Evaluation of an autonomous sensor swarm for fermentation reactor applications

Advanced monitoring of the spatio-temporal distribution of process parameters in large scale vessels and containers such as storage tanks as well as stirred chemical or bioreactors offers a high potential for enhanced investigation and further optimization of plants and embedded processes. This pertains especially to fermentation biogas reactors, where a number of process parameters, such as the temperature profile, distribution of pH, gas-liquid fraction in the substrates as well as flow characteristics, such as velocity profiles, dead zone locations and short-circuits of liquids, are of interest to engineers and operators. Autonomous sensor concepts enable the metrological acquisition of spatially distributed parameters by means of intelligent instrumented flow followers.
We developed and tested the concept of an autonomous sensor swarm that can be introduced into a process vessel to track the long-term spatial distribution of process parameters [1]. The prototype swarm comprises of robust and neutrally buoyant capsules (diameter 42 mm) each equipped with a measurement electronics that autonomously measures and records the output from miniaturized onboard sensors for temperature (0 to 70°C), pressure (0 to 200 kPa with immersion depth in the range of 0 to 10 m) and 3D-acceleration (±6g).
The performance of the sensor capsules were firstly evaluated in a fermentation reactor environment . A swarm of seven capsules was deployed in a 1000 L vessel of a stirred model fermenter. A highly viscous aqueous solution of straw was used with a dry mass concentration of about 5.5%, density 950 kg*m-3, viscosity 250 mPa*s at a shear rate of 10 s-1 and constant ambient temperature T = 19°C. The central three-blade impeller stirrer with a diameter of 0.324 m was adjusted at a rotation speed of 4.4 s-1. Thus, the capsules faced a maximum rotational speed of 4.5 m*s-1. After one hour of operation, the impeller was shifted from 200 mm above vessel ground to 324 mm along the mixer’s shaft to simulate varying mixing conditions. The sensor swarm was recovered after two hours of residence in the process environment.
All acquired data from the seven capsules were analyzed and they properly represent the conditions in the vessel. Temporal evolution of the vertical flow component can be observed from the capsule’s immersion depth which is calculated from the measured pressure. As mentioned above, the process temperature was kept constant at 19°C which was captured by the swarm correctly. However, also vertical temperature profiles may be extracted using the measured immersion depth, which was not reasonable under these isothermal conditions. Additional information about the fluid dynamics, the mixing behaviour and the distribution of dead zones are obtained from the recorded acceleration data. Changes in the setup such as the modification of the vertical impeller position are also reflected in the data.