In the beginning, as usual, a few words about the heroes.
ClimateGuard is a hardware startup built from the scratch to implementation of climate monitoring projects. Initially, it was focused on the custom device for creating a supportive and healthy environment at home and at work. But destiny (or market) decided otherwise.
Year of development: 2018
Dislocation: Russia, Moscow, Electrozavod (MELZ)
Participants:
- technical analyst-PM
- technical director
- designer
- circuit designer
- C/C++ developer
- web developer
- layout designer
- android developer
- 2 UI/UX designers
- engineer
- fullstack manager (manager, sales rep, accountant and storyteller)
Achievements since the previous article:
- the project is alive for more than 3 years (!!!)
- 10+ products made
- b2b, b2g and b2c sales
- our devices been bought in the USA, Germany, Italy, Poland, Japan, India
- made friends with a large number of manufacturers, electronic stores and enterprises.
So, heroes characteristics are given. Let's talk about deeds.
StagingAt almost the same time, two different companies reached out to ClimateGuard with a request to develop a pilot device that could measure the air flow rate inside a ventilation pipe.
The first client is the SOK coworking network. They make a great effort to create comfortable working conditions for tenants. Every office must have the right temperature, humidity, clean air. The request from them was like: "We want to get a system that will allow real-time monitoring of the air flow rate and consumption in ventilation for each floor".
The second customer, in addition to the air flow rate, had to assess the humidity and the amount of dust in the ventilation pipe. And based on the values of measured climate parameters, control the fan coil units and replace the filters.
Taking into account all the nuances of the installation, the following requirements were imposed on the system:
- data transfer in 24*7*365 mode;
- low installation costs;
- power supply 12 V;
- displaying data in a single interface;
- the ability to visually compare the values and dynamics of data for different monitoring objects;
- availability of API for centralized data collection from all devices;
- low cost for subsequent scaling.
At the time of the request, colleagues were checking the air flow rate “in manual mode”. Meaning, they carried out regular one-time measurements at control points. Testo professional anemometers were used as equipment. Manual measurement takes a lot of time and requires well-developed skills, since the readings of the hot-wire anemometer can significantly vary depending on the location of the tube with the measuring sensor in the ventilation pipe. For the same reason, it is difficult to achieve high repeatability of values during repeated measurements.
ArchitectureTo minimize costs and simplify installation, it was decided to make the devices as autonomous as possible.
Anemometer:
- is installed directly in the ventilation pipe on ties or screws, with the output of the sensor element inside the pipe (to reduce the number of wires and nodes in the system and facilitate installation);
- independently collects, processes and tranfers data to the cloud server;
- receives power through a cable;
- uses wifi as a data transfer interface, connecting to a hidden corporate network (to reduce the cost of creating an additional data transfer infrastructure);
- has a status indication system based on a diode (to facilitate installation and performance testing);
- has a button for switching between operating / setting modes.
The disadvantage of the chosen architecture is the lack of power autonomy. However, there is a constantly working wifi module on board of the device, as well as a large number of sensors. The anemometer sensor requires constant heating. As a result, keeping the device running on battery power seems to be an extremely difficult (expensive) task.
The process of a device installation within the chosen architecture is described by the following steps:1. The installation spot is determined; in the absence of a sensor output channel, a hole is drilled in the ventilation pipe.2. A 12V wire is connected to the installation site.3. Before installation, the wifi-network parameters are configured on the devices.4. In the personal account, the devices are tied to the installation sites, the operation with the network and the fact of data transfer are checked.5. The device is installed by the "sensor" into the hole and is fixed by metal ties or by self-tapping screws and, if necessary, it is sealed.6. The power cable is connected to the device.7. The device turns on. After the control diode on the device turns green, the engineer proceeds with the installation of the next anemometer.
Components and costThe following set of components was used:
- esp32 controller;
- esp32 CG-Board_mini development board;
- CG-Anem air flow speed sensor;
- PMS7003 dust particles sensor;
- bme680 climate sensor.
CG-Board_mini
The first thing that comes to mind when developing a standalone IoT sensor is a board from the TTGO ESP32 family. The only problem is that there is no easy way to connect several sensors to it. Therefore, the team decided to go with the CG-Board_mini solution. The board has a compact size 46 * 76 mm. ESP32 on board provides wifi support out-of-the-box. The board has 9 outputs for connecting sensors and relay, a watchdog timer (heals one of the most important sores of esp32 - uncontrolled freezing), 2 RGB diodes, a button and 3 power connectors (USB, DC and a terminal) with a range of 5 and 12-48 V.
