The most efficient Utility-Scale Solar Farm for EPC Contractor

Recently, I have written a blog post about the most efficient building quests in the AEC industry. The main idea was about different points of view on what does the most efficient building mean (you can read a full version here). This time I would like to narrow the topic and dive deeper into details regarding Utility-Scale Solar Farms (PV farms, to be precise).

The best is the enemy of the good.

I would change this quote to The best is the enemy of the good enough.

I started this blog a few times, and I struggled to publish things because of their imperfections, although I promised to write three blogs every two months. It is an excellent skill to have a stopword for 'good enough' things. Thank you in advance for your comments and messages. It helps me a lot to improve the content I create.

Assumptions

The AEC industry is a tricky one. Before jumping into any reasonings, we need to build assumptions and stick to them. Otherwise, we stumble at once, and we cannot go far ahead.

1. The primary market is Utility

2. The primary service is EPC

3. The point of view is Contractor

EPC Contractors are responsible for engineering, procurement and construction. They step into a process at a very early stage, after the Developer/Client created a Request For Proposal (RFP), in other words, requirements specification. Usually, many contractors bid on the same RFP and provide their Quotes. Then Developer/Client chooses only one of them and finalises the contract. Two of the most common contract types are lump-sum contracts and time & material contracts. For simplicity, let's choose a lump sum contract (a single price for all the works is agreed upon before works begin). After the contract is signed off, EPC Contractor becomes a single point of responsibility for design and construction.

Where to start?

Regarding our previous assumptions (I know I overconstrained the task, sorry), we have EPC Contractor working on Utility-Scale PV Farms that just signed off a lump sum contract.

There are no more discussions about most long-term energy, lowest operating costs and CAPEX efficiency. The only constraints we have are the contract price and the safety requirements. The contract includes hard, soft, overheads, and profit margins. The biggest issue here is that when EPC Contractors agree on the price with the Developer/Client, they have a minimal amount of information, so it is pretty crucial to be able to deliver the farm within assumed spending and profit:

Contract Cost = Hard Cost + Soft Cost + Overheads + Profit Margin

I suggest we focus on a hard cost (the most significant part of spending) and go through all trade-offs EPC Contractors need to analyse when designing a PV Farm Layout.

Fake PV Farm creation

Let's find the place where we always wanted to build a solar farm.

I hope you understand that google maps don't define solar farm property boundaries in real life. The boundaries come from ALTA land title surveys. But we don't care; we create imagined fake PV Farm (kind of freedom 😁).

Again, it is cosplay, and we are an EPC contractor - to make this experiment close to real life, please, keep it in mind. And we just signed off the contract for the parcel of the land above (surely, we built some assumptions and did some calculations, so it is not an unfounded number) with a lump sum price of $50M. Thus we have some limits we cannot cross with the hard costs, and we have requirements specifications from the Client. These conditions make the definition of the most efficient Utility-Scale PV Farm relatively straightforward: to minimise the hard costs with respect to clients' requirements and safety.

Before we start having a lot of fun with the actual experiment, let's define what information we need to have about the site and what Client's requirements are?

Then we will break down hard costs into categories and speculate on what trade-offs EPC Contractor typically needs to go through?

What information do we need to have about the site?

Obviously, EPC Contractors need this info long before the contract is signed off. But in a real-life, you never have enough time to mine a piece of accurate information or analyse it properly. Moreover, the data flow is coming and constantly updating without giving any respite.

I would separate all the data we need about the site into two categories:

• Data defining immovable boundaries (you cannot cross those boundaries at any cost)

• Data defining movable boundaries (you can change them at cost)

In most cases, boundaries are closed polylines or area spots.

Let me provide a few examples of data defining immovable boundaries:

• ALTA survey (legal document) is a detailed drawing of the property that shows boundaries and their relations to a title.

• Zoning requirements include required height limitations, setbacks, landscaping, site visibility, enclosure and screening.

