Page 68 - FoodFocusThailand No.247 November_December 2026

STORAGE, HANDLING & LOGISTICS

Freezing can also be studied using advanced analytical warning 30 minutes before a temperature deviation occurred,
tools, such as digital simulation, confocal laser scanning with a root mean square error (RMSE) of about 0.31 °C.
microscopy (CLSM), near-infrared spectroscopy (NIRS), It also reduced errors by 18–71% compared with conventional
nuclear magnetic resonance (NMR), and ultrasound. models (Duan et al., 2026).
These tools can be used to assess heat transfer, water 2. Dynamic Shelf-Life Prediction (DSLP): Instead of
distribution and changes in ice crystals during freezing. relying on passive monitoring and fixed expiry dates, this
On the advanced thawing technologies, microwave, approach predicts shelf life based on actual conditions. Smart
ohmic (electrical resistance) and radio-frequency packaging works together with electronic noses, electronic
thawing technologies can make the process faster and tongues and gas sensors to detect volatile compounds
more uniform, with less damage to product quality. These released by food as it spoils. The approach combines three
freezing and thawing technologies are not stand-alone tools. kinds of information: environmental conditions, the food’s
They are being linked with smart logistics systems, so the current quality state and risk limits. Deep learning models
cold chain can follow a continuous cycle: monitor conditions, (such as 1D-CNNs, TCNs and CNN-LSTMs) then estimate
analyse data, predict problems, make decisions and adjust the remaining shelf life, changes in quality and the probability
controls. Once freezing and thawing are under control, the that the product will spoil.
next challenge is monitoring conditions while food is stored 3. Digital Twins and Closed-Loop Control: A digital twin
and moved. This is where prediction comes in. is a virtual copy of a real storage environment. Paired with
model predictive control (MPC), it can simulate and manage
Smart Technologies for Predictive Cold Chain cooling conditions for fruits that continue to ripen after
Management harvest (climacteric fruits). Computational fluid dynamics
There are four main ways to predict problems affecting chilled (CFD), a simulation technique, helps determine where
and frozen foods during storage and transportation, so you to place sensors and shows how cold air actually moves
can act before the stock is lost. through the storage space. Predictions can then be used
1. Hybrid AI Guided by Physical Knowledge: to control ventilation, ethylene gas levels and temperature
This approach combines the laws of thermodynamics with automatically. The result can be better quality, less loss and
deep learning to warn you early about temperature problems lower energy use. Choosing the right technology means
during delivery. The Knowledge-Guided CNN-BiLSTM weighing accuracy, speed and suitability for real-world
model has two components. The first, a Convolutional operations (Wei et al., 2026).
Neural Network (CNN), identifies patterns in airflow and 4. IoT and Blockchain for Proactive Management:
heat. The second, a Bidirectional Long Short-Term Memory Two tools make this possible. Electronic noses connected
(BiLSTM) network, analyses time-dependent data, such as to cloud-based AI can measure gases released by food
cooling cycles and vehicle door openings. The model also in real time, estimate remaining shelf life and flag risks.
draws on several data sources, including the food’s starting A “digital freshness passport” can use blockchain to
temperature (ITF) and the total time the door has been record a product’s history and quality predictions, making the
open (COT). In the researchers’ tests, the model provided a whole cold chain more transparent. This information supports
practical decisions, such as dynamic pricing, prioritising the
distribution of products nearing their expiry dates to nearby
markets, and alerting staff to adjust storage conditions before
problems escalate. Taken together, these four approaches
turn the cold chain from something you check after the fact
into something you can manage ahead of time.
A smart cold chain improves your ability to monitor food
conditions, anticipate risks and make timely decisions, from
temperature control and transport planning to distribution
strategies based on a product’s remaining shelf life. That said,
even the best system is only as good as the data behind it.
Models need to be tested under real operating conditions,
and the people using them need to understand what the
system is telling them. Applied sensibly, these technologies
do more than reduce food loss. They save the resources
used to produce food and protect its quality and value from
farm to consumer.

More Information Service Info C011

68 FOOD FOCUS THAILAND NOV-DEC 2026

27/10/2569 BE 15:01
65-68_Storage_Logis_Ravis.indd 68
65-68_Storage_Logis_Ravis.indd 68 27/10/2569 BE 15:01