The intelligent container - Cold Chain Management: Home

Supervision of Banana Transports

by the Intelligent Container

Coldchain Manangement. 4th International Workshop,

27./28. September 2010, Bonn

Reiner Jedermann, University Bremen, IMSAS / MCB

Axel Moehrke, Dole Europe Import BVBA, Belgium

Autonomous logisticsSFB 637MCBMCB 2

The banana challenge

Introduction (by Axel Moehrke) Bananas can produce up to 800 Watt of

heat per ton during ripening

Hard to control after a certain point

Improve processes along the transport and supply chain by permanent monitoring of quality changes

Untimely ripening is the greatest risk

Can be triggered in the field, in the packing plant or during transport

Or by insufficient cooling / temperature changes

Full and accurate temperature supervision and control is a precondition for improvements in the banana chain.


Online monitoring by the intelligent container

Real-Time remote monitoring of local temperature

deviations

Idea first presented 2006

Transfer project 2008 + 2009

Online Access

Focus on results from field tests

Outlook to future research

Sensor

Nodes

External

Communication

Gateway

Pre-Processing


Outline

Set up for field test

Observed temperature deviations

Over length of container

Between different containers

Inside one box

Required sensor density

Intelligent data processing

The future of the intelligent container


Installation in Costa Rica

Gateway


Installation in Costa Rica

4 pallets equipped with 4

sensors each S

S

S

S

S

Tier 1

Tier 8

Air ducts

Foam Block


External Communication

Forwarding data from sensors to web-server

Using the vessel’s email system

Web-

Server

WLAN

Satellite

Wireless

SensorsGateway

Internet

Email Server

ShipContainer


Webinterface

Mote-ID Last message: Temp. Humidity Voltage

1 2009-09-23 14:00:02 14.79 °C 93.0 % 2.8 V

2 2009-09-23 14:00:02 13.96 °C 94.0 % 2.81 V

3 2009-09-23 14:00:02 14.34 °C 95.0 % 2.84 V

4 2009-09-23 14:00:02 13.69 °C 81.0 % 2.8 V

5 2009-09-23 14:00:02 14.36 °C 102.0% 2.86 V

6 2009-09-23 14:00:02 15.2 °C 82.0 % 2.86 V

7 2009-09-23 14:00:02 15.57 °C 100.0% 2.75 V

8 2009-09-22 02:00:02 15.3 °C 97.0 % 2.82 V


Signal attenuation by the fruits

Distance 0.5 meter ⅓ of all links

completely failed

⅓ of all links was not available part of the time

⅓ of all links worked well most of the time

Partly compensated by network

New hardware?0 2 4 6 8 10 12

0

0.5

1

1.5

2

2.5

Pallet

19Pallet

14Pallet

8Pallet

1

Door

End

Refr

igera

tion U

nit

Gate

way

1 2 3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

Distance to cooling unit in [m]

Dis

tance to flo

or

in [m

]

Packet Rate: >0.75 0.33< <0.75 <0.33

Water-containing goods hinder the radio communication of wireless

sensors @ 2.4 GHz


The necessity for core temperature measurement

The internal sensors of the aggregate help only little to

estimate the banana temperature

4 6 8 10 12 14 16 18 20

14

16

18

20

22

24

26

28

Days in August 2010

Tem

pera

ture

in [°C

]

Pallet core reefer side

Pallet core door side

Pallet 11 mid

Air supply (Aggregate sensor)

Air return (Aggregate sensor)


Temperature difference over length of container

Bananas are cooled down by the container

Large differences in time required to achieve 17 °C

10 12 14 16 18 20 22 24

14

16

18

20

22

24

26

28

Time in [days] in September 2009

Pa

llet

Co

re T

em

pe

ratu

re in

[°C

]

Reefer Side Container 1

Door Side Container 1

Reefer Side Container 2

Door Side Container 2

Air Supply

Reefer side

2.4 days

Door side

6.35 days

Door-end

Reefer end

Two containers of

same type and

year of manufacture


Comparison of 4 Experiments

Large variations in age of containers

Newer equipment cools faster, but local variations

cannot be avoided

The hot-spot can be at the door end or somewhere in

the middle of the container

0 2 4 6 8 10 12

2008

2009(A)

2009(B)

2010

Age 9 years

Age 12 years

Age 12 years

Age 2 years

Time to cool down to 17 °C in [days]

Reefer End

Door End

Maximum


Location of hot-spots

Horizontal cut through the container

Average pallet core temperature in tier 5 (1.25 meter above floor)

Pallet 11 in the middle / left side is the hottest, but neighbor pallet

on the right side almost normal.

0 2 4 6 8 10 1215.4

15.6

15.8

16

16.2

16.4

16.6

Palle

t 1

Palle

t 1

Palle

t 3

Palle

t 11

Palle

t 11

Palle

t 10

Palle

t 19

16.20°C

15.80°C

16.42°C

16.59°C16.53°C

16.21°C

16.11°C

August2010


Avera

ge tem

pera

ture

in [°C

]

Hot-spot

Left side

Right

side

Door

Air supply


Required sensor density

4 Sensors are not sufficient

Pallet at aggregate and door side / lowest and highest tier

But where to place

the extra sensors?

