René Elfrink, project coördinator GROW!...Analytics Easy access & applications IOT CONNECTED...
Transcript of René Elfrink, project coördinator GROW!...Analytics Easy access & applications IOT CONNECTED...
GROW!René Elfrink, project coördinator GROW!
Programma
Programma:
10.00 uur Ontvangst
10.15 uur Opening en toelichting op project GROW!
10.30 uur Toelichting op de ontwikkelde sensoren
11.00 uur In gesprek met elkaar over huidige ontwikkeling per sensor en wensen voor
toekomst
12.15 uur Terugkoppeling op de wensen
13.00 uur Afsluiting met aansluitend de gelegenheid de proef in de HAS Kas te
bezoeken en te lunchen.
Goal
“to monitor, optimize and enhance the yield and quality of horticulture
products by developing and deploying smart wireless sensor- and
datanetworks (Internet of Things, IoT”
Innovative sensor systems and big-datanetworks will be developed within the GROW!-
project and plant growth models will be optimized.
Implementation of both
• Currently available cost-effective sensors
• Newly developed sensors
To monitor other processes (nutrient monitoring).
The sensor- and datanetworks will be tested and valorized in fieldlabs together with SME
horticulture companies.
Partners
1. Nederlandse organisatie voor Toegepast Natuurwetenschappelijk Onderzoek TNO
2. Universiteit Antwerpen
3. LTO Glaskracht,
4. Katholieke Universiteit Leuven,
5. Stichting IMEC Nederland
6. HAS Hogeschool
7. Proefcentrum Hoogstraten
8. Vlaams Centrum voor Bewaring van Tuinbouwproducten
Samenhang
Technology R&D integration Test & growdiseases/plagues,
growth models
Quality / validationLink to agriculture
Beta system deployment at PCH : 10 sensor
boxes with Temperature and Humidity
7
Results :
input for KU-leuven → initial plant grow models
Input for WP3 → sensor selection and greenhouse
environment
15.0
17.0
19.0
21.0
23.0
25.0
26-Feb 5-Mar 12-Mar 19-Mar 26-Mar 2-Apr
tem
pera
tuur
(°C
)
Sensoren UA
box1box2box3box4box5box6box7
Sensors Temperature
RH
Light: 3 sensors present op 1 plot
Nutrients (in development):
─ EC
─ Nitrate
─ K, N, Ca, Mg
─ pH
CO2
Three types of sensors being tested (38 in total)
Sensoren UASensirionsSensoren KUL
15
20
25
30
35
40
23
-mei
9u
23
-mei
18
u
24
-mei
3u
24
-mei
12
u
24
-mei
21
u
25
-mei
6u
25
-mei
15
u
26
-mei
0u
26
-mei
9u
26
-mei
18
u
27
-mei
3u
27
-mei
12
u
27
-mei
21
u
28
-mei
6u
28
-mei
15
u
29
-mei
0u
29
-mei
9u
29
-mei
18
u
30
-mei
3u
tem
per
atu
ur(
°C)
Sensirion
Temp PRIVA afd 15
temp Sensirion 9
temp Sensirion 11
temp Sensirion 12
temp Sensirion 13
Temp Sensirion 14
Plantmodel
Measured properties of plant growth and development:
─ Stem thickness
─ Length
─ Bunch settlement
─ Yield (weight, dry mass)
─ Fruit Growth
─ Ripeness
─ Fruit quality (color)
Measured at 42 plots
Pathogenmodel
Pathogens measured:
─ White fly
─ Macrolophus (with and without enthomophthora)
─ Nesidiocoris
─ Powdery mildey
Counts with sticky plates and in the plants
Total 100 plots
SHORT SUMMARY 20182018
▪ Snelle start met enkele 10-tallen T en RH sensoren in PCH tomaten teelt▪ Gecombineerd met
▪ plantmetingen (Stengeldikte, Lengte, Troszetting, LAI, Opbrengst,Groei van de vruchten, Rijping van de vruchten,Vruchtkwaliteit)
▪ Pathogeen metingen▪ Model opgesteld▪ Technology ontwikkelingen :
▪ Sensor platform : Octa2.0▪ Printed sensors▪ EC, pH and Ion sensors▪ CO2 sensors
▪ Kontakten opgebouwd▪ Telers▪ Kassenbouwers▪ Klimaat besturings systeem bouwers
KOMENDE MAANDEN (PRACTICAL)2018-Q4
▪ Ion sensing : Laboratoriumtesten in voedingsmedium voor stabiliteit enkruisgevoeligheden
▪ Eerste laboratoriumproeven voor tweede ion▪ Prepare 100 Octa nodes for the deployment + Deployment and field testing at PCH
and HAS▪ Voorbereidingen nieuwe teelt PCH and HAS Q1-2019 (inzaaien oktober, planten in
December)
2019▪ Plant model validatie (model opgesteld 2018, toegepast op teelt 2019)
▪ Samenwerkingen met enkele telers starten
▪ Demo dag + kennis sessie (14 februari)
2019-Q2▪ integration with printed strip sensors + test in kas▪ Eerste proeven ion sensoropstelling in kas + Tweede ionconcentratie testen in lab
2019-Q3▪ Tweede ionconcentratie testen in kas
GROW!