Below you can see some hand-drawn sketches for our fake PV farm site of what is usually pointed out in the ALTA survey.

For an actual 3000 acres project, the ALTA document is a 10 A1 pages survey that contains lots of notes, legal descriptions, certifications, adjoining ownership tables, keymap, legend and detailed plans with boundaries, roads, fences, flood zones (divided into categories like '1% annual chance flood', '0.3% annual chance flood', etc.), overhead electric lines, power poles, telephone lines, gas lines, fibre optic lines, wash lines and even saguaro cactuses positions 🌵.

Zoning requirements come from different levels of authority, and they vary regarding country. As we are building a fake PV farm in Arizona, we need to find an authority responsible for planning permissions - for our site is a Yuma County AZ. Then we can discover Yuma County Zoning Ordinance, where we will explore all requirements to meet to get planning/building permission. After that, we submit the application with the proposed site plan and narrative and get them approved (after several attempts 😉).

Getting a building permit is the Client's responsibility, and EPC Contractors don't need to submit the application, but they need to take into account the approved site plan and narrative. This is exact data defining immovable boundaries and conditions.

And again, under 'immovable boundaries' we mean boundaries we cannot cross at any cost. Here, we are not talking about all the money in the world 😈 or building permits obtained through threats of nuclear weapons ⚛️ (actual and not funny joke).

Let's talk about the data defining movable boundaries or any other data we need to understand our hard costs:

• Topography (elevations representing terrain model) or so-called digital elevation model (DEM)

• Geology data (geotechnical report and geotechnical evaluation)

• Hydrology data (elevation analysis, flood maps, flood depth, drainage summary)

• Georeferenced photos

You might have relevant questions: why does the data above define movable boundaries? or why those boundaries are movable?

We need to do a quick sketch about how site slope can affect necessary cut & fill to understand this. The drawing below shows how single-axis trackers (without undulated capabilities) manage site slope. As you see, we can provide the same amount of modules installed by focusing on different trade-offs. For example, in the first scheme, we decided not to move any dirt, but as a result of it, we got many non-unified and long piers. So we decided to be more respectful with pier types and steel tonnage in the second scheme while moving lots of dirt. Then if we apply the second scheme to an actual site with a specific topography, we will end up in a situation where we will need to move too much dirt (which becomes expensive). In that case, we might want not to install trackers in a very sloppy zone. Thus, terrain model conditions created boundaries where we don't wish to trackers installed. But those boundaries are movable because when we need to hit a specific energy target and don't have too much space, we can be pushed to use those unfavourable zones.

What the clients' requirements are?

A short answer - they are always different. Assumptions and schemes are our bread and circuses for this blog.

Typically, clients' requirements are a set of documents like Electrical Specification (substation requirements, low voltage requirements, medium voltage requirements, etc.), Electrical Equipment Specification (design life requirements, exterior fire rating requirements, operating temperature range, HVAC system requirements, etc.), Main Power Transformer Specification and many more. Each of those documents contains 20-30 pages of information. But let's choose the most important ones and aggregate them in 7 bullet points:

1. PV Farm Power = 120 MWdc, which means that PV Farm should be capable of generating 120 MWdc power at high noon on a sunny day

2. Design Life = 30 years, Exterior Fire Rating 2 hours

3. PV Module: First Solar FS-6460-P

4. Inverter SKID: SG3425/3600UD-MV

5. 30 ft road width (glass-to-glass)

6. Maximum Voltage Drop: 1.5% for low voltage wiring and 2% for medium voltage wiring

7. Mounting system: Nextracker Horizon, 60/66/72 mods tracker variants, trackers heigh preference = 2 (trackers length maximisation)

For the sake of clarification, some of the requirements above came initially with the Client's specifications, and some were made during pursuit phase conversations. I would like to emphasise that creating a clear list of requirements is a big deal, and it is a separate big chunk of work to negotiate and make all clear. I think this fact is something people with superficial knowledge about the industry underestimate in their judgments about 'how everything is inefficient'.

