Two temperatures describe two different positions in heat transfer

Inside a roaster, the heated air or environment and the coffee beans continuously exchange heat. So ambient temperature and bean temperature aren't the same quantity. A sensor placed in the airflow shows the conditions acting on the bean; a sensor reflecting bean temperature describes the state of the material being roasted.

Research by Hernández and colleagues simultaneously tracked internal bean temperature, external temperature, and the heating environment's temperature. The curves don't overlap, and that very difference is information about the heat-transfer process.

Thermal lag explains why the two curves aren't alike

When green beans enter a hot machine, the bean absorbs heat and its temperature can't instantly equal the ambient temperature. Heat has to transfer through the surface and into the internal structure. At the same time, water evaporates and roasting reactions consume or release energy.

Bottazzi and colleagues built a numerical model of the roasting process based on heat and mass transfer, in which bean temperature is calculated from the machine's operating conditions. This shows bean temperature is a result of the whole system, not a copy of ambient temperature.

Recording only one quantity can lose context

If a profile only has ambient temperature, the reader knows what thermal condition the machine is supplying but struggles to know how the bean responded. If only bean temperature exists, it's hard to assess the level of environmental impact and the machine's controllability.

In some systems, the bean-temperature sensor can carry error due to placement or measurement method. So the profile's value doesn't lie only in the reading, but also in knowing where the sensor sits and how it's calibrated.

The heat curve needs to be read alongside time and rate of rise

The same final temperature doesn't guarantee two batches share the same thermal history. Two curves can reach the same endpoint but differ in time, rate of temperature rise, and when roasting markers appear.

Hernández and colleagues also used time alongside temperature to describe the bean's progression. When analyzing a roast profile, it's therefore worth looking at least at temperature, time, and rate of change, rather than saving just one final value.

What data has value when comparing two batches

If the goal is checking consistency, a useful profile should include ambient temperature, bean temperature if reliably measurable, time, rate of temperature rise, and important operating markers. Charge weight and raw-material conditions also need to be kept or recorded.

This data lets you distinguish two situations: the machine supplied different conditions, or the machine was similar but the beans responded differently. With only one heat curve, tracing back becomes harder.

Don't turn one temperature into a "standard" for every machine

The recorded temperature depends on the equipment, sensor position, heat-transfer method, and operating configuration. A value measured on one machine can't automatically be considered equivalent on another.

What can be standardized better is the measurement and reading method. You need to define sensor type, position, recording timing, units, sampling frequency, and how deviations are handled. When measurement method is consistent, data across batches becomes comparable.

A good roast profile has to show the relationship between environment and bean

The value of the two temperature groups lies in the gap and relationship between them. Knowing how the environment changes and how the bean responds gives the operator more basis for understanding heat-transfer rate, identifying deviations, and adjusting the process.

So instead of picking one temperature number to represent the whole batch, build a profile that lets you see the progression. This is how thermal data shifts from an operating log into a process-control tool.

Clearly recording the measurement method also helps explain unusual changes. A temperature jump could reflect a real fluctuation in the process or a change in sensor contact. Without information about the measurement, it's very hard to distinguish these two cases.

Two sensors both measuring "temperature" but placed in two different positions can give different results because they're observing two different zones of the system. So when building a roast profile, the quantity's name should be accompanied by its position and sensor type. If a sensor or its position changes, the new data should be flagged to avoid comparing it directly with the old series.

Sensor and measurement position determine what a number means

When analyzing the heat curve, sampling time should also be recorded. Two systems with the same start and end points but different recording frequencies will produce profiles with different levels of detail. A consistent recording method helps calculate rate of temperature rise and compare markers between batches more reliably.

When thermal data is used to investigate a deviation, it's worth looking at the shape of the whole curve rather than just taking the highest or lowest temperature. A batch can reach the same final temperature but rise too quickly at one stage or too slowly at another. These differences can relate to the heat-transfer progression and the bean's state, so a time-based profile is more valuable than a few isolated numbers.

References

• Hernández et al. (2007). Analysis of the heat and mass transfer during coffee batch roasting. Journal of Food Engineering, 78, 1141–1148. https://doi.org/10.1016/j.jfoodeng.2005.12.041
• Bottazzi et al. (2012). A numerical approach for the analysis of the coffee roasting process. Journal of Food Engineering, 112, 243–252. https://doi.org/10.1016/j.jfoodeng.2012.04.009
• Hernández, Heyd & Trystram (2008). Prediction of brightness and surface area kinetics during coffee roasting. Journal of Food Engineering, 89, 156–163. https://doi.org/10.1016/j.jfoodeng.2008.04.026