10 12 14 16 18 20 22 24

14

16

18

20

22

24

26

28

Container BH

Time in Days in September 2009

Tem

pera

ture

[°C

]

Full Temperature Range

Range covered by corner sensors

Corners Sensors: [ 5 8 17 19 ]

Extra Sensors: [ 7 13 ]

0 2 4 6 8 10 120

0.5

1

1.5

2

2.5

Pallet19

Pallet14

Pallet8

Pallet1

Do

or

En

d

Co

olin

g U

nit

Ba

se

16.0

17.8

18.5

17.9

16.7

16.8

16.7

16.9

15.8

16.3

16.5

---

15.9

16.2

16.1

---


Dis

tan

ce

to

flo

or

in [

m]

Average Box Temperature in [°C]


4 6 8 10 12 14 16 18 20

14

16

18

20

22

24

26

28

Days in August 2010

Tem

pera

ture

in [°C

]

Accuracy of temperature measurements

How accurate can we measure temperature inside a

banana box?

Cooling air flows through the boxes variations

Even loggers in similar positions (distance ~ 5 cm)

≈ 0.1 °C at end of transport

Up to 1 °C during cool down

Temperature

difference in

one Box


Accuracy of temperature measurements

Pulp temperature measurement

Hurts the banana

Comparison: Logger temperature before opening of pallets

Pulp temperature Logger in average 0.05 °C too high (σ =

0.085 °C)

Pulp temperature in centre of box 0.2 °C … 0.4 °C higher than

side of box


Innovation Alliance

New project starts in September 2010

13 partner companies and 6 research institutes

Bremer Institut für

Produktion und Logistik

an der Universität Bremen

Sea-Containers

(Bananas)

Truck-Transports

(Meat)

Software RFID

and Electronics

Federal Ministry of

Education and

Research

Ethylene

Sensor


Intelligent data processing

Not only forward data, but pre-

process

Different algorithms (OSGi

Software-Bundles) on demand,

similar to App-Store for mobile

phones

Elements of the decision

support tool

Spatial interpolation of

temperature by Kriging

Prediction of future

temperature curve

Shelf life models

High

sensor

tolerance


10 15 20 25 30 35

14.5

15

15.5

16

16.5

17

15

.14

°C

15

.83

°C

16

.07

°C

15

.80

°C

16

.05

°C

16

.27

°C

16

.20

°C

15

.73

°C

16

.15

°C

16

.06

°C

16

.21

°C 16

.50

°C

16

.56

°C

15

.42

°C

15

.56

°C

16

.57

°C

16

.53

°C 16

.85

°C

16

.35

°C

16

.59

°C

15

.68

°C

15

.87

°C

16

.38

°C

16

.11

°C

15

.82

°C

15

.84

°C

16

.42

°C

Cro

wn

Rot

Ro

tten

Fin

ge

r

Ro

tten

Fin

ge

r

Button Number

Te

mpe

ratu

re in

[°C

]

Relation temperature and quality

No relation between box temperature and defects per box

Further influence factors have to be checked

O2, CO2, Ethylene

Age at harvest

Micro biological load

Mechanical damages


0 0.2 0.4 0.6 0.8 1

2008

2009(A)

2009(B)

2010

Average defective fingers per box

Rotten Fingers

Crown Rot

Anthracnose

Relation temperature and quality

Relation to average quality per container

But no statistical evidence

Only on container level

No relation inside one container

0 2 4 6 8 10 12

2008

2009(A)

2009(B)

2010

Time to cool down to 17 °C in [days]

Reefer End

Door End

Maximum

No data available


Summary

The pallet core temperature can show large variations

Differences in cool-down time

Temperature is tricky to measure

The hot-spot can be anywhere, will be missed if only 4 sensors

Even inside one banana box ΔT ≈ 0.5 °C

Relation of quality and temperature needs further

evaluation

Other factors likes gases and biological variance

The system will help to reduce losses by unwanted

ripening

Accurate temperature monitoring is the basis for the further steps

Compensate different quality levels by FEFO planning


The End

Thanks for your attention

www.intelligentcontainer.com

Dr. Ing. Reiner Jedermann

University Bremen, FB1Institute for Microsensors, -actors and –systems (IMSAS)

Otto-Hahn-Allee, NW1

D-28359 Bremen, GERMANY

Phone +49 421 218 62603, Fax +49 421 218 98 62603

Email [email protected]

Jedermann, R: Autonome Sensorsysteme in der Transport-und Lebensmittellogistik, Dissertation Universität Bremen, Verlag Dr. Hut, 2009.

http://www.intelligentcontainer.com/


Effects of interruptions of the power supply

Interruption of power supply during transshipment

(Harbor Costa Rica)

Only small rise

inside pallet

< 0.2 °C per hour

1d 1d6h 1d12h 1d18h 2d 2d6h 2d12h 2d18h 3d

14

15

16

17

18

19

20

21

22

23

24

25

Time in [Days/Hours] in May 2008

Tem

pera

ture

in [°C

]

Pallet Core

Pallet Surface Wall

Return Air Grid

Air Ducts Floor (Supply)

The intelligent container - Cold Chain Management: Home

Documents

Transcript of The intelligent container - Cold Chain Management: Home