What trade-offs does an EPC Contractor typically needs to go through?

Before talking about trade-offs, let's do a hard cost breakdown. There are plenty of ways of doing a cost breakdown, and many classification systems exist for this purpose (to name a few of them, MasterFormat, UnitFormat, Uniclass, NRM1/2, SMM7). Some are system-based, others are material-based, but the idea is the same - to group costs in buckets, understand buckets' proportions and how any changes affect them. As we are talking about EPC contractors, I suggest we focus on the activities or scope of work they provide when building utility-scale PV farms:

1. Equipment (inverters, transformers, DC/AC combiners, disconnects, etc.)

2. Electrical wiring (low voltage wiring, medium voltage wiring, modules, sectionalising cabinets, fiberoptic cables, terminations, DC/AC trenching, etc.)

3. Civil (earthwork grading, roads, fencing, gates, etc.)

4. Mechanical (piles and trackers)

Different people are responsible for the pre-construction phase of the categories above:

• Solar performance modelling engineers are responsible for energy model and equipment selection and negotiations with the clClientlectrical engineers are responsible for all electrical wiring design.

• Civil engineers are accountable for cut & fill design, roads, fencing, gates, basins, etc.

• Structural engineers are accountable for piers.

Besides the people above, there are clients, pre-construction managers, consultancies and sub-contractors. The interactions of the all above end up with many different trade-offs. And when there is enough time, EPC Contractor can consider all necessary trade-offs, compare different farms options and choose the best configuration of a utility-scale PV farm (utopia 🤏). But there is never enough time. ⌛

It is time to go back to our fake PV farm site with created immovable boundaries:

On the screenshot above, you can see inclusion boundaries and exclusion boundaries. But in a real-life, boundaries are not inclusion and exclusion or black and white. For example, exclusion boundaries can carry different functions: do not put trackers, do not put equipment, do not do wiring above, do not do wiring below, do not do roads, do not trench, etc.

Let's configure our equipment and see what we can get out of this site:

As you see on the second screenshot above, we got 132,168.64 kW of DC power (which includes 39 inverters, 4053 trackers, 47877 strings and 287322 modules) for our fake PV farm site with immovable boundaries.

We are cosplaying EPC Contractor, and we just confirmed that we could hit our energy target for this particular site. And now, we can cut off excess power in the most costly zones.

Hold on a sec!

Before we go further, I want to bring up some framework philosophy here.

I genuinely believe that the right feedback loop for EPC Contractors should be:

1. Test if you can hit the energy target with the maximum available space (site boundaries minus immovable boundaries - boundaries you cannot cross at any cost)

2. Bite off the most costly zones out of site (create some movable boundaries) and see if you can still hit the energy target

3. Presumably, you will get different scenarios with combinations of movable boundaries that will provide the required energy target. Every scenario will take into account specific trade-offs. So play around with those scenarios and trade-offs until you get the cheapest option.

If #1 doesn't work, you shouldn't have bided on this project. 😶‍🌫️

I love frameworks because they are much easier than real life. 🤹‍♂️

There are many scenarios ('what ifs') with hard choices of trade-offs. Having the right tools that can help you evaluate one scenario vs. another with appropriate hard cost implications is the key to success in the construction.

Back to our fake PV farm and trade-offs around it

To make this blog too detailed, I propose to operate with four enlarged categories of costs (equipment, electrical, mechanical, civil) for each scenario that we will consider. All the numbers will be fake (from the bottom of my heart 💞), so don't take them seriously, please. All scenarios will be developed concerning the client requirements we pointed out above.

Initial option or starting point (excessive, only immovable boundaries)

PV Farm Power = 132.2 MWdc (12.2 MWdc more than required)

Inverters SKID number = 39

Trackers number = 4053

Modules number = 287322

Equipment Cost = $5.85M

Mechanical Cost = $10.14M

Electrical Cost = $13.65M

Civil Cost = $9.36M

Total Cost = $39.00M

Option #1: reducing earthworks

The site has several prominent hills. We will try to avoid them. At the same time, while doing that, we should always keep in mind that we cannot go below 120 MWdc.

