Supervision of Banana Transports by the Intelligent Container - PowerPoint PPT Presentation

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

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. MCB MCB SFB 637 Autonomous logistics 2

Online monitoring by the intelligent container  Real-Time remote monitoring of local temperature deviations External  Idea first presented 2006 Communication  Transfer project 2008 + 2009 Gateway Pre-Processing  Online Access Sensor Nodes  Focus on results from field tests  Outlook to future research MCB MCB SFB 637 Autonomous logistics 3

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 MCB MCB SFB 637 Autonomous logistics 4

Installation in Costa Rica Gateway MCB MCB SFB 637 Autonomous logistics 5

Installation in Costa Rica  4 pallets equipped with 4 sensors each S Foam Block S Tier 8 S S S Tier 1 Air ducts MCB MCB SFB 637 Autonomous logistics 6

External Communication  Forwarding data from sensors to web-server  Using the vessel’s email system Satellite Ship Container Web- Server WLAN Internet Wireless Sensors Gateway Email Server MCB MCB SFB 637 Autonomous logistics 7

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 MCB MCB SFB 637 Autonomous logistics 8

Signal attenuation by the fruits  Water-containing goods hinder the radio communication of wireless sensors @ 2.4 GHz y a w 4 2.5 e t a 1 2 3 Distance 0.5 meter G 17 13 9 5  ⅓ of all links 2 Refrigeration Unit Distance to floor in [m] completely failed  ⅓ of all links was not 18 14 10 6 Door End 1.5 available part of the time 19 15 11 7  ⅓ of all links worked 1 well most of the time 20 16 12 8 0.5  Pallet Pallet Pallet Pallet Partly compensated 1 8 14 19 by network 0 0 2 4 6 8 10 12  New hardware? Distance to cooling unit in [m] Packet Rate: >0.75 0.33< <0.75 <0.33 MCB MCB SFB 637 Autonomous logistics 9

The necessity for core temperature measurement  The internal sensors of the aggregate help only little to estimate the banana temperature 28 Pallet core reefer side 26 Pallet core door side Pallet 11 mid Temperature in [°C] 24 Air supply (Aggregate sensor) Air return (Aggregate sensor) 22 20 18 16 14 4 6 8 10 12 14 16 18 20 Days in August 2010 MCB MCB SFB 637 Autonomous logistics 10

Temperature difference over length of container  Bananas are cooled down by the container  Large differences in time required to achieve 17 °C Door-end   28 Reefer Side Container 1 Door Side Container 1 Reefer end 26 Reefer Side Container 2 Pallet Core Temperature in [°C] Reefer side Door Side Container 2 2.4 days Air Supply 24  22 Two containers of Door side same type and 6.35 days 20 year of manufacture 18 16 14 10 12 14 16 18 20 22 24 Time in [days] in September 2009 MCB MCB SFB 637 Autonomous logistics 11

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 Reefer End Door End Age 2 2010 years Maximum Age 12 2009(B) years Age 12 2009(A) years Age 9 2008 years 0 2 4 6 8 10 12 Time to cool down to 17 °C in [days] MCB MCB SFB 637 Autonomous logistics 12

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. Average temperature in [°C] 16.59°C August 16.6 16.53°C 2010 16.42°C 16.4 Hot-spot Left side 16.21°C 16.20°C 16.2 Door 16.11°C Right 16 side 15.80°C 15.8 Pallet 11 Pallet 10 Pallet 11 Pallet 19 Pallet 1 Pallet 1 Pallet 3 15.6 15.4 0 2 4 6 8 10 12 Air supply Distance to cooling unit in [m] MCB MCB SFB 637 Autonomous logistics 13

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? Full Temperature Range 28 Range covered by corner sensors Corners Sensors: [ 5 8 17 19 ] Base Average Box Temperature in [°C] 2.5 26 Extra Sensors: [ 7 13 ] Temperature [°C] 15.9 15.8 16.7 16.0 2 24 Container BH Distance to floor in [m] Cooling Unit 16.2 16.3 16.8 17.8 22 1.5 Door End 20 16.1 16.5 16.7 18.5 1 18 --- --- 16.9 17.9 0.5 16 Pallet Pallet Pallet Pallet 1 8 14 19 0 0 2 4 6 8 10 12 14 Distance to cooling unit in [m] 10 12 14 16 18 20 22 24 Time in Days in September 2009 MCB MCB SFB 637 Autonomous logistics 14

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 28 26 Temperature Temperature in [°C] 24 difference in one Box 22 20 18 16 14 4 6 8 10 12 14 16 18 20 Days in August 2010 MCB MCB SFB 637 Autonomous logistics 15

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 MCB MCB SFB 637 Autonomous logistics 16

Innovation Alliance  New project starts in September 2010  13 partner companies and 6 research institutes Sea-Containers Truck-Transports Software RFID Ethylene (Bananas) (Meat) and Electronics Sensor Bremer Institut für Produktion und Logistik an der Universität Bremen Federal Ministry of Education and Research MCB MCB SFB 637 Autonomous logistics 17

Intelligent data processing  Not only forward data, but pre- High sensor process tolerance  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 MCB MCB SFB 637 Autonomous logistics 18

Relation temperature and quality  No relation between box temperature and defects per box  Further influence factors have to be checked  O 2 , CO 2 , Ethylene  Age at harvest  Micro biological load  Mechanical damages 16.85°C 16.59°C 17 16.56°C 16.57°C 16.53°C 16.50°C 16.42°C 16.38°C 16.35°C 16.27°C 16.21°C 16.20°C 16.15°C 16.11°C 16.07°C Temperature in [°C] 16.5 16.05°C 16.06°C 15.87°C 15.83°C 15.84°C 15.82°C 15.80°C 15.73°C 15.68°C 16 15.56°C 15.42°C 15.14°C 15.5 Rotten Finger Rotten Finger Crown Rot 15 14.5 10 15 20 25 30 35 Button Number MCB MCB SFB 637 Autonomous logistics 19

Relation temperature and quality  Relation to average quality per container  But no statistical evidence  Only on container level  No relation inside one container Reefer End Rotten Fingers 2010 2010 Door End Crown Rot Maximum Anthracnose 2009(B) 2009(B) 2009(A) 2009(A) 2008 No data available 2008 0 2 4 6 8 10 12 0 0.2 0.4 0.6 0.8 1 Time to cool down to 17 °C in [days] Average defective fingers per box MCB MCB SFB 637 Autonomous logistics 20

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 MCB MCB SFB 637 Autonomous logistics 21