Marcel Zevenbergen, imec-nl, Program Manager gas & liquid sensor solution
Marieke Burghoorn, TNO, GROW! werkpakket lead WP3
Toelichting ontwikkelde sensoren
Imec sensor network technologies:
2nd KFAS-imec Symposium, April 17-18, 2018
“Big data is like crude oil”
Andreas Weigend
former Chief Data Scientist at Amazon
in an interview with Business Week
Costeffective
air & liquid sensors
Sensor network
&
Calibration
Cloud Storage
&
Monitoring
Visualization
&
Analytics
Easy access
& applications
IOT CONNECTED SENSOR NETWORKS
Nutrient analysis by cost-effective liquid sensors
15
pH: 2-10 pH range, 0.1 accuracy
ORP
Ion-selective electrodes 10-4 M – 1 M range, 10% accuracy
Microfluidic
Reference electrode
World’s first long-term stable multi-sensor chip
pH, Cl-, Na+, K+, Ca2+, NO3-, DO, extension to other ions possible → Mg2+, NH4
+
Integrated with temperature and conductivity: 5 – 100000 µS/cm
QUALITY DATA ENABLED BY LARGE-SCALE LIQUID SENSOR NETWORKS
First field tests in greenhouse
17
Various samples obtained from irrigation lines which were
subsequently analyzed for pH, EC, and potassium level
Preliminary results
22-23-1 22-23-32 26-27-1 26-27-32 28-29-1 28-29-32 Drain Voeding
0
1
2
3
4
5
6
7
pH
pH1
pH2
pH3
Com
Difference pH < 0.1, EC < 0.1 mS/cm compared to state of the art sensors
Reproducible results for different samples from the irrigation lines, as expected
Drain EC and potassium significant higher
Next challenge: integrate these sensors in an IoT network for continuous and real-time data!
CONFIDENTIAL
FIRST DATA FROM THE BLANKAARTUSING IMEC’S CONDUNCTIVITY SENSOR
(a) (b)
imec analytics and imec low cost sensor
brings equal accuracy as 10’sk$
professional equipment
500 µm
CONFIDENTIAL
SUBSTRAAT METINGEN
20
Druppelaar
Sensor 1
(achter
druppelaar)
Sensor 5
Sensor 4 Sensor 2
Sensor 3
Sample 1 : 5x 20 mL Voeding
Sample 2 : 15 min na meting1
Sample 3 : 30 min na meting1
Sample 4 : 5x 20 mL Voeding (3h 30 min na meting 1)
Sample 5 : 5x 20 mL Voeding (3h 50 min na meting 1)
1 2 3 4 5 Voeding
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
µS
/cm
Sample
Sensor1
Sensor2
Sensor3
Sensor4
Sensor5
▪ Het substraat moet eerst goed verzadigd zijn om iets te meten, sensoren moetenvochtig zijn.
▪ Pas na de 2e voeding (sample 4) is er iets te zien bij sensor 1.
▪ Na de 3e voeding (sample 5) lijkt sensor 1 goed te meten, deze zit direct naast de druppelaar.
▪ Sensoren 2 t/m 5 meten veel ruis, hier is het niet vochtig genoeg.
IMEC CO2 SENSORMINIATURE, LOW COST GAS SENSOR PLATFORM
V~
• Interdigitated electrodes covered by
ionic liquid film
• Impedance change in ionic liquid film
250 ppm CO25001000150020002500
0 200 400 600 800 1000 1200 1400
122
123
124
125
126
127
128
129
130
131
132
133
|Z| (
kOhm
) @
1H
z
t (s)
moving average
plane fit and
vertical shift
Controlled environment
Training Testing
Real Life deployment in imec offices
Simple Multi-Variate
Model
Printed temperature sensoren (TNO)
Nearly no dependence on relative
humidity till 65%
Measurement in climate chamger at
95% RH conducted
Sensirion SHT31 as benchmark
Material optimization in progress
Printed sensor strips
Foil strip dimension 50 x 4 cm
Printted temperature sensors
Sensirion sensors as benchmark and for RH
Lightsensor will we added
Easily connectable: tailored length
Connected to readout unit that transfers data wirelessly
Sensor strips today
Interconnected PCB connected to Octa-platform
Thank you!
contact
René Elfrink
R&D manager – Sensors4IoT
Project coordinator GROW!
T +31 40 4020440
M +31 6 52598548
https://www.grensregio.eu/projecten/grow