PR 00264: verschil tussen versies
(Nieuwe pagina aangemaakt met ' {{Project config}} {{Project |Supercontext=PR 00220 |Project type=Standaard |Name=Smart meter data analysis |Start date=2019/01/25 |End date=2019/06/30 |Summary=Th...') |
Geen bewerkingssamenvatting |
||
| Regel 1: | Regel 1: | ||
The Dutch association for managing the electricity network (Netbetheer Nederland, NBNL) would like to analyze data about failures, or malfunctioning in “smart meters”, to gain insights that help them to better and more efficiently carry out their tasks. Tasks have been described by NBNL in a so-called value profile. | |||
Data about smart meters are available at the “population” level. A population is a group of meters with common attributes, like the manufacturer and the components used. Attributes and size of the population are known in advance. Apart from that, data are collected when there is a malfunction in a smart meter. And every five years, a sample is drawn from every population to check and verify the quality of the meters. When a population does not meet the quality standards, the network manager needs to replace the meters within a year and a half. Sometimes that is challenging, because population sizes can be quite large. Based on the analyses, NBNL would like to closely follow the technical status of the meters when time passes. Based on trends and risk profiles, the network managers hope to improve their insights and planning. The big question is how the currently available data can be used to do that. | |||
HZ University of Applied Sciences made a value proposition for NBNL, which describes various ways to analyze the data. In this trajectory a specific assignment for students is defined: | |||
''The trends and profiles that are developed initially, have a relatively simple target: malfunctioning or not, and rejected or not. But not all malfunctions are the same, and there can be different reasons why a meter does not meet the quality criteria. Further research can discover those differences, and enable the development of more specific risk profiles and trends. There is limited information about specific malfunctions and rejections, and it is hard to estimate in advance what opportunities the current dataset offers.'' | |||
The specific assignment is therefore: find out how information about specific malfunctions and quality problems | |||
* can be recognized in the data | |||
* can enrich the analyses | |||
{{Project config}} | {{Project config}} | ||
{{Project | {{Project | ||
Versie van 25 jan 2019 17:24
The Dutch association for managing the electricity network (Netbetheer Nederland, NBNL) would like to analyze data about failures, or malfunctioning in “smart meters”, to gain insights that help them to better and more efficiently carry out their tasks. Tasks have been described by NBNL in a so-called value profile.
Data about smart meters are available at the “population” level. A population is a group of meters with common attributes, like the manufacturer and the components used. Attributes and size of the population are known in advance. Apart from that, data are collected when there is a malfunction in a smart meter. And every five years, a sample is drawn from every population to check and verify the quality of the meters. When a population does not meet the quality standards, the network manager needs to replace the meters within a year and a half. Sometimes that is challenging, because population sizes can be quite large. Based on the analyses, NBNL would like to closely follow the technical status of the meters when time passes. Based on trends and risk profiles, the network managers hope to improve their insights and planning. The big question is how the currently available data can be used to do that.
HZ University of Applied Sciences made a value proposition for NBNL, which describes various ways to analyze the data. In this trajectory a specific assignment for students is defined:
The trends and profiles that are developed initially, have a relatively simple target: malfunctioning or not, and rejected or not. But not all malfunctions are the same, and there can be different reasons why a meter does not meet the quality criteria. Further research can discover those differences, and enable the development of more specific risk profiles and trends. There is limited information about specific malfunctions and rejections, and it is hard to estimate in advance what opportunities the current dataset offers.
The specific assignment is therefore: find out how information about specific malfunctions and quality problems
- can be recognized in the data
- can enrich the analyses