Rethinking electrical water heaters

Nel, Philip Johannes Cornelis (2015-12)

Thesis (MEng)--Stellenbosch University, 2015.

Thesis

ENGLISH ABSTRACT: South Africa is, at the time of writing, in the midst of an energy crisis as the national utility is unable to meet the nation’s energy demands. Electrical water heaters (EWHs) remain one of the main contributors to residential energy consumption in South Africa and other countries where they are used. Although educational material has been published to create awareness of energy saving actions for EWHs, it is unclear if users understand the content and efficiently control their EWHs. Additionally, insufficient feedback of usage data makes it difficult for consumers to understand their consumption patterns and make informed decisions regarding their future water and electricity use. This work presents a mobile based eco-feedback system for the energy and water consumption data of residential EWHs. The system consists of several components: an EWH model; an event detection algorithm; and an Android mobile application. The physics based EWH model was developed in order to accurately simulate the energy input and output of an EWH for various control settings, usage profiles and orientations (i.e. vertical and horizontal). The accuracy of the model is validated against six datasets, four comprising 900 hours with multiple usage events and two with only standing losses. The results show that measured energy usage is modelled with an estimation error of less than 2% and 7% for schedule control and thermostat control respectively. As well as being accurate, the presented model has a low computational complexity, taking only 100 milliseconds to complete a 10 day simulation on a standard desktop machine, making it ideal for use in mobile devices. A novel and non-invasive hardware solution and matching algorithm were developed to support the identification and classification of warm water usage events without the use of invasive and expensive water metering technologies. The algorithm was tested using 49 days of data which included 127 usage events and was found to accurately detect usage events with an accuracy of 91%. Additionally, the algorithm was able to detect very small usage events (0.5 litres was detected successfully). However, the estimated duration of events is within 2 minutes accurate 79% of the time. Additionally, the outlet temperature and water meter data were used as inputs to the EWH model for estimating the energy consumption under various control settings. The outlet temperature data was used to estimate both the total volume of warm water consumed and the energy input for the EWH with an error of less than 10% for 3 of the 4 datasets considered. An Android mobile application was then created to allow consumers to remotely monitor and control their EWH from their mobile device. The EWH model was implemented as part of the functionality of the mobile application to provide a user with instantaneous feedback on the impact of changes in control settings and usage profiles. For example, this functionality in the mobile application allows users to determine how switching their EWH off intermittently will affect their energy consumption. Additionally, the event detection algorithm was utilised by the mobile application to establish usage profiles and provide recommended schedules for users, based on their consumption data. Finally, a usability study was conducted in order to evaluate the ease with which users are able to utilise the mobile application and to improve on any areas of difficulty that may exist. Several areas of difficulty were determined and these results were used to implement various changes to improve the application by making it more user friendly. The results of the study indicate that the system is user friendly and that participants had a positive overall experience with the mobile application.

AFRIKAANSE OPSOMMING: Suid-Afrika is, tydens finalisering van hierdie manuskrip, in die middel van ’n energiekrisis, aangesien die nasionale voorsiening nie in staat is om aan die energie behoeftes van die land te voldoen nie. Elektriese warmwatersilinders (EWs) bly een van die grootste bydraers tot residensiële energieverbruik in Suid-Afrika asook ander lande waar dit gebruik word. Alhoewel opvoedkundige materiaal al gepubliseer is om bewustheid van energiebesparende maatreels vir EWs te skep, is dit onduidelik of gebruikers die inhoud verstaan en hul EWs effektief beheer. Onvoldoende terugvoer van die gebruiksinligting maak dit ook moeilik vir verbruikers om hul verbruikspatrone te verstaan en ingeligte besluite te neem oor hul toekoemstige water en elektrisiteit verbruik. Hierdie werkstuk bied ’n selfoon-gebaseerde eko-terugvoerstelsel aan vir die energie en water verbruik data van residensiële EWs. Die stelsel bestaan uit verskeie komponente: ’n termiese EW model; ’n gebeurteniswaarnemingsalgoritme; en ’n Android mobiele toepassing. Die fisika gebaseerde EW model is ontwikkel om die energie toevoer en afvoer van ’n EW vir verskeie beheerverstellings, gebruiksprofiele en EW oriëntasies akkuraat te simuleer. Die akkuraatheid van die model is bevestig met ses datastelle, vier wat 900 ure van verskeie gebeurtenise behels, en twee met slegs staande verliese. Die resultate toon aan dat gemete energieverbruik lewer met ’n beraamde fout van minder as 2% en 7% vir skedule beheer en termostaat beheer onderskeidelik. Sowel as ‘n hoë akkuraatheid, het die model ’n lae berekenings-kompleksiteit en neem slegs 100 millisekondes om ’n 10 dag simulasie te voltooi op ’n standaard rekenaar, wat dit ideaal maak vir gebruik met mobiele toestelle. ’n Unieke en nie-indringende hardeware oplossing en ’n bypassende algoritme is ontwikkel wat die identifisering en klassifisering van warm water verbruiksgebeurtenise ondersteun sonder die gebruik van versteurende installasies en duur watermeting tegnologie. Die algoritme is getoets met 49 dae se data wat 127 gebruiksgebeurtenisse behels, en dit was bevind dat die algoritme gebeurtenise akkuraat kan waarneem met ’n akkuraatheid van 91%. Verder het die algoritme klein gebruiksgebeurtenise waargeneem (0.5 liter is suksesvol waargeneem). Tog is die duurteskattings van gebeurtenise binne 2 minute akkuraat 79% van die tyd. Daarna is die uitlaattemperatuur en water meter data gebruik as insette tot die EW model vir die beraming van die energieverbruik onder verskillende beheerverstellings. Die uitlaattemperatuur was gebruik om die totale volume warm water verbruik en energie-inset te skat met ’n fout van minder as 10% vir 3 van die 4 datastelle wat beskou was. ’n Android mobiele toepassing is geskep om verbruikers afstand monitering en beheer van hul EW te gee deur ‘n mobiele toestel. Die EW model is as deel van die funksionaliteit van die mobiele toepassing geïmplementeer om ’n verbruiker van oombliklike terugvoer aangaande die impak van veranderinge in beheer verstellings en gebruiks profiele te voorsien. Byvoorbeeld, verbruikers kan bepaal hoe die tussenpose afskakel van hul EWs hul energieverbruik beïnvloed. Die gebeurteniswaarnemingsalgoritme is daarbenewens deur die mobiele toepassing gebruik om gebruiksprofiele te bepaal en skedules vir gebruikers aan te bevel op grond van hul verbruiks data. Ten slotte, is ’n bruikbaarheids studie uitegevoer om te bepaal hoe maklik gebruikers die gebruik van die mobiele toepassing vind om te gebruik om sodoende enige probleme wat mag bestaan te verbeter. Verskeie probleme is geidentifiseer en die resultate is aangewend om verskeie veranderinge aan te bring om die toepassing meer gebruikersvriendelikheid te maak. Die resultate van die studie dui daarop dat die stelsel wel gebruikersvriendelik is en dat in die geheel deelnemers se ervaring met die mobiele toepassing ’n positiewe een was.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/98076
This item appears in the following collections: