Maximizing profits from PV installations in 2023 – part 1
Maximizing profits from PV installations in 2023 – How to increase the amount of energy produced while optimally managing operational costs. Interview with Michał Frys.
We are entering a new year, and new challenges for the photovoltaic and renewable energy market are ahead of us. The latest changes concern not only statutory limitations in increasing energy sales prices, but also the operating costs of maintaining farms are increasing with inflation. Are the owners of PV installations able to find a way out of this situation?
In my opinion, yes.
Considering the limitations of the market right now, you need to go beyond your current comfort zone. When energy sales prices were breaking records and the costs of financing and operating PV installations were lower than today, few investors paid attention to the efficiency of solar farm production. Sometimes we met an almost absurd approach of owners, who argued that the installation pays off better than it was expected in the business plan, so there is no need to pay attention to efficiency. There was also no will to look at the accelerated degradation of farms in the long-term, that is a result of unrepaired defects.
You say these sentences in the past tense. Does this mean that this period has already passed?
I’m not saying it’s over. There is a rising trend where the attitude is changing – both of the owners and of the companies that have been entrusted with managing these photovoltaic installations.
With the change in the economic situation and with the introduced restrictions on energy sales prices, participants of the photovoltaic market noticed that it is worth looking into the topic of efficiency analysis.
In my experience, maximizing the efficiency of solar farms is the most underestimated area of increasing the return on investment.
And when it comes to the significant increase of operating costs?
Yes, the second element that greatly affects the profitability of a photovoltaic farm is the issue of incurred costs. It is necessary to consider here how we manage the budget allocated to maintaining the efficiency of the farm. But this is a more complex issue and I think that we should not mix the two topics.
Let's go back to the issue of maximizing energy production efficiency for photovoltaic installations. Is there one solution that is best in this area?
No, I think that we cannot say that there is one solution that can be applied to every situation. The owners we talk to have different levels of awareness in the area of performance analysis. However, the first question I almost always ask is, “Do you know how much energy your farm produces, and how much energy it should be producing?”
Is this question so difficult to answer?
I can say yes and no. This is seemingly a very general question, which, nevertheless, is not easy for my interlocutors. They are unable to give a definite answer. Regardless of whether I am talking to the owner of a small installation of 1 MW or to an Asset Manager who oversees a large portfolio of photovoltaic installations. The answer is never easy.
Shouldn't this matter be obvious to every energy producer?
Of course, yes, but the photovoltaic market is very young, and the issue of reliable performance analysis is not standardized.
I recently talked to a client and asked about the efficiency of the generation. He answered that in the second year of operation of the farm it increased by over 20%. This was compared to the calculations received in the planning phase from the company that was designing and constructing the installation. From this particular owner’s view, this was really positive news, but it raised many questions in me.
What sort of questions?
Let’s start with this – what sort of data we should use to measure the potential of a farm.
According to the latest standards, the key here is to have qualitative data on irradiance, i.e. the degree of insolation that occurs over our farm at the analyzed moment. Only with reliable, up-to-date, and complete data, obtained for a given location, can we prepare calculations on the energy generation potential. And here we touch on one of the main challenges, which is the use of historical data.
A significant part of investors are not (yet) experts in the photovoltaic industry. Often, they base all their knowledge on information obtained when deciding whether it is worth “entering” this industry. Here I mean the stage of designing and building a farm in a given location. This is when they think about how profitable it is, and how much you can earn on a photovoltaic farm. Companies prepare a design of the PV installation for them and then calculate the return on investment. However, they have no way of estimating how much the sun will shine over the installation for the next 20 or 25 years. Therefore, they use averaged historical data.
In many cases, investors enter the operational phase, accepting this as the only one possible state. They do not think that there are tools available on the market, that can analyze the level of efficiency of PV farms – ongoingly and automatically, based on current data.
Are these figures so different? The data is chosen for the given location.
Yes, they are. Firstly, because of climate change, the weather is different every year. There are years of drought, there are years with floods, and here we are talking about data that is averaged from over 8, 12 or even 16 years. Based on such averaging, resulting both from the lack of taking into account weather variability, but also from a significant generalization in the field of data collection for the location, relating the current production volume to such data is simply unreliable.
