TTK Links: Home Page | TTK Moodle | ÕIS
TTK/UAS Repository For Digital Teaching and Learning Resources

Päikeseelektrijaamade tootlikkuse prognooside paikapidavus

Andreson, Deve (2018) Päikeseelektrijaamade tootlikkuse prognooside paikapidavus. [thesis] [en] Reliability of Solar Power Plant Productivity Forecasts.

[img]
Preview
PDF - Published Version
Download (1MB) | Preview
[img] Other (lihtlitsents- bdoc) - Other
Download (76kB)

Abstract

Sissejuhatuses sai autori poolt püstitatud probleem ja küsimused ning graafikute analüüsimise ja järelduste tulemusena on saadud kõikidele uurimisküsimustele vastused. Samuti sai tulemuste põhjal ülevaate prognooside paikapidavusest ning erinevate arvutusmeetodite prognoosimise täpsusest. Töös analüüsiti kokku üheksat erinevat päikeseparki, mis olid kõik erineva võimsuse ja mitme erineva paigaldusviisiga. Tulemuste saamiseks analüüsiti ettevõtte Solar4you OÜ poolt edastatud päikeseparkide tegelikke ja prognoositud tootlikkusi ning lisaks võeti võrdlusesse ka PVgis ja PVsyst programmiga arvutatud prognoosid, mis andis tootlikkuse prognooside paikapidavusest parema ülevaate. Võrreldud on tootmismahtude andmeid kuude lõikes, mille põhjal on tulemuste peatükis koostatud graafikud. Samuti sai analüüsitud tootmismahtusid kogu aastase toodangu lõikes, mille kohta on koostatud tulemuste peatükis tabel ning sarnane tabel sai koostatud ka eritootlikkuse kohta. Probleemi peamine püstitus töös oli, et kas prognoosid vastavad tegelikkusele ning kas erinevaid arvutusmeetodeid tootlikkuse prognoosimiseks saab usaldada. Vastuseid saadi ka küsimustele, kas TTÜ poolt loodud arvutussüsteem Solar4you OÜ-le on täpsem, kui internetis leiduvad prognooside arvutusmeetodid ning kas prognoosid on liialt optimistlikud või vastupidi. Antud küsimustele on vastused välja toodud ka iga graafiku all ning kogu tootlikkuse ja eritootlikkuse tabelite all. Tulemuste analüüside ja järelduste põhjal saab järeldada, et 100% täpsusega ei ole võimalik tegelikku tootlikkust prognoosida mitte ühegi programmi ega arvutusmeetodiga. Selleks, et saada täpsed tulemused, peaks ette teadma, kui efektiivne on päikesekiirgus, paneelide mustuse täpsed kaod ja pargile langevad varjud, kuna need kolm faktorit mõjutavad PV-süsteemi efektiivsus tugevalt. Kuna aga neid faktoreid ei ole võimalik ette ennustada ning ka kõik päevad, kuud ja aastad on päikesekiirguse efektiivsuse osas erinevad, siis on tegeliku tootlikkuse ja prognoosi väikesed erinevused aktsepteeritavad. Küll sai aga graafikute põhjal järeldada, milline arvutusmeetod prognoosis tegelikkusele kõige täpsemini. Aasta esimeses pooles (jaanuar-aprill) domineeris pigem Solar4you arvutusmeetod, suvekuudel (mai-september) andis kõige täpsemad tulemused PVgis. Aasta lõpp kujunes üldiselt üsna umbmääraseks, vahe tegeliku ja prognoositu vahel oli märgatav kõikide prognooside puhul. Vaadates kogu aastase tootlikkuse ja eritootlikkuse tegelike ja eri meetoditega prognoositud tootmismahtude erinevusi, saab järeldada, et internetis olevad arvutusmeetodid (PVgis ja PVsyst) on täpsemad. Kogu aastase tootmismahu võrdluse põhjal saab öelda, et kõige paremini prognoosis PVgis programm, mille ennustus oli võrreldes teistega kõige täpsem kuue pargi puhul üheksast. Teisele kohale jäi PVsyst programmi arvutusmeetod, mille prognoosid olid suhteliselt lähedased PVgis prognoosile. Kolmandale kohale tuli Solar4You arvutusmeetod, mis tõusis prognoosi täpsuse poolest tegelikkusele esikohale kahel korral, kuid üldiselt olid prognoosi aastase ja tegeliku aastase tootlikkuse vahed suured. Vaadates suurt pilti, siis võib tõdeda, et üldjuhul on prognoosid ikkagi suuremad kui tegelik tootlikkus ning kohati ka liialt optimistlikud ning seda eriti just suveperioodil ning jaanuaris ja detsembris. Kõige rohkem jäi optimistlike prognoosidega silma just Solar4you arvutussüsteemi prognoosid suveperioodil, mis oli enamike parkide puhul tegelikkusest tunduvalt suuremad. Kokkuvõttes võib öelda, et töös püstitatud küsimustele saadi vastused ning töö eesmärk sai ka täidetud. Ootus, et Solar4you prognoosimeetod on kõige täpsem ning on võimalik näidata, et ka teised päikeseenergialahenduste pakkujad peaksid võtma selle vaeva, et luua ettevõttele oma arvutussüsteem, ei täitunud ja lõpptulem oli hoopis vastupidine ehk päikeseenergia lahendusi pakkuvad ettevõtted võivad julgelt kasutada internetis leiduvaid prognoosi arvutusmeetodeid, eriti just PVgis programmi. Kokkuvõtteks on tulemused ettevõtte Solar4you OÜ jaoks siiski pigem negatiivsed ning nende arvutussüsteem tuleks teadlaste poolt üle vaadata ning parandused sisse viia.

