Proceedings of KGCOE-MD2004: Multi-Disciplinary Engineering Design Conference Page 3

MD2004-04013

MICRO TURBINE DEVELOPMENT

Simien Lin
Project Manager / Charles Daze
Lead Engineer
Matthew Davis
ME / Thomas Cague
ME / Margaret LaRochelle
ME
Ream Kidane
EE / Lucas Lessa
ISE / Donald Slate
ISE
Dr. Jeffrey Kozak
Advisor

Proceedings of KGCOE-MD2004: Multi-Disciplinary Engineering Design Conference Page 3

Abstract

The technical paper presented will summarize the design considerations, analytical and experimental progress completed by Senior Design Team 04013 in regards to the fabrication and testing of a miniature turbine currently in development at the Rochester Institute of Technology (RIT). The main goal of the project is to design, fabricate, and test a housed miniature turbine capable of 5 watts of electrical power generation, while massing under 40 grams. The power output by the miniature turbine system will provide power for the electrical components of a micro air vehicle (MAV). Existing batteries are nearly as efficient as possible, meaning higher power to weight ratios are more difficult to develop. While the use of miniature turbines as a power source is relatively new, room for improvement in regard to efficiency is quite large.

Upon fabrication, the turbine will operate within a speed range of 40,000 to 120,000 rpm and develop a minimum of 5 watts of power, enough to power the controls of a MAV. Theoretical analysis of the system shows that operating speeds may reach 140,000 rpm, while producing approximately 75 watts of electrical power. Testing is currently in progress to minimize mechanical losses and maximize efficiency of miniature turbines.

introduction

Miniature turbines came as a result of the search for a lighter power source for Micro Air Vehicles (MAVs). MAVs are air vehicles with a maximum linear dimension of six inches used for surveillance and scouting. Due to size constraints, it is imperative that MAVs be lightweight. The main power source for MAVs is currently batteries, which may account for more than 50% of the total weight. By improving the power density of batteries, MAVs could be made even smaller, carry more instrumentation, or have greater endurance. Due to the chemical makeup of batteries, it is believed that battery efficiency is at its peak. Advancements in battery technology lead to minimal power density gains. Many institutions have begun researching micro turbine technology as a feasible replacement to batteries.

Research in micro turbines at the Rochester Institute of Technology (RIT) has been ongoing for more than a year. To date, a proof of concept and one original turbine design study has been conducted. The proof of concept research was able to produce 18 watts using a commercially available dental turbine. The original turbine design scaled down this design, but produced less than one watt. The purpose of this project is to continue this research and bring the micro turbine design closer to MAV applications.

Nomenclature

CG - Gradient Factor

CL - Load Factor

CS - Surface Factor

J - Polar moment of inertia

P - Power

Q - Volumetric flow rate

R - Overall turbine radius

R - Resistance

Sn - Endurance Limit

Sn’ - R.R. Moore Endurance Limit

SUS - Ultimate Torsional Shear Strength

T - Torque

V - Voltage

c - Radial distance from center

d - Diameter

n - Number of blades

r - Pitch radius

vjet - Turbine impinging jet velocity

b - Efficiency correction factor

r - Density

t - Shear stress

w - Angular velocity

FINAL DESIGN

Housing

The housing, machined from MDS filled cast nylon, is of cylindrical design comprised of two pieces; a cylindrical shell consisting of two flow paths, turbine and bearing seats, outlet holes, an inlet, and a cap that holds a vegetable fiber gasket along with the remaining bearing. Nylon screws run through both the cap and housing to press the two together, and function as alignment for the motor mount plate, which is also made of MDS filled cast nylon. Standoffs provide compression to the plug and act as spacers for the motor plate. The motor mount plate is circular with a hole in the center for the motor shaft. Surrounding the center hole are three set-screw holes to fix the motor to the plate.

Upon completion of machining for the housing, burrs were removed from the housing and the cap using fine grit sand paper. When the flow channels were cleared, the shaft/bearing/turbine assembly was fitted within the housing. The tolerances of the bearing seats were not held, leaving a slip fit between the bearing and seat. Initially, adhesive was considered to fix the bearings in the seat, but was decided against due to issues concerning seizing of the bearings. After being inserted into the housing, the cap was determined to provide enough constraints to properly align the shaft and hold the bearings within the bearing seats. Although the cap provided proper alignment, leaks occurred between the outer rim of the cap and the cap bore in the housing. Putty was initially placed around this area to seal leaks. While the putty did improve sealing, some air still escaped. Teflon tape was adhered to the outer circumference of the cap to act as a seal. The Teflon tape effectively lowered the flow rate, and enabled the turbine to operate at higher pressures.

In the preliminary design it was estimated that the housing would mass around 44 grams, four grams above the design parameter. Plans to remove excess material were discarded when the housing was found to mass 36 grams. Measurements were not made with the motor and motor plate included, as they were neglected in the scope of the project.

Shaft

The motor selection was the predominate factor in choosing our shaft, which has an output shaft diameter of .059 inches. To keep components as standardized as possible, a shaft with .059 inches was chosen. In order to validate the shaft selection, stress and fatigue analyses were conducted. The calculations required power output values, which were taken from turbine performance plots. Formulas corresponding to the analysis are shown below in equations (1-5)13.

To determine the fatigue life of the shaft, the ultimate yield strength of stainless steel was needed. For steels, this is 0.8 times the ultimate strength. The infinite life strength was found using equation (xxx)13.

CL, CG, and CS are all correction factors based on geometry of the specimen. Further analysis yielded a maximum torque of less than 1.5 ksi, nearly fifty times less than the ultimate yield strength. The infinite life strength of the shaft was determined to be 34 ksi; 23 times larger than the maximum torque found in the shaft.