In this situation, can investors who want to reliably analyze the performance of their installations - obtain quality data and perform analyzes on their own?
Of course, but it is worth considering whether it is worth it.
What do you mean?
There are three key factors that are the foundation of a reliable analysis of production potential. These are reliability, completeness and, above all, accuracy of the data based on which we want to analyze performance.
The first thing that comes to our customer’s mind is to obtain data from a weather station installed directly on the farm. But this solution is far from being reliable and dependable.
Often, the data collected in this way deviates from reality, for example due to the dirt on the station or its bad location. Even the inaccuracy of the measurements stated in the device specifications can be between 3 and 6%, and this is a really significant error.
In addition, a common problem that occurs on farms are periodic interruptions in communication. This results in gaps in the data and, consequently, the inability to make a reliable analysis and answer the question whether the farm produced as much as it could in certain weather conditions.
Two other significant factors are the issues of competence and workload.
Are you suggesting that conducting such analyses on your own is ineffective?
In my opinion, that is exactly the case.
Ensuring that the irradiant level data over the selected installation is of good quality requires it to be obtained from many sources, such as meteorological data, data from the local weather station, historical data and, in some cases, satellite data. – And this does not exhaust the topic, because the next step must be a cross-analysis of this data and verification of their fit to a given installation, at a given time of the year, month, or even day.
What do you recommend in such a situation?
Taking into account the required level of competence and the workload of the process, and according to the experience of our clients – it is best to entrust this work to specialized companies that, apart from competence, have created software that can cope with such a large number of calculations.
Working with such a partner, we can easily obtain a high-quality analysis of the performance of our installation, without exposing ourselves to high costs of its preparation, delays related to the time-consuming calculations, as well as inaccuracies related to human errors.
An equally important issue as the reliability of the analysis is to obtain information about detected performance drops in the shortest possible time. And here we touch on the challenge of the photovoltaic industry in a broader sense, which is the lack of constant supervision over the efficiency of an operating farm.
But many photovoltaic installations have something like SCADA software, so it is possible to monitor the efficiency on an ongoing basis.
I am glad that we have raised this issue.
It is true that many installations are supported by a more or less extensive SCADA solution, but most of these solutions have two drawbacks that make them unable to conduct a reliable analysis of the efficiency of photovoltaic installations.
The first issue is measuring the level of farm efficiency in relation to the average historical irradiance data from several previous years, to which these systems compare the currently analyzed production readings.
The second element that makes the SCADA solution insufficient in the category of a good performance analysis tool – is the lack of using “intelligent” algorithms, crucial for correct interpretation of the acquired data.
What do you mean by an “intelligent” solution?
Today’s technology allows us to “teach” software not only to transmit data, but also to be able to interpret it. SCADA systems do not have “sewn-in” intelligence to interpret the acquired data, they are limited only to displaying them.
For example, when a photovoltaic installation is shaded by a cloud moving over it, the SCADA system will read a sudden drop in the level of energy production on the given strings of the installation and will sooner or later display an “alarm” to the dispatcher about such an event, not being able to interpret it. When the sky is cloudy and the weather is changeable, such a system “floods” the dispatcher with worthless notifications, which as a result of their multitude are often turned off.
The situation is different with intelligent solutions, where each acquired data is analyzed in terms of the characteristics of its occurrence, as a result of which only valuable notifications about performance drops resulting from defects are displayed to the dispatcher.
To simplify, we can call this step of the analysis – cleaning the source data from useless information noise. However, this is a much deeper issue, related to competences in the field of artificial intelligence and machine learning, which I could talk about for hours.
In conclusion, we can be sure the photovoltaic market will be facing new challenges in 2023.
For now, we can see that the majority of difficulties will be mainly connected to limitations on energy sales prices and a substantial increase in operating costs.
However, owners and managers of photovoltaic installations can maintain profitability and continue to succeed by keeping track and by properly analyzing their solar production.
We believe it is crucial to maximize the efficiency of solar farms and to carefully control maintenance budgets. – And all this can be done thanks to the smart use of innovative performance analysis tools, which we will be happy to tell you more about in part two of our interview.
