Optimize Extraction Progress Using Bionetics Thermal Property Sensors
This study examines the feasibility of an inline thermal property sensor as an indicator of extraction progress. It is proposed that this non-intrusive sensor can provide important process information at a lower cost and with an easier integration than other currently available sensors.
A recirculating ethanol extraction was performed to compare the outputs of an Bionetics Thermal Property Sensor (TPS) and a Coriolis meter acting as a control. The density, thermal properties, and flow rate were recorded throughout the test. Results showed that an Bionetics TPS can effectively track extraction progress and provide important information to the operator including flow/no flow conditions, if oil is present in the solvent, the rate of extraction, and most importantly, when extraction is completed. A TPS can provide real-time measurement which allows an extraction operator to increase yield and decrease waste while having minimal effect on the system design.
Introduction
An efficient and profitable extraction operation is achieved by maximizing the amount of extracted material and minimizing the cycle time of the process. This is accomplished by accurately identifying completion of the process – the point at which additional time does not result in significantly more yield. Fixed-time extractions based on trial-and-error test runs are unable to adjust to varying process conditions that affect extraction rate, such as biomass potency, temperature, or milling fineness. A change in any of these process conditions without a corresponding change in extraction time either decreases cycles per day by running the extraction too long, or increases waste by stopping before the extraction is complete. Therefore, an improved process indication that identifies completion of extraction is desirable to maximize yields, increase throughput, and decrease waste.
There are several types of common sensors that can measure fluid properties. These include tuning fork type sensors that oscillate a paddle in the fluid to measure density and viscosity, Coriolis sensors that measure mass flow and density by oscillating the flow tube itself, and optical sensors that measure light absorbed by the fluid. These sensors all have significant drawbacks when considered for use in an extraction operation. Coriolis type sensors are very expensive and unable to operate at high pressures, while tuning fork and optical type sensors are prone to sedimentary fouling and calibration drift.
A thermal property sensor has the unique advantage of measuring the fluid from outside the flow tube, allowing it to operate with a low pressure drop and a resistance to sediment buildup. This unobstructed approach to fluid measurement is why TPS style sensors have been used in nuclear and aerospace applications that require a high degree of reliability and long lifetime. Unlike many other fluid property sensors, a TPS can withstand high-pressure and high-temperature operating conditions.
Thermal property sensors operate by measuring the rate of heat loss to the fluid. This heat loss rate is affected by flow rate, density, thermal conductivity, viscosity, and heat capacity of the fluid. A typical application of this technology is to measure the flow rate of constant property fluids. When volumetric flow rate is constant, TPS technology can be used to measure change in fluid properties. Many extraction processes naturally operate at a steady volumetric flow rate because they use constant displacement pumps such as gear, vane, piston, or diaphragm pumps, allowing easy integration of TPS sensors after the extraction vessel. In addition, since the properties of the solvent and extracted oil are so different, a TPS is very sensitive to changes in fluid composition during an extraction operation. In fact, because the TPS measures a combination of thermal properties it is often more sensitive to changes than a density sensor alone.
Experiment
An experiment was conducted to test the ability of a TPS to measure extraction progress by indicating changes in fluid thermal properties. An ethanol extraction system was built, shown in Figure 1. The system contained an open extraction vessel where biomass was continuously washed with ethanol, a vane pump operated at a constant speed, a TPS sensor, a Coriolis sensor, and a bypass.
The bypass was included to ensure that the extraction was done slowly, only allowing a fraction of the recirculating ethanol to contact the plant matter at a time. This allowed close tracking of the extraction process.
The TPS sensor reads flow rate accurately in lbs/hr when flowing pure ethanol. The flow rate of pure ethanol prior to extraction is measured and recorded as the baseline reference. Any deviation from the baseline value is in proportion to the amount of oil extracted.
During the experiment, TPS data was collected as well as flow rate and density from the Coriolis meter. This data was compared to validate TPS reading as an effective way to track extraction progress.
