75 lines
4.8 KiB
Typst
75 lines
4.8 KiB
Typst
#import "/metadata.typ": *
|
||
#pagebreak()
|
||
= Context <sec:spec:context>
|
||
|
||
#let spec_context = [
|
||
The @ofsp emphasises that insufficient ventilation can lead to increased @co2 levels #cite(<ofsp_ventilation>), which can negatively affect cognitive performance and health.
|
||
|
||
It is advised to monitor indoor air quality in classrooms using @co2 concentration sensor. The @co2 level should not exceed a threshold of 1400 @ppm. When this threshold is exceeded, the room should be ventilated until the concentration falls below 1000 @ppm. For reference, outdoor air typically contains around 400–500 @ppm of @co2.
|
||
These thresholds are summarised using @tab:co2_levels.
|
||
|
||
The @ofsp recommends ventilating classrooms effectively before each lecture and during breaks by opening the windows fully to renew the air quickly. To minimise heat loss and prevent the cooling of walls in winter, the ventilation period should remain short (a duration of 3 to 5 minutes is generally recommended).
|
||
|
||
Minimising heat loss reduces heating demand, contributing to the objectives of Switzerland’s Energy Strategy 2050, namely reducing energy consumption, improving efficiency and lowering greenhouse gas emissions.
|
||
|
||
An air quality simulation tool named SIMARIA #footnote[https://simaria.ch/fr/simaria] has been developed to assist teachers in planning the timing and duration of window ventilation, based on classroom size, number of students, and class schedules.
|
||
|
||
In this context, HES‑SO focused on this issue and installed @co2 sensors in the classrooms. Indeed, the long duration of class periods (2h30), combined with windows often remaining closed before, during, and after lessons, can result in high @co2 concentrations. Moreover, some classrooms can accommodate a large number of students, which also contributes to the increase in @co2 concentrations.
|
||
|
||
Sensors have been installed to alert teachers and students to the @co2 levels in the room, encouraging ventilation. However, it has been observed that some students are unaware of their presence, and both students and teachers often do not pay attention to them. The timing and duration of window opening depend on the individual, often without regard to the @co2 levels indicated on the sensors and without consideration of the resulting energy loss.
|
||
|
||
#figure(
|
||
table(
|
||
columns: (auto, auto),
|
||
align: center,
|
||
table.header("co2 level [ppm]", "influence"),
|
||
[400],[Normal outdoor air],
|
||
[400-1000],[Typically found in ventilated rooms],
|
||
[1000-2000],[Medium air quality \ usually reached during a 2 hour course without airing the room],
|
||
[>2000],[Poor air quality \ Reported headache, loss of focus, poor concentration, ...],
|
||
),
|
||
caption: [Noticeable CO2 level threshold]
|
||
)<tab:co2_levels>
|
||
The @tab:co2_levels considers recent review #cite(<co2_levels_review>).
|
||
|
||
#let top-level-architecture-get-data = [
|
||
#figure(
|
||
image("/resources/img/PI-get_data.drawio.png"),
|
||
caption: [Top level architecture to get measurement],
|
||
alt: "Thingy52 --BLE secure advertising--> Raspberry Pi --MQTT--> RabittMQ --GO--> InfluxDB"
|
||
) <fig:top-level-architecture-get-data>
|
||
]
|
||
// #top-level-architecture-get-data
|
||
|
||
#let top-level-architecture-lstm-notify = [
|
||
#figure(
|
||
image("../../resources/img/PI-lstm.drawio.png"),
|
||
caption: [Top level architecture to analyse data and notify],
|
||
alt: "Influx + MeteoSwiss --> LSTM --> Teams notification"
|
||
) <fig:top-level-architecture-lstm-notify>
|
||
]
|
||
// #top-level-architecture-lstm-notify
|
||
|
||
#let top-level-architecture = [
|
||
#figure(
|
||
image("../../resources/img/PI-top-level.drawio.png"),
|
||
caption: [Top level architecture],
|
||
alt: "Thingy52 --BLE secure advertising--> Raspberry Pi --MQTT--> RabittMQ --GO--> InfluxDB | Influx + MeteoSwiss --> LSTM --> Teams notification | InfluxDB --> MatLab --> InfluxDB"
|
||
) <fig:top-level-architecture>
|
||
]
|
||
|
||
As indicated in @fig:top-level-architecture, the planned system consists in nodes retrieving the environmental data.
|
||
These data are sent to a gateway, running on a raspberryPI, with BLE.\
|
||
The gateway then aggregates the various data and sends them, added by a timestamp, to a database.
|
||
The database offers an API allowing the insertion and the retrieval of the stored data.
|
||
A physical model ensures that each point of data is coherent with the one coming before and after it.\
|
||
These data can be either displayed, with regards to a given timestamp, of fed to a forecasting model.
|
||
Ultimately, the model generate prediction on the upcoming points of data and sends notifications using Microsoft teams to the room users.
|
||
The use of Microsoft teams is motivated by the shared use by the students and teachers.\
|
||
The notifications are indicating when to open, respectively close, the room windows to ensure that the air quality remains proper for a course while reducing the heat transfers with the outside to the minimum.
|
||
#top-level-architecture
|
||
|
||
]
|
||
#spec_context
|
||
|