New symbiotic starter culture investigated for production of original yogurt

Yogurt production is traditional branch in our food industry. Yogurt is a fermented milk product that contains the characteristic bacterial cultures Lactobacillus bulgaricus and Streptococcus thermophilus. One of them is invented and used for the first time in our country and that's why it is named after Bulgaria. Yogurt starter culture is a symbiotic combination of strains of bacteria mentioned above without genetic modification. One of the main parameters that characterize starter culture quality is pH, which is measured on-line in industry. A control system of starter culture pH value is proposed. To guarantee the quality of pre-fermentation product, the pH value is maintained in admissible range. The proposed system is a part of contemporary high-technology in the manufacture of fermented dairy products.

A new highly effective symbiotic starter culture is investigated. For production of the original yogurt, natural strains S.thermophilus 13а and Lb. bulgaricus 2-11 are used. They are isolated from home-made yogurts manufactured in Rhodopes - mountain region in Bulgaria. The new control system is developed on the basis of experimental data - laboratory measurements of lactose, lactic acid and biomass concentrations, as well as on-line measurements of starter culture pH. A mass-balance model is derived using experimental data. The dynamics of each variable is presented by differential equation. The procedure applied for model identification is realized using program package on the basis of Evolutionary algorithm that minimizes mean square errors between model and experimental data. From expert's point of view, a mathematical relationship between lactose concentration and pH value in the reactor is determined. This relationship gives the possibility the dynamical equation of pH to be derived. In this way, the relationship between the control output (pH) and control input - the dilution rate (D) is received.

A software sensor of pH variations rate is synthesized using pH on-line measurements. The software sensor is a program module with one input (pH on-line measurements) and two outputs. The first of them gives the estimates of pH, and the second one - the estimates of unknown parameter pH variations rate. The software sensor tuning is realized off-line through appropriate choice of its design parameters. The tuning includes stability analysis and optimization procedure with criteria minimum mean square error between input values and its estimates. This procedure guarantees the good observation of the unknown parameter. A new adaptive and indirect control algorithm is designed. The adaptation comes from the new on-line information (received by software sensor) included in the structure. Yogurt starter culture acidity (pH values) control by feeding with lactate characterizes the algorithm as indirect one.

In the figures the simulation investigation results of proposed pH control system are shown. The process variables dynamics is predicted by mass balance model simulations (black lines). Step changes of set-point of control output are applied according to the process technological requirements. A comparison between the model and estimated pH and pH variation rate values (red points) are shown in figures a) and c) respectively. As can be seen, the observation's values coincide with the model's ones. In figures b) and d), the control input and the concentration of lactoce in the reactor are shown. The curves follow the step changes of the control set-up. Step control is realised in the admissible range of pH variations: 0.95 corresponds to 5.7 pH and 0.85 to 5.1 pH. As it can be seen in figure a), the output pH follows the set point changes in both directions with high accuracy.

The good simulation results of proposed quality control system for yogurt starter culture production are precondition for testing of the proposed algorithms in laboratory conditions and after that to be proposed for industrial approbation.