issue: April 2009 APPLIANCE Magazine
Appliance Engineer Feature
Designing Better Ovens—Faster
Email this Article
David Ward, R&D director, Fulgor Appliances (Gallarate, Italy); and Carlo Brutti, Luca Andreassi, and Marco Evangelos Biancolini, Università di Roma Tor Vergata (Rome, Italy)
This article discusses three possible methodologies for designing domestic ovens.
Figure 1. Simulated ducts (above) and oven cavity (below).
Correct oven cavity design is at the heart of
all good self-cleaning ovens. In fact, this design element is often the
starting point for new oven projects. Some of the key challenges facing
engineers include complying with cost budgets (especially those
concerning materials and investments) and performance standards, as
well as many other nontrivial constraints such as reliability,
certification, noise, and manufacturing processes.
paper describes design methodologies applicable both for self-cleaning
and non-self-cleaning domestic ovens based on the following three-step
simulation of door and air-cooling systems/devices (external portion of
the oven cavity such as the electronics bay) so as to establish, as a
minimum, the external thermal conditions of the door (the largest heat
dispersion source for ovens) and areas where important cooling
- The thermal-fluidynamic
response of the cavity in loaded and unloaded conditions and using real
cooking modes (e.g., convection roast, thermal bake, etc.).
thermal-mechanical response of the cavity structure coupled with the
fluidynamic simulation. In this way, engineers can appreciate the real
response of the system with and without thermal load (i.e., real food).
approach is not dependent upon the firm’s product development process
(e.g., technology-push, market-pull, concurrent design, etc.) since,
once in place, it can be iterated on demand. Interestingly, the results
not only provide better knowledge of the product, but also can
strengthen intellectual property and verify or quantify tacit knowledge.
Initial Design Process
general, domestic ovens are designed from the outside-in: First,
aesthetics are established and then product characteristics are fixed.
Once this is complete, product specs are issued, and the design
engineers start work. A typical draft spec will usually include the
For example, this could include flat brushed stainless-steel trim with
black glass for oven door and control panel; black speckled enamel and
silver door gasket; and large rounded handle with brushed stainless
Possible control specs might include a single multifunctional display
that provides clear, clutter-free feedback for the end-user. Modern
oven controls nowadays are electronic with oven settings established
using one of the following three methods: touch control panel with a
full keypad and no knobs, i.e., nonmechanical selectors or encoders;
semitouch control panel, i.e., a reduced keypad and a series of
selector switches and/or encoders with knobs; and no touch control with
only selector switches and/or encoders with knobs.
Overall Measures and Weights.
This might include cabinet cut-out size; cavity volume and dimensions;
service requirements such as installed power, energy consumption; and
Possible specifications could be warm-up time, self-clean cycle
duration, self-clean cycle temperature, energy class, cooking
performance, and noise, etc.
This might include enameled or stainless-steel racks, cookie trays,
meat probe, telescopic cooking racks, and additional halogen lighting.
For example, this could cover launch and rollout dates, quantities
(volumes for each market), final product cost, selling price, and
Product Development Phase
those details are finalized, it is customary to start the product
development phase. In the suggested approach, this implies using one of
the following three methods.
feedback and in sequence. In simulating the door and key externalities
(e.g., electronics, rear external wall, etc.), the authors conducted a
standard fluidynamic simulation (CFD) of the cooling system, with
special attention to the door, the electronics bay, the catalyzer duct
and cooling of oven fan(s), and control panel. This work was conducted
with FLUENT software, but was also supported via a specific program
developed in MathCAD. All mechanical drawings, including the cooling
ducts and parts, were drawn with Pro-Engineer Wildfire 3.0.
the next phase, the thermal-fluidynamic conditions were simulated with
FLUENT using a finite volume methodology. This was conducted in empty
cavity conditions (unloaded cavity) and in loaded conditions (i.e., a
simulated standard energy consumption “refractory brick”). With the
thermal-fluidynamic response in hand, it was possible to move to the
third and final step, which consisted of simulating the
thermal-mechanical response of the cavity structure (using NASTRAN)
under the conditions established in the previous step (FLUENT). In the
authors’ opinion, this approach is the most complete and beneficial,
but it requires dedicated resources (outsourced or in-house) for a
specific allocated time, which may not always be possible for all
Coupled, without feedback and in
sequence. This approach is almost identical to the previous method. The
difference is that there is no feedback, and the MathCAD worksheet can
be avoided, provided that essential duct and door operation data are
known or set (e.g., temperature contours, air velocities, pressure
Uncoupled and stand-alone
(results can be grouped or pooled at the end). The third approach
splits the design steps into at least two parts (i.e., steps 1 and 2,
or steps 2 and 3, or step 1 only). Because each of the three steps can
be conducted separately, this approach affords two main advantages.
First of all, it can significantly reduce development time and costs
(since each step can be outsourced separately), although it still
implies that the results should be pooled together. In essence, this is
typical of the “work-package” approach. Secondly, it can be applied for
singular design purposes (e.g., determining mechanical stress on cavity
walls, improving catalyzer duct performance, predicting energy
consumption, warm-up time, etc.). This is convenient when the oven is
already far down the development process.
are, however, two main drawbacks. The first is that the knowledge is
pooled at the end rather than shared and coupled during the development
process. In addition, tacit knowledge tends to remain within each
department involved. Both of these aspects may add cost and extend
development time for future projects.
Figure 3. Detail of typical mesh of oven cavity.