Bearings and Coupling

The bearings chosen were ABEC 3, pre-lubricated, unflanged bearings from the Alpine Bearing Company. The desired operating speed of the shaft was chosen to be 100,000 rpm, which is within the rated 120,000 rpm of the bearings.

Alignment, a major issue with last year’s design, was a prime focus of the current design project. A pre-manufactured coupling would be best suited, as a means of standardizing the design, and making replacement of failed parts simple. Two companies were found which produce micro-couplings; but were ruled out due to bore size and rated speed limitations. A Belgian research group working with micro turbines used heat shrink tubing as the coupling for their system, with estimated losses near 1%16. Rigid heat shrink tubing was chosen to couple the two shafts on the basis of low cost, light weight, and ease of assembly and disassembly.

Turbine

Pelton type impulse turbine technology has not been under further development for some time. Industry has changed to axial impulse turbines or more advanced reactive turbines. Even large hydro power plants have changed over to a cross flow turbine, leaving few sources of Pelton type turbine design. The available sources use water as their working fluid, which causes numerous problems. The fluid property change also affects all design parameters, making any established relations invalid. The only design relations that are still true are the general equations for estimated power outputs. A design basis had to be formed for the turbine and then validated with these power output equations.

Turbo machinery power density scales with the square of the peripheral speeds6. With this compact design, it is optimal to have the power density as large as possible. Therefore, the tip speeds of the blades should be approaching sonic velocity. For a selected diameter of 0.312 in, this would mean that the optimal angular velocity would be approximately 400,000 rpm. Such speeds would require a tremendous amount of alignment and bearing focused work. The literature review and previous studies at RIT determined 100,000 rpm to be the largest practical velocity for the turbine.

To obtain the largest torque from the jet, the pitch diameter should be as large as possible. The pitch diameter should also be approximately in the middle of the blade. The best performing blades are at a balance between impact area and height10. It is also ideal for the jet to always be impinging a blade at the exact tangential location on the pitch diameter. This logic relates to the number of blades. As soon as one blade is being pushed passed this tangential location, another blade must be moving into the jet. If there are too many blades, a blade will begin to impinge the jet before the previous blade reaches the tangential location. Turbine blade design is a balance between blade area, height, and number.

Most modern Pelton type turbine designs have a ridge that runs down the center of each blade radially. A cup is formed from this ridge to the outside that runs the entire length of the blade. On the micro turbine scale, a design similar to this could be machined on a 4 axis CNC. However, the cost for each turbine would be multiple thousands of dollars. This is outside our budget, and impractical for reusable and replaceable application on MAVs. To reduce machining costs, the design must be a 2-D design that can be made by wire Electro-Discharge Machining (EDM). This means that the turbine will have to be curved from the bottom of the blade to the top.

Using this basis the turbine was designed. The design rotational speed is 100,000 rpm, and the outside diameter is 0.312 in. The blade height is 1/3 of the radius, 0.052 in. The pitch diameter is the center of the blades at 0.27 in. Equation 6 was used to determine the number of blades using these dimensions10.

(6)

Solving the equation resulted in a value for n ranging from 6.24 to 8.43. For balance, n must be an even number, therefore eight blades was selected as the best design.

The rest of the turbine was designed using curve fits and tangent curves. The blade thickness was set at 5 degrees, which was determined to be the best in balance between curve effectiveness and thin blade design. The blade on the pitch diameter must be perpendicular to the jet velocity at the top of the turbine. To blend the tip of the blade and the flat pitch diameter, a series of curve fits were utilized. The back of the blade had to remain slim. The low profile design was used to keep the blade thickness down and to keep the back of the blade from hitting the jet flow before the front. The final turbine design is shown in figure 2.

To validate the design and determine design mass flow rates, estimated power output was calculated10. To begin, the expected torque and power comes from equations 7 and 8.

(7)

(8)

Only using a small correction factor, these equations are finding the maximum power of the turbine. The actual output is dependent on many more variables. Our actual output will be affected by bearing, alignment, friction, flow, and generator losses. To be a robust design, the results of these equations must be much larger than our design specifications. The results are plotted in table 1, illustrating the design’s robustness. The design goal is a turbine that operates around 40-50 psig at 100,000 rpm. At this point, the maximum potential output is approximately 30 watts much more than our desired 5 watts, which will allow for our design to overcome the other losses in the system. Even if this isn’t sufficient, the turbine and housing will be designed to operate up to 100 psig, giving us an estimated turbine output of 50 watts.

Testing

Test Set-up

The test stand developed is shown in figure 4. Due to the large mass flows required, the main source of air is the shop air provided to the RIT Wind Tunnel. The shop air maximum deliverable pressure varies between 70 and 90 psig. High pressure air tanks will be used as a secondary air source to provide consistent pressures above 70 psig. The incoming air enters a stagnation chamber or plenum to dampen any perturbations from the shop air.

Budget limitations forced the team to purchase a rotameter type flow meter. Therefore, the flow has to be turned vertically for the meter to operate correctly. The turbine/generator system is connected by a flexible hose that allows the system to be removed quickly and easily with limited wear to the nylon parts. The hose also allows the housing to sit level for optimum turbine performance.

A three phase motor was chosen to be used as the generator. This motor was connected to a delta circuit, so that it could be loaded and power outputs could be read. The delta circuit utilized resistor boxes, to include variable resistances. This allowed flexibility in our test set-up and turbine operating conditions. The voltage across one resistor box is measured using a volt meter. Knowing this and the resistance applied, the power output can be calculated using equation 9.