It took over 3,000 pouches of spaceflight food, but Timothy Goulette and Hang Xiao ultimately created a mathematical model that NASA will soon use to ensure that its astronauts are eating well.

After preparing and storing the thousands of packets of spaceflight food according to NASA's specifications, the two UMass Amherst researchers came up with a tool that predicts the degradation of vitamins in spaceflight food over time.

Their model, funded with a $982,685 grant from NASA, will help the agency to more accurately and efficiently schedule resupplying trips.

NASA’s food preparation process requires specific formulation and packaging considerations. To study how the vitamins in the space-food degrade over time, Goulette and his colleagues stored the thousands of meal pouches according to the exact recipes.

"We wanted to mirror NASA's process as closely as possible,” lead author (and food science Ph.D. student) Timothy Goulette told Tech Briefs.

"This enabled us to better understand the true degradation of their spaceflight food items and minimize deviations between what we observed and what NASA should expect to see."

Goulette, Xiao, and colleagues investigated how vitamin B1, also known as thiamine, holds up in three options on NASA’s menu: brown rice, split pea soup, and beef brisket. The team discovered that while the brown rice and split pea soup stored at 20 °C demonstrated resistance to thiamine degradation, the thiamine in beef brisket was much less stable, retaining only 3 percent of the vitamin after two years.

a chart showing thiamine content levels in split pea soup from NASA
Thiamine content measured within the Split Pea Soup product stored in three environments: refrigeration, room temperature, and temperature abuse, over a two year period. Fractions of these data were used to develop the predictive model that allows the prediction of thiamine content in the future at any storage temperature.

“Proving the model was as simple as comparing these measured values from two years of storage to what was predicted as early as 12 months prior,” said Goulette.

The modeling tool will be especially important as NASA plans for the first human mission to Mars.

"On their longest duration missions, the need to understand the nutritional content of their foods is paramount,” said the lead researcher. “NASA will be able to use a minimal amount of data to quickly and accurately predict the vitamin content of a given food at any given time at a reasonable temperature.”

In an interview with Tech Briefs, Goulette explained just how accurate and precise his team's mathematical model can get.

Tech Briefs: A UMass press release  referred to the preparation and storage part of the test process as “painstaking.” How was this part of the process challenging?

Timothy Goulette: We wanted to mirror NASA's process as closely as possible. Their process encompasses specific formulation, processing, packaging, testing, and storage considerations, which necessitated a multi-level analysis on our part to plan and execute this process successfully.

Tech Briefs: Why was making that kind of painstaking effort so important?

Timothy Goulette: This enabled us to better understand the true degradation of their spaceflight food items and minimize deviations between what we observed and what NASA should expect to see.

Tech Briefs: How did this idea for a mathematical model come about? What was NASA’s need?

Timothy Goulette: The model was thought up as a tool that could significantly reduce the time, effort, and cost commitments that are required as part of the conventional approach to determining the loss of nutrients over time. NASA is focused on understanding the nutritional content of the foods that they provide to their crew, and how this nutritional content evolves during a mission. Our model allows them to accurately predict those changes within their foods using a fraction of the data required by the conventional method.

Tech Briefs: What were your most important findings from the mathematical model?

Timothy Goulette: We found that the stability of a given vitamin varies based on the food that it is contained in, and that the effect of temperature on that stability is not always the same among foods. From this, NASA will be able to carefully select meals for the crew which will be the most stable in a given environment. This becomes particularly important if refrigeration is not available on the spacecraft, or the food experiences abuse at high temperatures.

Tech Briefs: What specific foods or “menu options” are most stable?

Timothy Goulette: Our research determined that, when it comes specifically to the stability of thiamine, carbohydrate-rich foods such as split pea soup really shined. Another staple such as brown rice offered similar stability, while fatty, protein-rich options such as our BBQ beef brisket were significantly less conducive to thiamine retention over time.