Ground and pelletized Hops were selected as a suitable plant material for testing. Hops are readily and legally available in Ohio as well as taxonomically very close to cannabis, both coming from the same family, Cannabaceae. The hops were loaded in a 25-micron mesh filter bag which was then placed in the extraction vessel. The TPS and Coriolis meter, along with a length of tubing, were placed in a room temperature bath for thermal stability. This bath ensured that any heat added to the fluid by the pump would dissipate and therefore not affect the experiment. The system was started and allowed to reach thermal equilibrium by running for about 30 minutes. The extraction was started by opening the valve above the extraction vessel, allowing the ethanol to flow over the Hops. During the experiment, hops were periodically agitated to ensure even washing. The fluid was constantly recirculated through the system, increasing in oil content and density until the extraction process was deemed complete.
Results and Discussion
To determine the amount of extract in the system, the TPS indication should be compared to the baseline flow rate of the system. Each unique system design will have a unique baseline flow rate which is determined by using a “dummy” load in the extractor. This dummy load provides a similar flow resistance to an extraction and can be either inert biomass (fully extracted of all oils), or another flow resistance such as glass beads. The TPS indication is calibrated to the mass flow of whichever pure solvent is selected for an extraction, in this case ethanol. When flowing the pure solvent, the TPS provides an accurate measurement of flow rate. The baseline flow rate is established by using the dummy load, then allowing the system to come to equilibrium and recording the TPS indicated flow rate.
During an extraction operation, as oil is dissolved into the solvent, the TPS indication deviates from the baseline flow rate. This deviation, or flow rate error, is proportional to the extracted oil in the line at a given time and is a result of the difference in the thermal properties of the pure solvent and the solvent/extract mixture. In addition, the rate of change of the TPS indication is proportional to the extraction rate, signifying a complete extraction when the rate of change approaches zero.
The results of the experiment are shown in figure 2. The baseline flow rate of the system was determined to be around 460 lbs/hr. The extraction began at point “A” around 500 seconds, with the initial density of ethanol at 0.787 g/cc. Soon after the extraction started, density began increasing as oil was extracted from the hops. In figure 2, the TPS indication is plotted with volumetric flow rate. The volumetric flow rate of the experiment is approximately constant due to the use of a vane pump, which was driven at a constant rpm therefore delivering a constant volume of fluid per second. During the experiment, the TPS indication decreased from the baseline in a proportional manner to the increase in density, successfully indicating the change in fluid properties during the extraction. The extraction was deemed complete when the rate of change of the TPS indication approached zero, at 5000 seconds and 410 lbs/hr.
Two external influences were seen to affect the test signals during the experiment. First, when the dry pellets begin absorbing ethanol, the pressure head on the pump decreased causing flow changes noticeable in Figure 2 at point “A”. Second, agitating the hops produced short duration signal fluctuations as seen at point “B” in Figure 2. These signal disturbances are of short duration and much smaller in magnitude than changes caused by oil concentration. To compensate for these disturbances and better represent a closed extraction system steady flow rate, the TPS indication can be plotted as the percent difference of the baseline flow rate as shown in Figure 3.
Representing the TPS indication as a percent difference of baseline flow rate as in figure 3 allows us to directly compare the TPS output with the measured density of the Coriolis meter. Both meters indicate the beginning of the extraction at 500 seconds and level off around 5000 seconds.
Conclusion
Figure 3 shows that the Bionetics TPS indication can be used to track extraction progress in the same way as other more established methods. The indication begins to change as soon as oil is introduced to the system and levels off as the rate of extraction slows. This allows tracking of important process parameters such as, flow/no flow conditions, when oil is present in the solvent, the rate of extraction, and finally, when extraction is completed. The TPS is naturally sensitive to changes in fluid composition due to sensing thermal conductivity, heat capacity, and viscosity, and because of its non-intrusive and cost-effective design it shows a great potential to be used in increasing the efficiency and decreasing the cost of extraction systems.