Here is what we got:

PV Farm Power = 121.9 MWdc (close to our target)

Inverters SKID number = 36

Trackers number = 3759

Modules number = 265062

Equipment Cost = $6.35M

Mechanical Cost = $9.18M

Electrical Cost = $13.42M

Civil Cost = $6.35M

Total Cost = $35.30M

Option #2: reducing the cost of trenching pathways and roads

The layout above has many zigzag roads. That happened because the algorithm tried to save 2-tracker height across the whole site.

We can create another layout with a 3-tracker height within the same buildable area and check what we get:

PV Farm Power = 122.03 MWdc

Inverters SKID number = 34 (here we accidentally improved inverters number)

Trackers number = 3762

Modules number = 265290

Equipment Cost = $6.19M

Mechanical Cost = $9.83M

Electrical Cost = $14.56M

Civil Cost = $5.82M

Total Cost = $36.40M

As we see above, despite some savings from civil and equipment spending, we increased the electrical wiring cost, and hence the total cost.

Option #3: reducing mechanical cost

We used three trackers configurations (60/66/72 mods) in previous scenarios. We can test the layout with only 72 mods trackers, which should reduce the number of external and motor piles and positively affect mechanical cost.

As you see, with our initial attempt (top left screenshot), we cut off some energy and didn't hit the energy target. So we decided to use different trackers positioning. In some cases, it can bring more power, especially for sites like ours. Thus we got:

Power = 111.28 - 14.7 - 26.76 + 50.3 = 120.12 MW

Inverters SKID number = 34

Trackers number = 3627

Modules number = 261144

Equipment Cost = $6.61M

Mechanical Cost = $8.56M

Electrical Cost = $15.17M

Civil Cost = $8.56M

Total Cost = $38.90M

Option #4: reducing wiring cost by using another low voltage wiring pattern

With option #3, we reduced the mechanical cost, but we increased electrical and civil expenses, so we would probably decide not to go any further with this option. Let's use option #2 as a basement for reducing wiring costs by using another low voltage wiring pattern. There are many different ways how current can go from modules to transformers, see some examples below:

In order to find the most efficient low voltage wiring EPC Contractor should go through many different options and be able of getting quick electrical wiring feedback on any layout changes. Let's leave this fascinating exercise behind the scenes and see what we got:

PV Farm Power = 122.03 MWdc

Inverters SKID number = 34

Trackers number = 3762

Modules number = 265290

Equipment Cost = $6.07M

Mechanical Cost = $9.68M

Electrical Cost = $11.15M

Civil Cost = $5.90M

Total Cost = $32.80M

Option #5: reducing equipment costs by using inverter skids more efficiently

There are many places we can refine equipment efficiency or electrical wiring efficiency, especially taking into account that we have 2033.4kWdc in reserve.

It is a fun process, and at the end of the day, we can be close to our target with minimum hard costs for the solar farm:

PV Farm Power = 120.013 MWdc

Inverters SKID number = 33

Trackers number = 3696

Modules number = 260898

Equipment Cost = $5.91M

Mechanical Cost = $9.43M

Electrical Cost = $10.86M

Civil Cost = $5.75M

Total Cost = $31.95M

Thus we found the best option:

Good analysis through pre-construction is the key to productive construction.

Summary

The idea of this blog was to go through the feedback loop, and trade-offs EPC Contractors typically face. I hope you understand that all our speculations were done in a conceptual way and real projects require a much more accurate and considerable way of analysing those scenarios. I have many other 'what ifs' in mind and would appreciate any ideas you can share with me. So feel free to write a comment below or to me personally.

Different solar PV farm parties need to consider so many exciting trade-offs. And the most efficient Utility-Scale Solar Farm is something everybody is striving for but not reaching.

Interaction and communication are the core of the AEC industry.