Skip to main content

A real-time forecast decision support system for the food supply chain

Periodic Reporting for period 1 - FreshProof (A real-time forecast decision support system for the food supply chain)

Reporting period: 2017-12-01 to 2019-11-30

Food security is a current worldwide challenge faced by the uncertainty of meeting future food demand for the increasing population. Increasing food production is challenging, and reports have revealed that reducing food waste has the potential to address the food security challenges. It is believed that about 30 to 50% of the food produced in the world never reaches the consumer. A key cause of commercial food waste is under-addressed supply chain issues. In the case of fresh fruit and vegetables, almost 50% of produce are lost due to temperature related spoilage.

Freshproof is an innovative systems approach aiming to address the existing food supply chain losses/waste and the overall shortcomings in food safety, integrity, and traceability. Current commercial product waste reducing strategies have many limitations (the interlinked complexity of qualitative and nutritive aspects of food is ignored, constant postharvest environmental conditions are assumed, and pre-harvest environment is completely disregarded). Current commercial systems offer fragmented solutions and lack the capability to apply a holistic perspective to supply chain integrity.
Freshproof aims to address these issues and provide a system capable of predicting the remaining shelf-life of products as they progress through a full supply chain. It will be based on First Expired First Out strategy (FEFO), incorporate pre- and postharvest conditions, and exploit novel data capture sensor units combined with advanced modelling algorithms.
The main objective of this project is to develop a cloud based forecast decision support system to deliver real-time food product shelf-life prediction along the farm-consumer supply chain. Initially FreshProof will be developed for strawberries due to their highly perishable nature; forming an ideal product to evaluate system robustness. However, the functionality of FreshProof, once demonstrated, can be applied across a wide range of agri-food products and related supply chains.
To achieve this objective, Freshproof will:
1. Explore the effect of pre-harvest conditions on shelf-life and quality of strawberries; using data capture sensor units to collect environmental data during harvest
2. Combine the effect of postharvest conditions on shelf-life and quality by simulating supply chain conditions
3. Monitor the biochemical properties (relating to the nutritional value) of strawberries as they progress through the supply chain and reach the consumer (simulation)
4. Use machine learning techniques to build accurate shelf-life prediction models

FreshProof will promote positive change across the European food industry by enabling stakeholders to proactively identify problematic loads and act to minimize losses. This research will positively impact: a) Primary producers: reduce monetary losses from rejected loads; b) Distributors/retailers: plan distribution based on product shelf-life; c) Consumers: establish a complete product chain of custody ensuring the highest quality, safe and nutritious products are supplied to consumer.

A system such as FreshProof will advance protection of the environment through promoting sustainability and increased efficiency in food distribution chains; strengthen food security by reducing existing waste and losses through the supply chain; enable sustainable food systems through digitization of the supply chain.
Work performed to date facilitated the completion of 3 specific objectives (Work packages 1-5) of the project as detailed in Annex 1 of the Grant agreement. The first specific objective (SO) of the project was to establish cyberphysical linkages to enable the collection of pre and postharvest data. This was achieved by the application of multi-sensor units at growing sites and in the lab. Also, a simulation of the supply chain was designed, based on temperature scenarios encountered in each step of the supply chain. The simulation was achieved using temperature and humidity-controlled chambers, monitored with sensors recording real-time temperature and relative humidity data. Pre-harvest and postharvest environmental conditions data were successfully collected and analysed.
The experimental work took place in Florida, USA, where strawberry harvest occurs between late November to mid-March. In accordance with specific objective 2 (Shelf-life mapping) strawberries were subjected to lab simulated shelf-life assessment including multiple temperature scenarios (Fig.1). Shelf-life was determined across a series of quality, nutritional and microbial parameters. Also, a state-of-the-art hyperspectral camera was employed to explore the potential of predicting shelf-life using hyperspectral imaging technology. Arising from this work a dataset was built consisting of quality and spectral data measurements that determine the shelf- life of strawberries. Machine learning algorithms and multivariate data analysis were employed to identify shelf-life limiting factors, exploit synergies between pre- and postharvest parameters on strawberries shelf-life and to deliver a robust shelf-life prediction model.
It is envisaged that FreshProof will progress the state of the art by integrating multiple environmental, quality, safety, and nutritional parameters of perishable food products into a single decision support tool. The unique approach of combining both pre- and postharvest conditions in a single decision support tool and implement prediction models to determine the remaining shelf life and nutritional value of a product to consumer level, will support smart agri-food supply and sustainability in food systems. The potential positive economic impact, stems from successfully addressing supply chain inefficiencies which can reduce preventable food waste (€235 billion p.a. potential saving), and address food security challenges relating to accessibility and availability of food for the increasing population.
In addition, the results from the hyperspectral imaging technology trial have the potential to establish this technology as a tool for quick, non-destructive, and accurate self-life estimation. This will be a turning point in supply chain management, as it will remove current subjective industrial practices of estimating shelf-life and flawed supply chain management systems. It will enable informed and accurate decision making and strategy planning that will benefit all stakeholders in food supply chain.
Hyperspectral camera
Supply chain simulation