Examples of Design Input and Output
MathCAD computation worksheet was set up to deliver three types of
information based on a monodimensional conceptualized system of ducts:
pressure drop and velocities in each section of the cooling duct(s).
Using the coupled with feedback and in-sequence approach, data were
verified to test the complete simulated model of the cooling system
(door included) simulated with FLUENT.
- Estimate temperature domains in each part of the cooling system and ducts.
- Estimate heat transfer in each part of the cooling system(s) and ducts.
the data included (and modifiable), there was the characterization of
cooling fans, duct geometry, etc. The geometric model was directly
obtained from the original Pro-E model of the complete oven. Figure 1
shows the simulated collection of ducts and cooling system (left) and
half of the oven cavity (right).
part of the MathCAD worksheet is the section concerning input and
output data. The input data array shown in Table I refers to one of the
ducts. The data array shown in Table II represents a typical output for
pressure drop calculations.
Output for the
cooling system thermal-fluidynamic simulation might look similar to
Figure 2. This kind of output can be conducted in steady-state
conditions or, more rarely, during transitory. Operating conditions can
be adjusted to take into account products working at altitude or in
very hot conditions.
While only briefly
summarized, these three examples of computed information are typical of
step 1 of the suggested approach. Clearly much more information can be
obtained and provided as input into the design process (i.e., the
characterization of moving cooling fans).
Figure 4. Simulated transitory vertical plane.
Taking It a Step Further
designing domestic ovens, this is usually where product manufacturers
stop and continue development using an experimental approach. However,
if the next two steps are followed, the design engineer has the
opportunity to discover much more about the full operation of the oven.
3 shows the meshed domain (finite volume) of a section of the oven
cavity. (The outline of the grilling or broil element is visible.) The
meshing was applied directly to the exported drawing file from
Pro-Engineer and, together with the data obtained during both
experimentation and the previous step, was used to set the boundary and
initial conditions of the thermal-fluidynamic simulation.
should be noted that it is also possible to include cooking racks and
thermal loads in the mesh, although for empty cavity simulations,
nothing was actually included. In load conditions, it is customary to
emulate the cooking rack as a plane, but as stated, it is quite
possible to include the rack(s) as well. This is worth noting because
rack mass is not negligible for certain thermal simulation conditions
since ovens often work at temperatures in excess of 325°F (163°C) and
typical mass is at least 1.5 kg for U.S. ovens.
typical requirement for design engineers is to predict warm-up time or
guarantee warm-up performance. Ideally, this implies predicting
performance for different cooking modes or for the self-clean cycle. In
order to achieve this, it is necessary to conduct simulations where the
relative transitory are assessed. This provides the design engineer
with not only a measure of the temperature rise gradient, but also how
the heat is dispersed and circulated within and across the system. In
this way, thermal currents are quantified and qualified. Figure 4
depicts a typical vertical plane showing thermal currents within the
As previously mentioned, when
this type of analysis includes transitory conditions, it is possible to
predict and compare oven warm-up performance, as shown in Figure 5
(experimental versus predicted). Similarly, this analysis can be
conducted at any plane in the oven cavity, as shown in Figure 6.
possible output concerns the structural response of the oven,
especially the oven cavity. As suggested, it is possible to carry out a
complete coupled simulation in which the thermal-fluidynamic response
is combined with a complete structural simulation both for steady-state
and transitory conditions. Such circumstances can also include
different structural conditions (e.g., hyperstatic or isostatic, simple
supported, fixed-end, etc.), as well as linear and nonlinear materials
Figure 7 shows a typical output in
two consecutive time frames for temperature; hence, this is a typical
transitory. Similarly, we can have the stress contours and deformation
contours, as shown in Figures 8 and 9. Since temperature, stress, and
deformation are obtained in steady-state and transitory conditions, it
is possible to plot, in time, these results and, therefore, conditions
for peak stress, yield conditions, etc. Figure 10 shows a typical
example of maximum cavity stress plotted versus time for two design
In a concluding note, it is also
possible to simulate the presence of thermal loads, such as the
previously mentioned “refractory brick” used for energy measurements in
Europe. Figure 11 shows a typical output for this type of simulation.
three approaches discussed all provide reliable and fast results, thus
saving product development time and costs. They also offer the benefit
of more knowledge and understanding of the product.
approaches are applicable to companies of all sizes. It is not
necessary to have in-house, expert engineers or a dedicated department
that specializes in such work. However, in order for the organization
to learn, it is necessary that the knowledge acquired is shared and
that any work outsourced uses appropriate project coordination and
This article was submitted for publication in January 2009.
Incropera, FP and De Witt DP. Fundamentals of Heat and Mass Transfer. 3rd ed. New York: Wiley, 1990.
Ward, D. “Thermal Modelling of Oven Cooking and the ‘Brick’ Test.” APPLIANCE Magazine—European Edition, January 2003, 16–18.
Young, WC. Roark’s Formulas for Stress and Strain. 6th ed. New York: McGraw-Hill, 1989.
Figure 5. Comparison of experimental and simulated transitory
Figure 6. Thermal-fluidynamic simulation of cooking planes.
Figure 7. Coupled thermal-fluidynamic and structural simulations (transitory).
Figure 8. Stress contours during thermal transitory.
Figure 9. Deformation contours during thermal transitory.
Figure 10. Stress versus time for two different cases.
Figure 11. Thermal response of refractory brick.