The most slippery assumption in designing PV farm layouts at the early stages
The nature of designing a layout presupposes many feedback loops and iterations, which shows a kind of evolution of the initial design. We don't have all the answers when we do the initial design, so we make assumptions about the future.
Everybody makes mistakes, so there is nothing terrible about wrong assumptions. Still, there is something sinister when the construction cost becomes 2X, where X is the initial cost we calculated because of our assumptions.
Let's speculate a bit about critical features of the PV farm layout and what makes this design unique and different from, say, residential building design.
PV farm layout features
The equipment is connected in hierarchical circuits
Think about a solar PV farm as a tree (connected acyclic undirected graph):
Modules are grouped into strings on the ground mounting systems (single-axis tracking, fixed-tilt, etc.)
Different electrical patterns are connecting strings to combiner boxes (for the 'Central Inverter' scheme)
Then the current is gathered into inverters, where the direct current becomes an active current.
After that, the current goes to transformers, where the low voltage system converts into a medium voltage.
Transformers are being grouped in medium voltage circuits with the help of sectionalising cabinets (when transformers cannot be simply connected one-to-one).
At the end of the day, all medium voltage circuits lead to the substation, which is the final destination for the current within the layout.
The high voltage system starts its way after substation, but it is already a bit different world.
The example above is a simplified one, and it covers only the 'Central Inverter' scheme. Nevertheless, it should give us a good understanding of the 'tree' concept for the PV farm models.
The layout contains many repetitive patterns
The number of strings connected to the combiner boxes is defined by amperage and some safety margins. Usually, those strings are hosted on trackers, and electrical wiring gathers the current from the strings on trackers to the combiner boxes. As a result of these complicated relations, we get a combiner box circuit - for example, 18 strings on 6 trackers connected to the combiner. Those circuits repeat themself across the layout.
MV circuits and blocks are repeatable circuits as well. Therefore, we can think of them as reusable components within a tree hierarchy.
The degree of reusability depends on site context and equipment variability. Thus in some ideal conditions, we might have no diversity at all. Which is a good thing from a predictability standpoint, isn't it?
PV farm layout design process
Typically, the PV farm layout lifecycle contains five main phases:
BRIEF At this phase, project owners define the goal, gather information about constraints and build up assumptions about certain things that cannot be known at the moment.
DESIGN 10% This is a tender phase, where engineers have BRIEF and many discussions as a basis to design the very initial version of the layout. On utility-scale projects, it usually lasts from 4 to 6 weeks. The main problem is to make the right decisions by reinforcing the correct assumptions because, technically, engineers don't have all the information that is needed.
DESIGN 30% Conceptual design phase where engineers have almost all the relevant information: geotechnical analysis, topography survey, equipment setup, etc. The output of this phase is a preliminary array layout which built with the most crucial considerations in mind: grading, wetlands, and equipment specification.
DESIGN 60% Developed design phase where engineers start delivering drawings for their disciplines: electrical engineering, structural engineering, civil engineering, etc. The output of this phase is a detailed array layout with electrical, energy performance, structural and civil considerations.
DESIGN 90% During the technical design phase, engineers create final construction drawings and submit them for review and approval. To design projects from 10% to 90% typically takes 10-12 weeks for utility-scale projects, and then it requires another 2-4 weeks for permissions.
I believe that the best concept to describe the design process is a spiral feedback loop:
Every phase of the PV farm layout lifecycle is a feedback loop epoch. Every epoch contains several feedback loops (read more about feedback loops here). The transition from the previous phase/epoch to the subsequent is a refinement, cultivation or evolution of the design (for every taste).
How PV farm layout features affect the design process at the early stages
Engineers have minimal information about PV farms at the DESIGN 10% phase. But it doesn't mean that process is frozen until the brief is fulfilled. All the gaps are covered with assumptions, and engineers try to get the maximum allotted time to find the best possible layout. As mentioned above, the layout consists of many hierarchical circuits and repetitive patterns. Also, we already know that engineers have a limited amount of time (we all do, but here we mean the time for design). So how do they optimize their work, trying not to lose quality?
The answer is pretty simple - they design thoroughly typical parts of the system and try reusing them to get the complete picture.
But the obvious problem with such an approach is that it creates an ideal world situation. For example, the jigsaw above contains 48 pieces. Even if we assume that every piece is a prototype of one solar farm block and, say, we defined 3 unique blocks and multipliers for the whole layout, we are still missing a considerable part of the entire picture.
After applying real context and constraints the PV farm layout might be much more complicated that it was assumed. For example, there are actually 15 different types of blocks in the layout-jigsaw we used a simplified model for the solar PV farm layout.
Making the wrong assumptions about the repeatability of a specific type of circuit is the most dangerous one when designing solar layouts. The nature of the PV farm layout multiplies the mistake, which can lead to substantial miscalculations. To avoid this, we need to consider nuances behind equipment: roads layout, electrical wiring concept, trenching, topography specifics, and construction sequencing. Well-educated high-quality guesses are our bread and butter. Amen.