Embedded Files
Sunfolding supplied single-axis trackers for the utility-scale solar energy industry.
I led the design of locking pneumatic rotational actuator, which is the core technology of the tracker system. The locking actuator enabled a step-change improvement to the cost of structural resilience.
Like all single-axis trackers, the T29 has a large surface area of solar panels mounted to a light-weight, low-cost steel structure. As such, tracker structural designs are often governed by wind pressure loads. An upcoming revision to the most widely cited structural design standard, ASCE 7, requires accounting for dynamic wind loads in addition to static wind loads. These dynamic wind loads can be caused when "exterior" trackers on the edge of an array shed vortices towards "interior" trackers within the array. If the vortex shedding frequency is close to an interior tracker's natural frequency of vibration, resonance can occur which "dynamically amplifies" the wind pressure loads - hence dynamic wind loads.
Exterior trackers are most likely to shed vortices corresponding to a Strouhal number of ~0.15 [Strobel- Banks 2014]
St = ( f_vortex[Hz] * L_panel[m] ) / ( U_wind[m/s] ) = 0.15
The resonance phenomenon is most powerful for the fundamental vibration mode of a structure (lowest frequency, most easily excited), which for a tracker is often a twisting mode about the longitudinal rotation axis. The frequency of this twisting mode is proportional to the following expression:
f_twist ~ √((tracker torsional stiffness)/(tracker rotational moment of inertia))
We were interested to find the lowest-cost tracker structure designed for dynamic wind loads. One common industry strategy is to "stow" the tracker at a steep angle when wind speeds are high, which is a stable orientation that does not shed vortices. But at steep angles, static wind loads are enormous, and driven-pile foundations must be very strong and expensive. Instead, we chose to minimize cost by using a flat stow position to reduce static loads, and by increasing torsional stiffness to reduce both vortex shedding on exterior trackers and resonance on interior trackers.
Tracker with a non-locking pneumatic actuator driven by bladders
Trackers supported by air bladders alone have a modest rotational stiffness limited by the bladder's pressure capacity (see picture at left). This is perfectly adequate for static wind loads. Designing a tracker to withstand dynamic winds loads at high wind speeds is also possible, but would require both a relatively high density of posts to increase stiffness and the addition of many fluid dampers to dissipate vibration energy. This approach leads to relatively high material and installation costs.
Another way to increase torsional stiffness is to use relatively fewer posts and actuators, each with a much higher rotational stiffness. We estimated that this flat-stow, high-stiffness, sparser-post design would be the lowest cost approach to designing trackers for dynamic wind loads.
Proof-of-concept with gate latch
In 2021 Q1, I began to explore design changes that would drastically increase the actuator rotational stiffness and strength. Product requirements including leaving the air bladders unchanged, avoiding additional sensors, and using only pneumatic power to drive mechanisms. All components and systems were designed for zero maintenance throughout a 30+ year life in all weather.
One promising concept was the addition of a lock to hold the tracker in the flat stow position. The lock would be actively "primed" by an air pressure signal when wind speeds exceeded a risk threshold, passively engaged when the tracker passed through the locking position, and actively disengaged by reversing the air pressure signal when wind speeds dropped back below the risk threshold. This control scheme allows for reliable lock engagement, open-loop control, and was tolerant of relatively slow control signals when using small-diameter air lines.
I designed a proof-of-concept for this concept using a ranching gate latch mounted to the output (moving) side of the actuator, and a pin mounted to the fixed side of the actuator. (see video at left) The lock is "primed" by extending the pin, and disengaged by retracting the pin.
(Note that the automotive turbocharger wastegate actuator used here results in a normally-retracted pin, which is not a failsafe mechanism.)
