The conclusion of the TAP project (converting industrial waste heat into electricity) carried out last year is that the thermoacoustic process, the conversion of waste heat into acoustic power, can be scaled up in power unlimited. This however does not hold for the conversion of this acoustic power into electricity by linear alternators. In linear alternators the stroke of the moving mass  is limited (by the springs). Consequently, they need to be operated at high frequencies (50-100 Hz) to get sufficient output power per unit mass.  Linear alternators works fine at small systems (≈ 1 kW) but fails to perform at industrial power levels  (→ 1 MW).

Size of the acoustic circuit (feedback, housing) goes up proportionally with power, not only because of the increasing system cross-sectional area but also because of longer bends. The later is a consequence from the acoustic and practical condition that the feedback tube diameter (D) << λ and bend radius (r) > 3D. Summarizing, the higher the power level of the thermoacoustic system, the lower the oscillation frequency.

For linear alternators a low oscillating frequency is disastrous because maintaining the electromagnetic induction (dΦ / dt) a larger stroke and/or a stronger magnetic field is required. This dramatically increase complexity and cost. Both because of the impossibility to maintain the seal gap requirements over large strokes and because of the increasingly cost of stronger magnets and mechanical issues associated with the high moving mass .

An alternative is to convert the acoustic power into rotation first subsequently driving an off the shelf generator. Previous work on the small bi-directional impulse turbines for converting acoustic power  into electricity was very promising. Therefore, as a next step in this development, a full scale bi-directional turbine is build, and tested in the existing 100 kW TAP. The turbine is based on a classic design for an OWC plant and scaled down (69%) to fit into the open space in front of the cold hex of one of the engine stages of the TAP. The turbine rotor and guide vanes are manufactured by 3D rapid prototyping and build together with an off the shelf car generator (28V, 80A). The assembled unit is depicted below.

The generator is provided with a rotational speed sensor and loaded with a fixed dc-load of 0.5Ω.  The rotor of the ac generator is disconnected from the rectifier bridge and connected to adjustable power supply to set the field current and with that the shaft torque (T).  Together with the rotational speed (ω) the mechanical power (=ω.T) delivered to the shaft (by the turbine rotor) can be measured.

For the test we have only one turbine available and this one is mounted in the second engine stage (#2) of the TAP in front of the cold hex as is depicted below.

The test is performed with compressed air at 0.8 MPa as working medium. Actually the TAP is designed to operate with helium and therefore the maximum drive ratio during the test was limited to about 1%. Nevertheless the acoustic power is sufficient to test the turbine under realistic condition like an oscillating flow at elevated mean pressure (higher gas density). The measured data as a function of the field current (shaft torque) is shown below.

The highest measured mechanical output power (Pm) is 478W at a rotational speed of 377 rpm. At this point the actual drive ratio (pa/P0) was 1.14%. The acoustic power dissipated in the turbine is 629W yielding a conversion efficiency of 76%.

This result is in line with the data that in general turbine efficiency goes up with size and gas or fluid density. To illustrate this, the small bi-directional turbines (80 mmØ) tested previously at atmospheric pressure reach 40%, large (OWC) turbines (> 800 mmØ) operating at atmospheric pressure could reach 60-70% while water turbines could reach >90% due to the high fluid density. The present turbine is tested at elevated mean pressure (0.8 MPa) so its conversion efficiency should be somewhere in between which indeed is the case.

The conversion efficiency of the current bi-directional impulse turbine is somewhat less than the (theoretical) efficiency of linear alternators. Some more experiments and tests are needed to optimize the performance of this basic turbine under acoustic single frequency conditions. The experiment however proves that bi-directional impulse turbines has the potential for successfully converting acoustic power into rotation and from there into electricity at unlimited power scale.

An other, even more important, aspect is the economic efficiency. The turbine operates at ambient temperature, runs at relatively low speed and has no critical tolerances. Therefore the rotor and guide vanes can be made from plastic (e.g. injection molding).  The generator is an off the shelf type. As compared to the initial cost of the linear alternators in the TAP project this concept reduce the cost per kW electric output by one order of magnitude down to less than 300€/kW. This all brings the TAP back on stage for waste heat recovery on industrial scale.

Thermoacoustic electricity generation is still under development. The emphasis hereto is on  increasing power level and efficiency and in reducing the cost per Watt electric output power.

Early May, Aster together with the Nottingham Score team managed to get the demo2.2 Score thermoacoustic engine to produce 36 Watts electricity continuously, and 45 Watts peak. The working gas was air at an absolute pressure of nearly 200 kPa. A high power speaker was used as alternator. Its acoustic to electricity conversion efficiency is about 50% which means that the engine delivers 70-90 Watt net acoustic output power. Increasing the mean pressure up to 250 kPa and/or using a more efficient alternator will suffice to achieve the 50 Watt project target.

This result was obtained by modifying the acoustic circuitry toward a serial connection of two individual (acoustic) matched engine stages. This yield both a higher power density and a reduction of feedback tube length. The test set-up is shown below.

The current (messy) shape of the tubing is for measuring purposes and for easy access. In a final version the tubing can be folded to fit into, or be part of, the wall of the stove.  Note the relatively short feedback tube length which allows for a higher frequency and proportionally higher electric output of the alternator.

At an input temperature of 99°C and heat rejection at 20°C this corresponds with 38% efficiency relative to the Carnotfactor. This is an encouraging result because it proves that the thermoacoustic process can be scaled from lab size up to power levels in real applications.

