Usually, it is desirable to install solar mounting systems on flat roofs without roof penetration or attachment. Due to the dead load from weight of modules and substructure being too low to prevent uplift or sliding failure due to wind actions, additional ballast is required to maintain static equilibrium. For oblique wind directions, vortices originate from the corners of flat roofs. These vortices which are also referred to as "delta wing vortices" cause the largest suctions on both, roof and solar racking, and reduce with increasing distance from the roof corner.

Wind tunnel studies conducted in the I.F.I. large boundary layer wind tunnel laboratory have shown that many parameters affect the wind loading on solar racking systems mounted on flat roofs. Typically, it is required to accurately scale models of solar panels including panel tilt angle, row-to-row spacing, aisles or gaps between rows, height above the roof, setback from roof edge, alignment of rows compared to the main axes of the building, deflector shapes, and where possible the geometry of the panel support structure. The models of generic buildings shall be large enough in plan area to capture the wind flow environment over different roof zones with at least six rows of panels per array. Therefore, the test matrix needs to include the range of building plan dimensions, roof height, and parapet height. Another important parameter is load sharing as ballast required to maintain static equilibrium of the solar roof mount system depends on the stiffness of connecting members.

According to ASCE 7-16 and SEAOC PV2-2017, peer-review of wind tunnel reports on solar ballasted roof mount systems is required. I.F.I. Institute for Industrial Aerodynamics is a City of LA Department of Building and Safety (LADBS) approved laboratory for wind tunnel testing of buildings and structures, Testing Agency License Number TA 24830.
 

It is tempting to calculate wind loads on flush-mounted solar PV systems simply by applying provisions for gable and hip roofs from codes such as EN 1991-1-4 or ASCE 7. However, simple approaches are often conservative. As wind tunnel studies conducted in the large I.F.I. boundary layer wind tunnel have shown, reductions in the range of 30% to 70% of code values are possible enabling ballasting instead of roof penetration. The reason is that in the wind tunnel the favorable effect of pressure equalization between upper and lower module sides is modeled, whereas code values simply represent external pressure coefficients.

Single-axis PV tracking systems and fixed-tilt ground mounted PV systems may be subjected to wind dynamic effects such as buffeting forces resulting from the wakes of upstream rows. Thus dynamic amplification of the mode shapes may occur below or at typical design wind speeds if the vortex shedding frequency matches the natural frequency of the structure. Structural damping of the mounting system acts to mitigate the structure’s response.

From the wind tunnel testing, pressure-time-series corresponding to peak structural load effects such as normal force, tracker moment, bending moment or cantilever moment are identified. Parameters affecting the wind loading are tilt angle, row-to-row spacing, ground clearance, chord length etc.. Rows and tables in the array interior generally benefit from sheltering effects.

With the results of the modal analysis (mode shapes, natural frequencies) and critical damping ratio, resonant dynamic amplification factors (DAFs) are calculated. From the combined effect of static and dynamic loads, the equivalent static wind load may be calculated, and e.g. fed into structural analysis software to calculate peak stresses, displacements or foundation loads.

Equivalent static wind loads have the advantage of being in a format that structural engineers are used to work with. Furthermore, equivalent static wind loads can be easily combined with other loads such as dead load, snow load etc. by applying load combinations outlined in building codes such as EN 1990 or ASCE 7.

In the past, flutter- or galloping-like instabilities have been observed on single-axis trackers relying on a highly flexible central torque tube driven from a single location. The aerodynamic instability is a combination between a static increase of the torsional angle and a sudden excitation of the first mode of vibration. This mode is a helical twisting which increases with distance from the torque motor. It is excited by vortices forming on, and then shedding from, the leading edge as it twists up and down. It is important to note, however, that the aerodynamic moment increases the angle of attack in a quasi-static manner without reaching the state of torsional divergence. The sudden release of torque as a vortex is shed triggers torsional motion that grows uncontrollably with wind speed once a critical wind speed is exceeded. The critical wind speed is defined as the limit state where the sum of structural and aerodynamic damping becomes zero. In general, the frequency of oscillation is lower than the frequency of the first mode of vibration, but vortex lock-in may occur at tilt angles above 10deg.

The instability is stiffness-driven at low tilt angles (0°-10°) and is referred to as torsional galloping. Large aeroelastic forces and rotations may occur which cannot be effectively mitigated by damping.

At mid tilt angles (20°-35°), the instability becomes damping-driven and is referred to as torsional flutter. Aeroelastic forces and rotations are still possible, but damping mitigation becomes more effective. Static loads are much higher than at low tilt angles.