In 2021 Q2, I designed a single-post test rig to simulate the rotational inertia and motion of a tracker. Then, I worked with the Reliability team on a test procedure that primed, engaged, and disengaged the lock thousands of times. This test provided rapid, valuable feedback:
Confirmed that a 2-sided passively-engaged lock is a promising concept to meet product requirements
Alerted us that lock engagement forces are higher than anticipated, and the prototype required reinforcement during the test
Raised concerns over any load-bearing surfaces that would slide without lubrication, due to impacts on lock reliability
Suggested that material hardness and yield strength will be critical design parameters for load-bearing components
Lock mechanism test rig
Latch face surface is indented and abraded after 10000 lock cycles
Rotational actuator with integrated gate latches
In parallel, I explored ways to integrate a fail-safe version of the gate latch mechanism into a rotational actuator assembly, in order to test behavior in a full-sized tracker.
The large air bladders and their air lines were a significant packaging constraint, and drove a design with two lock mechanisms - one on either end of the fixed portion of the actuator. The lock pin is mounted on a curved steel plate which moves with the output side of the assembly. The gate latches are actuated with a horizontal rod spanning the driven by another pneumatic diaphragm linear actuator, this time from the HVAC industry.
This mockup confirmed that a lock mechanism could be integrated into the rotational actuator assembly.
I used this prototype to explore behavior of a multi-post tracker while engaging and disengaging the locks. Locks engaged reliably during a simulated high-wind event where one or more of the tracker's fundamental modes approaches resonance. But, locks did not disengage reliably when the air bladders caused a small net torque, as sometimes happens due to stiction or air pressure signals moving slowly through small-diameter tubing.
With an assembly layout that seemed plausible and results from accelerated life testing in hand, I designed a second prototype with improved lock mechanism reliability and wear characteristics.
Most load-bearing interfaces that had sliding contact in the gate latch were replaced with rotating contacts and bushings. Other load-bearing interfaces that would slide unavoidably were designed with high-hardness steels (HRC 20+) for minimal wear.
Lock mechanism exploded view, showing rotational interfaces
Demonstration of lock mechanism. Bladders are unpressurized to exaggerate effect.
Bushings show some wear after 1000s of cycles
In 2021 Q3, we repeated the same lock reliability test on the rotating-catch prototype and found that it performed very well. Some bushing wear was noticeable, likely due to the cantilevered rotating catches applying a moment to the bushings, but the wear did not have any negative impacts to lock reliability.
So, we moved onto static load testing. The Reliability team applied maximal wind loads to the actuator, represented by a force-moment couple (normal forces of 1,000s lbf, moments of 10,000s in-lbf). Consistently, the moment reaction loads on the cantilevered lock pin would cause the curved carrier plate to deflect out of plane enough to pull the pin out of the lock mechanism. The actuator would suddenly become unlocked, and rotate 10s of degrees.
I explored eliminating this risk by increasing stiffness of the pin carrier plate, but it proved too costly. So the twin-lock, cantilevered-pin concept was dead.
Locking actuator design with midplane lock load path
I made a final design iteration to the lock mechanism in 2021 Q4.
The lock mechanism was moved to the midplane of the actuator assembly, to eliminate out of plane bending and for much a more direct load path from input forces (wind pressure acting on solar panels) to reaction forces (contact between lock components).
Then, I designed the components for production using die casting, forging, and stamping. I used Solidworks FEA to check strength requirements while ensuring manufacturability and minimizing cost.
Stamped lock mechanism components
We spent 2022 Q1 evaluating risks using DFMEA as a team, and I made design refinements through 2022 Q2 according to our findings.
Alongside manufacturing partners, I refined each component's design for manufacture using the selected process (die casting, powdered metal, stamping). Component tooling was completed in 2022 Q3, and assembly tooling in 2022 Q4.
I assembled five prototype actuators at the end of the year.
(Pictures of assembled actuators have been redacted due to privileged IP)
Powdered metal lock mechanism components
Dies for high-pressure casting
Unfortunately, this development program was terminated before full production ramp-up, after company leadership changed priorities for the Engineering team.
A subset of the actuator development team was successful in using the processes, knowledge, and tools built in this program to execute other design programs.
Google Sites
Report abuse