör en klassificering baserad på potentialen för EE eller D/R-program. Vi tar upp frågan om attributdetektering med hjälp av fördelningen av TLP hos kunderna. I synnerhet jämförs flera klassificeringsalgoritmer för att upptäcka TLP som är karakteristiska för EV. De erhållna belastningsformerna visar konsumtionstoppar på natten, vilket kan motsvara laddningstiden för EV. Metoden diskuteras, särskilt valet av antalet belastningsprofiler som ska ingå i ordlistan för TLP. Metoden visar sig vara användbar för att gruppera konsumenter med liknande förbrukningsprofiler och öppnar dörren för tillämpningar som prognostisering av enskilda hushålls förbrukning.

The transformation of the legacy electric grid to Smart Grid (SG) poses numerous challenges in the design and development of an efficient SG communications network. While there has been an increasing interest in identifying the SG communications network and possible SG applications, specific research challenges at the network protocol have not been elaborated yet. This dissertation revisited each layer of a TCP/IP protocol stack which basically was designed for a wired network and optimized their performance in IEEE 802.11s-based Advanced Metering Infrastructure (AMI) communications network against the following challenges: security and privacy, AMI data explosion, periodic simultaneous data reporting scheduling, poor Transport Control Protocol (TCP) performance, Address Resolution Protocol (ARP) broadcast, and network interoperability. To address these challenges, layered and/or cross-layered protocol improvements were proposed for each layer of TCP/IP protocol stack. At the application layer, a tree-based periodic time schedule and a time division multiple access-based scheduling were proposed to reduce high contention when smart meters simultaneously send their reading. Homomorphic encryption performance was investigated to handle AMI data explosion while providing security and privacy. At the transport layer, a tree-based fixed Retransmission Timeout (RTO) setting and a path-error aware RTO that exploits rich information of IEEE 802.11s data-link layer path selection were proposed to address higher delay due to TCP mechanisms. At the network layer, ARP requests create broadcast storm problems in IEEE 802.11s due to the use of MAC addresses for routing. A secure piggybacking-based ARP was proposed to eliminate this issue. The tunneling mechanisms in the LTE network cause a downlink traffic problem to IEEE 802.11s. For the network interoperability, at the network layer of EPC network, a novel UE access list was proposed to address this issue. At the data-link layer, to handle QoS mismatch between IEEE 802.11s and LTE network, Dual Queues approach was proposed for the Enhanced Distributed Channel Access. The effectiveness of all proposed approaches was validated through extensive simulation experiments using a network simulator. The simulation results showed that the proposed approaches outperformed the traditional TCP/IP protocols in terms of end to end delay, packet delivery ratio, throughput, and collection time.