| 1. |
- Janestad, H, et al.
(författare)
-
Modelling of dynamic flavour properties with ordinary differential equations
- 2000
-
Ingår i: Food Quality and Preference. - 0950-3293 .- 1873-6343. ; 11:4, s. 323-329
-
Tidskriftsartikel (refereegranskat)abstract
- The most common way to analyse sensory dynamic measurements (time-intensity, TI) is to extract some characteristic parameters from the resulting curve such as 'intensity maximum' and 'area under the curve'. In order to get more information from TI data, a general mathematical model was developed. The model was based on the theory for ordinary differential equations. The solutions were characterised by their eigenvalues, which might be correlated to recipe and process. As an example, the temporal perception of chocolate flavour has been measured and modelled. In addition the classical characteristic TI parameters could easily be calculated by the model.
|
|
| 2. |
- Wendin, Karin, 1963-, et al.
(författare)
-
Modelling and analysis of dynamic sensory data
- 2003
-
Ingår i: Food Quality and Preference. - 0950-3293 .- 1873-6343. ; 14:8, s. 663-671
-
Tidskriftsartikel (refereegranskat)abstract
- Time intensity (TI) data from earlier reported studies on cream cheese and salad dressing were used to develop models based on both polynomials and ordinary differential equations (ODE) that can be used to describe and interpret TI-data. Polynomials were thus fitted to experimental data. By taking the first and second derivatives of the polynomials one gets new polynomials that express how the perceived intensity changes with time. By integrating the original polynomial one gets a new polynomial that expresses how the classical TI-parameter "Area Under the Curve" is accumulated with time. Graphical display of all these types of polynomials gives an immediate and easily interpretable impression of the influence of different experimental factors on the time dependent perception. In the ODE models experimental factors, both formula and process conditions, were taken into account. Thus it was possible to develop equations that can be used for prediction of TI-curves for intermediate experimental settings.
|
|
| 3. |
|
|
| 4. |
|
|