While not published in this work, we also researched how freeze-drying altered the kinetics of thiamine degradation in our foods compared to non-dried counterparts, using our model.

The goal here was to help NASA determine whether or not adding a freeze-drying process to their food production regime would increase nutrient stability significantly. It's traditionally believed that reducing the moisture content of a food system slows down the degradation within that system. Our findings challenge this presumption in some interesting ways. Hopefully, we can discuss that further when that work is published.

Tech Briefs: Will NASA use this tool? And if/when NASA does, how will astronauts use it, do you think?

Timothy Goulette: NASA plans to use this tool as part of their strategy to routinely provide our astronauts with nutritious food on long-duration missions such as the anticipated manned expedition to Mars. On the ground, NASA will use the model to determine the degradation of essential vitamins in the foods that they send with the crew as well as load on resupply shuttles to support the crew over time (which are prohibitively expensive and need to be carefully and deliberately scheduled).

Functionally, this would look like the following: The Mars-bound shuttle is stocked with enough food, calorically, to last the entirety of the trip to the red planet. However, NASA has preemptively used our model to predict the degradation of vitamins B1 and C in the supplied menu options and found that vitamin C will degrade to an unsafe level months before arrival, and that vitamin B1 overall will only last a few months after touch-down within the foods provided. Knowing this, NASA institutes supplementation, tailored dietary interventions, and schedules resupply missions well in advance of the threat of malnutrition. Alternatively, they may decide to install refrigeration aboard the shuttle in order to slow down the degradation found at ambient temperature in their foods. Our model allows them to estimate the benefit, if any, of this strategy, as a factor in the energy utilization and cost logistics of the mission.

Tech Briefs: How are you able to predict vitamin degradation with “high precision?”

Timothy Goulette: The answer is in our approach to modeling. We are able to accurately predict the loss of a given vitamin within these foods over time because we start with certain reliable assumptions about the kinetics of degradation, and then we use real-life experimental data to tweak these kinetics for each food.

By digging into the literature and establishing a set of axioms for how certain compounds break down, a relationship between degradation rate, temperature, time, and the concentration of that compound can be written. That relationship is then modified given the unique degradation that we observe analytically to produce our predictive model. The beauty of this approach is that this modification requires minimal data, and predictions can be made within a wide range of times and temperatures.

Tech Briefs: What is the model exactly? Is it software? Is it an equation?

Timothy Goulette: The model is a predictive equation for nutrient degradation that we first inform and then visualize using a program that we developed through Wolfram Mathematica.

We start with a set of two rate equations with five unknowns in each: the momentary concentration of a compound at a given time relative to time = 0 (when the concentration is 100%), the time when the concentration is measured relative to t = 0, the storage temperature as a constant, the degradation rate of the compound, and the temperature sensitivity of that compound.

We experimentally set or determine the first three parameters, but the other two need to be determined through solving the two differential equations. The program in Mathematica allows the user to determine the degradation parameters (degradation rate and temperature sensitivity) of their compound of interest visually, by virtue of the fact that only one set of degradation parameters will explain the concentration observed at two different times and two different temperatures.

These parameters are then used to inform the model where predictions can be made mathematically as well as visually.

The degradation rate (k) and temperature sensitivity (c) are already determined. When input into the program, the model is then "informed" and the trajectories are drawn in the bottom module. A solution for k and c is found when those trajectories span through both points in the bottom module, representing the experimentally-determined concentrations following different times and temperatures.

The informed model can then be used to show the degradation trend of the compound at any inputted time and temperature.

Tech Briefs: Why is this modeling tool so important?

Timothy Goulette: As was mentioned, the model that we have developed enables NASA to enhance their knowledge of the stability of the nutrients in food provided to the crew for informed mission and contingency planning with the end goal to optimize crew nutrition and support their success. The tool greatly reduces the time, effort, and cost required to achieve this by the conventional method, and offers NASA superior predictive capabilities.

What do you think? Share your questions and comments below.