For this measurement the TAP did not run at the proposed configuration of a 4-stage engine driving four loads (alternators). This is because on the moment no correct linear alternators are available. Therefore, to be able to judge the performance of the thermoacoustic part of the TAP, one of the four engine stages is used as an artificial acoustic load. This is done by disconnecting the high temperature hex of this stage from the heat source.  Net acoustic output power of the remaining three engine stages in that case is the  difference between the acoustic power at the in- and output of the disconnected stage.

The measured values for a 3-stage engine at these low power levels (relative to the design values) are found to agree well with the simulations. The 1.64 kW output power is reached with helium at a mean pressure of 750 kPa and at only 1.7% drive ratio. At this pressure amplitude the acoustic power in the feedback loop is nearly 6 kW.  Simulation for the initial 4-stage engine running at a drive ratio of 5% shows that for an input temperature of 140 °C the acoustic output power will be about 11 kW.

An important parameter to judge the performance of an thermoacoustic engine is the relation between applied temperature difference and acoustic power in the feedback loop.  The “steepness” of the curve versus the temperature difference across the regenerator is a measure for the internal viscous losses in the acoustic feedback or resonance circuit, the heat exchangers and regenerator. Plotting acoustic power versus the input temperature difference also includes the effect of temperature drop across the high and low temperature heat exchangers. For the 4-stage engine the measured data for both air and helium as working medium is plotted below.

The plot shows that the onset temperature depends on the mean pressure and that for helium at 750 kPa oscillation start at 40°C temperature difference. For air the onset temperature is even below 30°C but the less steep curve indicate higher viscous losses with respect to the acoustic loop power (which for air at the same pressure amplitude is lower than for helium due to the higher value of ρ.c). Note that these figure holds for the current implementation only and that, for example,  inserting alternators or changing heat exchangers or regenerator material or structure will modify this figure.

In the current TAP the first energy conversion step, from heat into acoustic power, is demonstrated to be feasible at relevant thermal input powers according to the project plan.  The second energy conversion step, of acoustic power into electricity, however fails to convert the acoustic output power available into the planned 10 kW level electric output power.

For the travelling wave multi-stage configuration the acoustic impedance at the pistons of the linear alternators needs to be real in order to extract maximum power at minimum pressure amplitude. A real acoustic impedance means that the mechanical resonance of the alternators should be close to the acoustic oscillation frequency set by the feedback loop length.

Initially the TAP was designed for running at 70-80 Hz.  In the end however, the high moving mass (magnet + springs) limits the mechanical resonance frequency of linear the alternators to no more than 40 Hz. This mismatch dramatically reduce the load to the engine and from that the amount of acoustic power that could be extracted. This issue and some other necessary re-engineering found from the construction of the current set-up will delay commercializing the TAP for the moment.  Next months focus will be on exploring and testing alternatives for converting acoustic power into electricity which can be  scaled up in power as well.

By the end of September the installation of the TAP was completed. The flue gas heat exchanger is inserted into the STEG exhaust ( cross-section 3 x 3 m) and  connected by an isolated water/steam circuit down to the intermediate heat exchanger located near the TAP.

From the intermediate heat exchanger a thermal oil circuit is used to transfer heat to the high temperature heat exchangers in the engines stages of the TAP.  The use of an intermediate hex for separation of both the thermal oil and water circuits was required here for concession reasons. In future systems this intermediate heat exchanger and pump has to be avoided because of the additional temperature drop which is harmful at low operating temperatures.

The low temperature heat exchangers in the TAP are connected to an existing  cold water storage tank in a stand about 80 meters away. All fluid circuits are equipped with pumps,   flow sensors and thermocouples for measuring thermal powers. Piezo resistive pressure sensors on the TAP are used for measuring mean pressure, pressure amplitude frequency, phase and acoustic power.

Testing the periphery was done by running the TAP, without alternators and with pressurized air (up to 750 kPa) to generate maximum heat flow trough all heat exchangers. At maximum heat flow the temperatures at the various sections are measured.

Conclusion from this test was that the effective temperature for the TAP to run at is limited mainly, but not only, by the heat transfer in the thermal oil circuit. Also the air sided temperature drop in the flue gas heat exchanger is higher than expected. As a result the effective input temperature of the TAP is not more than 100°C  which means that about 60 °C is lost in the periphery.

This observation stresses again the conclusion that at low and medium temperatures the design and performance of external circuits, to supply and reject heat to the TAP, are as important as the performance of the TAP itself.

After a lot of preparative work by the consortium, finally  the construction of the thermoacoustic  part of the TAP was completed by last week. The high- and low temperature heat exchangers and regenerators are assembled to single units, provided with thermocouples, mounted in the pressure vessel and connected to the high and low temperature fluid circuits. Also the data-acquisition system was installed for real time measuring temperatures and acoustic power.  The complete test setup of the TAP is depicted below.

At the construction location no 100 kW (waste) heat is available therefore a controllable heat source (9 kW electric oil heater) is used simulating the flue gas heat exchanger. This temporarily heat source allows for measuring static heat loss, onset temperature  and the increase of acoustic loop power with   temperature difference in the low power range.

Based on well known scaling rules the power levels in the TAP will be one order of magnitude less but the system will be thermoacoustically similar to the final system (using helium at 600 kPa) when filled with air at 240 kPa. This “similitude” allows for judging the performance of the TAP and for validating the simulations.

As a first result, the onset temperature difference  of the TAP was measured to be 29 K between the high temperature (49°C) thermal oil and the low temperature (20°C) water circuit which agrees well with the simulated values.

This is an encouraging result  because a low onset temperature is a key parameter and absolute requisite for successful operation of waste heat recovery systems. So the first test is passed and more test will follow.