From pv magazine Spain
We begin with a “real world” case study: At a 70 MW solar plant in Spain, 20 to 30 modules are being blown off of the trackers every few weeks. The plant is located in “wind zone C” – one of the windiest in Spain’s classification system, as we’ll find out shortly.
The tracker manufacturer claims that the incident is due to “extraordinary phenomena,” despite the fact that the anemometer registers values below the regulations, which is 29 meters per second for that area. This is the maximum pressure that the module anchors are capable of withstanding due to the wind.
The maximum pressure depends on several factors: the dimensions of the module, the length of the rows or rail panels, and whether staples or screws are used. The maximum pressure that the manufacturer accepts varies based on the design of the system, but is usually 2,400, 2,100, 1,800 pascals, or less.
The engineering, procurement and construction provider that purchased the trackers has blamed the anchoring system. The module manufacturer approved the design, but the tracker is not specifically named in the installation manual. The EPC contractor said that only a few modules have been blown away, but the reality is that the entire plant is theoretically exposed to potential wind damage.
“The problem is not only that some panels are blown away, but that those that have not been blown away (yet) are suffering mechanical fatigue in the anchor joints, weakening them and increasing the probability that they will start to be blown away en masse in the future,” said Asier Ukar, senior consultant and managing director of PI Berlin S.L.
Ukar has analyzed the case and has agreed to share his findings with pv magazine. He said the worst thing is that this happened after the EPC warranty. In this project, a few months remain until the end of the guarantee period, and after unsuccessful attempts to reach an agreement, an arbitration period begins.
This is a case that will sound familiar to more than a few readers. As prevention is usually better than finding a cure, Ukar identifies the design factors to take into account in similar cases.
Static analysis
“You need to ask the manufacturer to guarantee the specific anchoring method used for a particular module and for the particular project loads,” said Ukar.
For example, you can anchor a module measuring 2,384 mm x 1,096 mm (210 mm wafers) with four screws to a 400 mm strap, assuming a maximum pressure of 2,400 Pa. But that's not the same as anchoring a module measuring 2,256 mm x 1,133 mm (182 mm wafers) stapled to a 1,400 mm strap, assuming the same pressure.
It is common to find that the “pressure-anchorage-strap” triangle is not correctly reflected in the module installation manual. Without a clear certification, the responsibilities begin to be diluted.
“It is almost never clearly reflected because it doesn't matter. If the calculation is very transparent, the owner can see that the design is weak and demand reinforcement, which implies an increase in price and then the tracker manufacturer loses competitiveness”, said Ukar.
The Iberian Peninsula is divided, according to the Eurocode, into three wind zones: A, B and C. Zone C covers the windiest areas, such as Tarifa or Jerez. In this area, the Technical Building Code (CTE) dictates that a tracker must resist winds of 29 meters per second (104 km per hour), registered at 10 meters from the ground for 10 minutes. Since both the tracker and the anemometer are always located at lower levels, it is necessary to correct said speed, generally at the level of the latter, to make an “apples to apples” comparison.
Hellmann's exponential law is used to make this correction. Depending on the type of terrain, one coefficient or another is used, distinguishing between “flat places with ice or grass,” “rural areas,” or “very rough terrain or cities.” The coefficient “α” will therefore vary between 0.08 and 0.25, thus the design speed of the tracker being higher or lower depending on the type of terrain chosen. If a tracker manufacturer assumes rougher terrain than it really is, it will work with a lower design speed and therefore any speed higher than this value could be classified as “extraordinary.” The correct classification of the terrain must be done by the tracker manufacturer, and it is a key issue.
This observation also extends to the calculation of the dynamic pressure through the use of an appropriate exposure coefficient. In other words, the pressure that the structure must withstand will depend on the force exerted by the wind on the surface of the modules. And this force, again, depends on the type of terrain, and specifically on its degree of roughness, which is also clearly tabulated by regulations and divided into five classes from land ranging from “by the sea” to land “in business centers of large cities.” Analogously to the previous case, choosing a degree of roughness greater than the real one favors the calculations of the tracker manufacturer.
Another important issue involves the plastic limit of the anchors – that is, the maximum pressure they can resist before deforming. For this, finite element calculations are used, which constitute a fundamental part of all static calculations. In these calculations, it must be verified that the limit of materials is not exceeded by the forces resulting from the static calculation. This does not happen automatically. It is not uncommon to observe finite element calculations that show joint areas between the strap and the screw or staple in dark orange or red, clearly pointing to a risk of fatigue and subsequent breakage in operating conditions.
Finally, before moving on to the dynamic stress, it is essential to understand that some tracker manufacturers perform the structural calculation in the operating position (not stowed) with temporary values of the speed of only 3 seconds when really, the time necessary for the tracker to move into its stow position is much longer, often around 30 seconds. This entails a clear structural risk.
Dynamic analysis
Dynamic calculations are “another world,” according to Ukar. Dynamic calculation does not have as many guidelines or pre-established rules, as is the case with static calculation.
“Here we are talking about dynamic phenomena such as the famous torsional gallop or flutter, which are more difficult to predict and depends on numerous factors,” he said.
In order to adequately model the dynamic behavior of a tracker, additional testing instruments and parameters are required. One of these resources are scale models tested in a wind tunnel, which serve to obtain information on the dynamic behavior of the structure. It is an open secret that no tracker manufacturer carries out a wind tunnel test for every project due to time and cost.
Consequently, the results of previous studies are usually “reused” for future projects, leading to a consequent lack of “customization.” This can be seen when the tracker manufacturer provides, as part of the dynamic calculation dossier, studies up to three years old, often done on previous models and with piling geometry, overhangs and module dimensions that differ from those that apply to the project in question.
With regard to torsional stiffness, the premises assumed by the manufacturer to reduce thickness in the “torque tube” must be analyzed. If everything is so clear, why do cases of wind damage continue to crop up?
Although tracker manufacturers have more than enough knowledge to design robust trackers, price pressures force them to take risks in order to be competitive. This sometimes leads to reckless designs based.
“The sector must be willing to absorb an increase in the cost of trackers of … [up to] €0.03 (per watt), depending on the geographical location and the type of module used, if it is really intended to have systems with structural guarantees at 30 years,” said Ukar.
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: editors@pv-magazine.com.
Thanks for explaining.
What’s the lift force on a 30-50m/s ground level wind speeds and average weight of modules on said power plant on m² base? Main wind directions statistics available?
(dynamic pressure ~0.5-1.5kPa, wind load <50-150daN/m² and depending on angle of inflow)
Maybe it's about shaping panel profiles (adding shaped containers for collecting condensing water beneath?) for offering countering forces to uplift winds, too?
Very good article! I would like to see more of these if possible, nice work