CG_Anem
Usually, hot-wire anemometers are used in industrial solutions due to their low cost, lack of mechanical components and compact dimensions. As a useful side effect, these sensors measure air temperature very accurately. But there is a lack of modular hot-wire anemometers in the market. At first, the team was thinking about Wind Sensor from Modern Device. But during testing, several critical blockers were found:
- in this particular solution, signal processing is carried out not on the module, but on the side of the external controller;
- sensors do not have factory calibration;
- based on the logistics, the sensors are made “by request”.
As a result, we decided to make our own hot-wire anemometer module - CG-Anem (a separate article will be devoted to the process of its development). The following features of the sensor turned out to be important for users:
- in terms of accuracy (0.3 m/s ± 10%), the sensor is not far behind professional measuring equipment (e.g. Testo device in the middle price category has an accuracy of 0.3 m/s ± 10%);
- the sensor is calibrated by the manufacturer and responds the controller with accurate data (which means it can be easily replaced in case of failure).
During testing of the IoT anemometer with the CG-Anem sensor, we got pretty good convergence of the results with the CEM DT-8880 reference anemometer.
PMS7003
An important advantage of the PMS7003 dust sensor is its compact size (it is certainly more compact than the sds011 popular among eco-activists). This allows you to place it inside the vent pipe. Also, the sensor is cheap, easy to connect. Its accuracy has been confirmed by numerous studies by non-British scientists (e.g. «Evaluation of Low-Cost Sensors for Ambient PM2.5 Monitoring»).
BME680
BME680 – very controversial decision. Configuring it to work with esp32 is tricky. Because of the heating element in the sensor required for VOC measurement, the sensor can generate inaccurate temperature data. But it's still a Swiss knife - it's cheap and measures temperature, humidity, pressure, VOC and AQI. It is difficult to find something to replace him.
Cost
Componets:
- CG-Board_mini - $35
- CG-Anem - $23
- PMS7003 - $17
- bme680 - $13
- cabels, screws, etc. - $3
- body - $3
Total:
- $64 for CG-Anemometr – basic version of anemometer;
- $94 for CG-Anemometr PRO – extended version supports of dust particle concentration, humidity and VOC.
The requirements for the device’s body have been presented above. Therefore, instead of boring text, this section contains pictures of early prototypes and a ready-made solution.
Just wanted to draw your attention to three points:
1. The cases were designed for 3D printing.
2. Together with the users, we considered the possibility of attaching the device to the ventilation pipes with magnets. The idea did not justify itself, since for an accurate and precise measurement of the climate parameters, the device must be rigidly fixed to the pipe
3. Anemometers were designed taking into account two options for installation of a sensitive element (air flow rate sensor) - at the edge of the pipe and in its center. For the second option, it was decided to use an extension stand. The size of this stand is selected based on the radius and geometry of the pipe.
All final 3d model for Anemometr are placed in attachment to this article.
Data collection and processingTo collect, analyze and visualize data, a proprietary IoT platform was used. It is simple, functional enough, and it covers common use cases. And most importantly, it's free.
A telegram bot is used for notifications, which is configured in the web application. An API is available for data extraction.
Project DurationIt took 3 months to develop the CG-Anem sensor. The actual development of the IoT anemometer was about a month. Stage of on-site testing (post-installation) was greatly simplified due to the device's support for OTA updates.
4 people participated in the project. The timing looked like this.
As an outcome, users received a device with the following specification.
General device characteristics
Measured parameters:
- Air flow rate, m/s
- Ambient air temperature, °С
- Air consumption, liter/min
- Humidity, %
- Pressure, mm Hg
- Particles pm1, pm2.5, pm10, mg/m3
- VOC concentration, mg/m3
- AQI
Power supply: 12 – 48 V
Interfaces: Wi-Fi
Fastening: ties, self-tapping screws
Operating temperature range: -20... +60 °C
Dimensions: 73х38х25 mm
Tube diameter: 29 mm
IoT-platform: ClimateGuard
Metrological characteristics of the anemometer
Anemometer type: Thermoanemometer
Flow rate measurement range: 0.1 … 25 m/s
Temperature resolution: 0.1 °С
Air flow velocity resolution: 0.1 m/s
Air flow velocity resolution
Accuracy:
- 0…2 m/s: 0.2 + 5 %
- >2 m/s: 0.4 + 10 %
Companions and sympathizers, that is the first article from the planned cycle. The small, but very excited team of the ClimateGuard startup is waiting for your comments and will be very grateful to the criticism. We will reply to all the comments and will definitely take into account the feedback of the community in the following articles.
If Hackster will be interested in this article we will write about the development of the CG-Anem gas flow rate sensor itself. There is a story to tell, and there will be a lot of science.
Now we take our leave, and give our creation to your judgment...
…and keep believing: «Climate matters!»
This story is an updated translation of our russian article on habr.com.
Comments