Abstract [en]

The introduction of solar energy is an ever-increasing trend in the world. Solar parks are built on the roofs or on the ground of private houses and production plants, but in all three cases the goal is same: to reduce costs of electricity consumption by maximizing energy production from renewable sources. Solar energy solutions providers give clients projections for each park in terms of monthly production (kWh), annual production (kWh) and specific production (kWh/kW). Different companies use various programs and computational methods to calculate the forecast. At the moment, there is no overview of exactly how precise predicted productivity is, and whether the prediction calculation methods can be trusted at all. The aim of this thesis is to compare the projected and actual productivity of solar power plants on the basis of solar power plants established by one solar power solutions providing company. The company transmitted the actual and projected productivity data of nine solar parks installed by them and, in addition, two other methods for calculating the forecast on the Internet have been compared. Comparable projection computing systems are Solar4you, PVgis, and PVsys computational methods. The goal is also to provide feedback to the company regarding the reliability of their computing system and the validity of the forecasts. In addition to the goal, several research questions have been identified in the work. First, whether the projected productivity reported by the exemplary company to customers is true, weather Solar4you's calculation system is more accurate than the calculation methods available on the Internet. Third, to make the appropriate conclusions, the validity of the estimates is analyzed by months to assess their relevance under different production conditions. The aim is also to answer the question of whether the forecasts are too optimistic or vice versa. In order to perform comparisons, analyzes and to answer the questions raised in the work, the actual production volumes and forecasts received from the company and forecasts calculated on the Internet programs were used. A total of nine different solar power stations were compared to 2017 data, Smarten Logistics Solar Park is listed in addition to the 2017 data also the productivity data for 2016. Parks have different peak power, orientation and installation. In this work three solar parks of ground mounted, three solar parks on tilted roof and three solar parks mounted on a flat roof have been studied. The smallest parks in the work are solar parks mounted on the tilted roofs, and the largest solar parks is on a flat roof with peak power of 375kW. Based on the results of the analysis and conclusions, it can be concluded that with 100% precision it is not possible to predict actual productivity with any program or calculation method. In order to obtain accurate results, it should be made aware of how effective solar radiation is, the exact losses in PV-panels coming from dirt and shadows falling on the park, as these three factors strongly affect the efficiency of the PV system. In the first half (January-April), Solar4you was more likely to dominate the calculation method, during the summer months (May-September) gave the most accurate results PVgis. The end of the year was generally fairly non-standard, with the actual and unpredictable difference being significant for all projections. Looking at the differences in production volumes estimated by actual and different methods of annual productivity and specific productivity, it can be concluded that the calculation methods on the Internet (PVgis and PVs) justify themselves better. On the basis of a comparison of the total annual production volume, it can be said that PVgis , which was the most accurate predictor of nine parks in the six parks, is best predicted. In the second place, the calculation method of the PVSyst program remained, with projections that were relatively close to the PVgis forecast. Unfortunately, Solar4You came in third of terms of being closest to actual and predicted production. Looking at the big picture, it can be said that, as a rule, forecasts are still higher than actual productivity and at times too optimistic, especially in the summer period and in January and December In sum, it can be said that the answers to the questions were answered, and the goal of the work was also met. However, cooperation with Solar4you OÜ was a small hope that it would be possible to confirm the reliability and validity of their calculation method, but the final result is different in general terms. If Solar4you now wishes to continue to use its own forecasting method, then it is important for scientists to update their computational formulas, especially in summer months (May-September) and winter months (January and December), in order to make predictions more realistic in these months. On the other hand, it would be wise, for example, to calculate and compare projections in the PVgis program with a view to making changes, if necessary. If continuation of the use of its own calculation method is not an objective, then the recommendation is to switch over to the calculation method for forecasting in PVgis in order to avoid the transmission of overly optimistic forecasts to customers.

Item Type: thesis
Advisor: Siret Talve
Subjects: Technoecology > Technology and Waste Management > Energy and the Environment
Technoecology > Technology and Waste Management > Sustainable Ecological Technologies
Divisions: Institute of Circular Economy and Technology > Environmental Technology and Management
Depositing User: Deve Andreson
Date Deposited: 05 Jun 2018 13:01
Last Modified: 05 Jun 2018 13:01
URI: http://eprints.tktk.ee/id/eprint/3638

Actions (login required)

View Item View Item