At high tilt angles of 45deg or above, the highest static loads occur. Except for negative tilt angles at low damping, aerodynamic instability is not observed to occur. However, small aeroelastic rotations are still possible.

To suppress self-excited oscillations which may seriously damage the structure, stowing at tilt angles at or above 30deg is a possible solution. Torsional stability depends on wind speed, natural frequency, torsional stiffness and damping ratio. It is expected that the most significant effects due to self-excited forces will occur from winds within +30° normal to the array.

Another way to eliminate the torsional instability when stowing for survival would be to lock the torque tube position at several positions along the row. This could be realized by using a rotational stop (torsional safety valve, torsional relief system) or by undertaking other counteracting measures.

The order of installation of tracking system components should be in a way which ensures that torsional instability cannot occur prior to commissioning of a solar plant. With regard to this, it is highly recommended that dampers be installed ahead of modules. Once installation is completed, it is also important to monitor the health of the dampers.

The adoption of National Annexes to EN 1991-1-4 of the CEN member states (“National members” and “Affiliates”) has evolved remarkably during the past couple of years.

“National members” are regular members of the Comité Européen de Normalisation (CEN). Regular members are national standardization bodies of the EU and EFTA member states. Additionally, the general assembly may offer a regular membership to candidates for an EU or EFTA membership.

“Affiliates” may be national standardization bodies that reside in the “European neighborhood” and are regular or corresponding members of the International Organization of Standardization (ISO). Historically, the term “European neighborhood” is interpreted quite widely. Besides Eastern Europe and the Caucasus, it includes larger parts of the Mediterranean and of the Middle East. Originally, this status was supposed to be a probationary membership for EU- and EFTA-candidates, but today it is also offered to other interested countries.

Current regular CEN members are: Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Macedonia, Malta, the Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey and the United Kingdom.

“Affiliates“ within Europe are: Albania, Belarus, Bosnia and Herzegovina, Moldova, Montenegro, Serbia and Ukraine.

“Affiliates” outside Europe are: Armenia, Azerbaijan, Egypt, Georgia, Israel, Jordan, Lebanon, Libya, Morocco and Tunisia.

Currently, an advanced state of adoption of National Annexes among the regular CEN members is observed.

• The current 33 regular CEN members with very few exceptions have adopted the EN 1991-1-4 and have published National Annexes. Exceptions are Macedonia, Malta, and Turkey. For Malta the estimated date of publication for the National Annex is unknown. For Macedonia no information is available. In Turkey for wind design a code is used which shows apparent similarities to a former German wind loading code.
• Among the “Affiliates” only Belarus and Serbia have published National Annexes to EN 1991-1-4. Bosnia and Herzegovina as well as Montenegro are currently developing National Annexes, but have not yet decided when these will be published.
• Armenia has no national wind loading standard and does not consider adopting the Eurocode.
• Ukraine has not published a National Annex. The valid Ukrainian standard shows a lot of similarities to the Russian standard, even though both codes are not identical.
• Russia and Azerbaijan are still using the standard of the former USSR.
• Israel does have an independent wind loading standard.

Official wind loading standards are neither available for Albania and Moldova, nor for Egypt, Jordan, Lebanon, Libya, Morocco and Tunisia. However, many of these countries either use the Eurocode or ASCE 7 alongside recommendations for wind speeds or nationally defined wind speed maps.

Wind load related standards in the US are ASCE 7-10 and ASCE 7-16 while a former version of the US code, ASCE 7-05, is still used in some federal states.

The versions of ASCE 7 also briefly define the requirements for wind tunnel testing. A more thorough general description on how to conduct wind tunnel tests in boundary layer wind tunnels is given in ASCE 49-12. ASCE 49-12 also defines appropriate methods for post-processing of wind tunnel data while ASCE 7-16 has its own chapter for wind tunnel testing on solar ballasted roof mount systems.

Requirements for wind tunnel testing of solar roof mount systems are given in more detail in the SEAOC PV2-2017 report published by the Structural Engineers Association of California.

Besides EN 1991-1-4 and ASCE 7, other advanced wind loading codes may be found all around the world. Examples are AS/NZS 1170.2 in Australia and New Zealand, the AIJ Recommendations on Loads for Buildings and the Building Standard Law of Japan, the 2018 edition of the International Building Code IBC or ISO 4354:2009(R2014), the code of the International Organization for Standardization. The aforementioned codes differ in averaging periods for design wind speeds, pressure coefficients for various structures and wind dynamic effects.