Digitalizing waste collection in Tinn Municipality : a case study
Master thesis
View/ Open
Date
2021Metadata
Show full item recordCollections
- Master's theses (HH) [1138]
Abstract
Municipal waste is an ever-growing problem in the modern world. According to The World Bank, 2,01 billion tons of municipal waste is generated annually. They expect global waste generation to reach 3,40 billion tons by 2050, which is more than twice the expected population growth for the same period (The World Bank, n.d.). This expected growth in waste generation will lead to increased pressure on waste collection, making the process of waste management even more important than before.
In Norway, 20% of the waste is generated by private households (SSB, 2020). Collection of waste from private households in Norway is done based on static routes, where waste is collected regardless of how full the containers are. By using sensor technology, one can monitor waste levels in the containers and adjust the collection routes accordingly. Studies conducted in other European cities have shown that a cost and emissions reduction by 30–60% can be achieved by implementing sensor technology (ISWA, 2019).
Together with Sensorita, the thesis aims to digitalize waste collection in the municipality of Tinn, Norway. Using Tinn as a case study, the goal is to create a contextualized data model for waste management which can be used by several municipalities. The objective is to improve the current waste management process by using sensor technology. The Sensorita System contains three components. A hardware, software, and a sensor. The sensor is located on the back of the truck and collects data from wastebins using radar technology. The contracting process in a small, rural municipality with a limited economic budget will also be studied.
The key objective is to look at the existing process within waste collection and evaluate whether it can be improved based on several criteria through digitalization. The criteria in focus are costs, efficiency, and environmental friendliness. The project team aims to initiate the digitalization process of Tinn municipality and provide valuable knowledge for Sensorita to further build and grow upon.
Tinn municipality has budgeted a total renovation cost of NOK 2 900 000 in 2021 (Tinn Municipality, 2020). This is a total of NOK 16 233 500 when adjusted for prices and wage increases from 2022–2027. A total of 32 524 kilometers are driven annually to collect municipal waste in Tinn, and there are a total of 3 133 municipal clients. This amounts to an estimated cost of NOK 1 036 and 10.38 kilometers driven per municipal customer. Total project costs and investments are estimated to be NOK 6 425 583 from 2022–2027. Three scenarios are presented when considering total savings potential. The realistic scenario includes an 20% reduction in costs and driving distance today. This reduces the cost by NOK 808 and annual driving by 8.30 kilometers per municipal customer. Considering these costs, an estimated savings potential of 22.0% can be achieved when compared to what Tinn municipality has budgeted. This is close to the rest of the digital waste collection market and what Sensorita´s competitors claim. The results are meant to illustrate potential effects by transitioning to a digital waste collection system.
More dynamic driving routes can lead to time savings as the sensors use machine learning to learn and grow independently, leaving human intervention excluded from the learning process. Sensors can cause more efficient time use by emptying bins that are actually full. To make sure that cost reductions are realistic, it is pivotal that the waste collection system is implemented correctly. This will require additional resources in the form of IT support, increased monitoring of the new process, and adjustments to the new workflow for the employees.
The reduction of climate gas emissions is difficult to determine based on the data provided. As is the case with the cost reduction, it is highly likely that the digital waste collection system will contribute to reducing emissions. A reduction will then come in the form of being able to empty more bins and collect more waste on the dynamic routes compared to the static routes. This means fewer stops and unnecessary trips to bins that have not reached their capacity.
The project group recommends that Sensorita starts by doing more pilot projects in order to gather more data and experience to build a solid foundation for further development of the machine learning algorithm. This can strengthen their position in a potential tender. Sensorita should look at how AI can be used to design driving routes and automate the waste plan. They should look at the possibility of analyzing weight as this is something many of their competitors already offer. In relation to the current situation, Sensorita should focus on collecting more data, verifying the technology they are using, document savings, and learning from the process. Kommunalt avfall er et voksende problem i dagens samfunn. Ifølge The World Bank blir 2,01 milliarder tonn søppel generert årlig. De forventer at den globale avfallsproduksjonen vil nå 3,40 milliarder tonn innen 2050, noe som er mer enn det dobbelte av forventet befolkningsvekst for samme periode (The World Bank, n.d.). Den forventede veksten i avfallsproduksjon vil føre til et økt fokus på avfallsinnsamling. Dette gjør avfallshåndteringsporsessen enda viktigere enn før.
I Norge blir 20% av søppelet produsert av private husholdninger (SSB, 2020). Avfallsinnhentingen fra husholdningene er basert på statiske ruter. Det vil si at søppel hentes uavhengig av hvor fulle søppelkassene er. Sensor teknologi gjør det mulig å observere hvor mye søppel en søppelkasse inneholder, og på den måten kan man hente søppel etter behov. Studier gjennomført i andre europeiske byer har vist at 30–60% av kostnader og utslipp kan reduseres dersom sensorteknologi benyttes i avfallsinnhenting (ISWA, 2019).
I samarbeid med Sensorita er målet å digitalisere avfallsinnhentingen i Tinn Kommune. Hensikten med oppgaven er å lage en kontekstualisert modell for avfallshåndtering som kan implementeres i flere kommuner. Hensikten er å forbedre dagens prosess ved bruk av sensorteknologi. Systemet til Sensorita består av tre komponenter. En hardware, software og en sensor. Sensoren plasseres bak på søppelbilen og samler inn data ved bruk av radarteknologi. Oppgaven vil også se nærmere på anbudsprosessen i en mindre distriktskommune med begrenset økonomi.