Temperature monitoring of non Temperature monitoring of non- - - PowerPoint PPT Presentation
DHL Solutions & Innovations Temperature monitoring of non Temperature monitoring of non- actively cooled pharmaceutical actively cooled pharmaceutical transportation transportation A Amir Mousavi i M i September 2010, Bonn Agenda
DHL Solutions & Innovations
A i M i Amir Mousavi September 2010, Bonn
Introduction Scope Methodology Results Summary Summary Forecast
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 2
I i t t t ll d hi t it
Market situation
Increasing temperature controlled shipment capacity Increasing volume of temperature sensitive goods e.g. life sciences and pharmaceutical products, chemicals and food e g e sc e ces a d p a aceut ca p oducts, c e ca s a d ood
ntity Cargo load
Swap bodies > 10 t
frozen : < - 20 °C
Temperature ranges
Quan
5 – 10 t 2 – 5 t < 2 t
chilled: + 2 °C to + 8 °C ambient: +15 °C to + 25 °C
Year
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 3
Status of temperature controlled transports in Germany Source: Federal Office for Motor Traffic (Kraftfahrt Bundesamt)
Year
Temperature controlled transportations
DHL cold chain services
Temperature controlled transportations FTL / LTL services Partially temperature monitoring vehicles End-to-end temperature monitoring very difficult to realize 87 % are ambient shipments focuses on end to end temperature monitoring
DHL SmartSensor temperature service
focuses on end-to-end temperature monitoring intelligent temperature sensor RFID interface Temperature monitoring on all levels of transportation
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 4
For frozen and chilled products temperature monitoring usually takes place on
Observed monitoring approaches
For frozen and chilled products temperature monitoring usually takes place on shipment level For ambient products temperature monitoring takes place on shipment, pallet, b d t i l l swap body or container level Ambient transportations are mostly not shipped with temperature controlled vehicles In cases one sensor is linked to all shipments in a swap body Temperature data not evaluated on swap body level
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 5
Introduction Scope Methodology Results Summary Summary Forecast
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 6
Questions
How many sensors are necessary to allocate the temperature in a swap body and conclude the whole environment temperature? Which relation is given between number of sensors and the quality of g q y monitored temperature?
Subject of analysis and Measuring device:
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 7
DIN EN 284:2006 Swap body DHL SmartSensor Temperature
DHL SST in a glance
Web portal Sensor Reading device Web portal Sensor Reading device
Data analysis Data logging Data read-out Wireless Wireless
Target Market
Life Sciences
Food and other perishables Chemicals
External Partners
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 8
Introduction Scope Methodology Results Summary Summary Forecast
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 9
Challenges for temperature monitoring of ambient products transported in swap bodies
high cost for temperature monitoring on shipment level high cost for data harvesting and management
Conflict of objectives
difficulty of placement of sensors
Conflict of objectives
Highest quality of temperature data Minimal numbers
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 10
Reason of conflict of objectives: Lack of information
Approach
S l i th l k f i f ti i i t l ti f t t d t Solving the lack of information via interpolation of temperature data Reducing the cost through replacing hardware sensors by software sensors
Method
Adaption of the geo-statistical Kriging method P Pro: Statistical interpolation method Usage of variography to optimize the results Usage of variography to optimize the results Able to solve 3D problems Best linear unbiased estimator (BLUE) Contra: Requests high quality of data Complexity in calculations
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 11
Complexity in calculations
Experimental and theoretical variograms*:
Experimental variogram: Experimental variogram:
p Theoretical variogram: Exponential variogram function
* The variogram is a location-independent method which indicates the mean statistical spread
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 12
g p p
spatial relation between these two variables.
63 sensors for temperature data harvesting
Gathering of temperature allocation
63 sensors for temperature data harvesting in a 3x3x7 matrix alignment 24 hours measuring duration for each test g 14 minutes measuring intervals
Swap body DHL SmartSensor 7300 mm Z DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 13 X
Introduction Scope Methodology Results Summary Summary Forecast
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 14
Mean temperature in the swap body
Measurement of 09.09. Measurement of 22.09. C ure in ° C emperatu Mean te
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 15
time
Mean temperature of swap body outer walls 09.09. – 10.09.
C ure in ° C emperatu Mean te ti
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 16
time
left right roof floor door rear
Optimization of measuring network according to max error- and Kriging-Variance ° C Error in ° no of sensors
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 17
no of sensors
(Error) mean error (Error)
(variance) mean error (variance)
Interpolation: Measurement of 22.09.2010 with 14 sensors
Interpolation error Interpolation error ° C time
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 18
time
mean error max error defined mean error
Introduction Scope Methodology Results Summary Summary Forecast
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 19
Temperature situation in the swap body
Temperature allocation at daytime not constant over daytime high temperature differences through the spatial dimensions h t t ll ti f i ti very homogenous temperature allocation from evening time Temperature differences for loaded swap body must be considered
Temperature monitoring with lowest numbers of sensors
Kriging-method to eliminate information lapse Optimization of measuring network according to the maximum interpolation error delivers good results Defined mean interpolation error is only excided very shortly Defined mean interpolation error is only excided very shortly Results under the given sensor tolerance of ± 0,5°C
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 20
Introduction Scope Methodology Results Summary Summary Forecast
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 21
Further analysis of the Kriging- method Methodology enhancements based on results y g g Analysis of different load situations in swap bodies Analysis of further condition parameter e.g. humidity and shock M d t il d d l b ti f it i
Technological enhancements
More detailed and longer observation of monitoring Risk of increasing technology costs
Mathematical methods can be used to reduce technology costs
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010 22
DHL | Page 4th International Workshop - Cold Chain Management | Bonn | September 2010