Solar power plants need regular maintenance. Business interruptions may occur due to staff shortages. The companies Moxa and Thingnario have therefore joined forces in a technological project to promote the first intelligent solar energy monitoring system “Photon”. The main benefit of this solution is that it uses durable industrial grade equipment and artificial intelligence (AI) to greatly improve the efficiency of outdoor solar.
At a glance
The Taiwanese government has set itself the goalto obtain 20% of the electricity from renewable energy sources by 2025. It is predicted that solar energy will contribute 20 GW to achieve this goal. That is fifteen times more than the 1.3 GW that are currently installed. However, before the Taiwanese government can achieve its goal, many obstacles remain to be overcome.
In reality, many solar power plants lack a good maintenance strategy, which often results in the energy generated being lost. In addition, when operations are interrupted, companies have to spend a lot of money sending staff to repair the equipment. It is therefore similar to many other areas a maintenance strategy that works without people is becoming more and more interesting and now also practically feasible thanks to further developments in AI. In this sense, Moxa and thingnario have combined their technological advantages to jointly promote the first intelligent solar energy monitoring system “Photon” to increase efficiency.
Recurring maintenance expenses
The owner of the plant told about the complex processes involved in generating solar energy on large areas. Each solar station transmits between 20,000 and 50,000 field data per minute. It was too time consuming to use the traditional operating system to manage these huge amounts of data, and data loss was a serious problem. In addition, the previous maintenance system was hardware-based and relied heavily on the performance of the inverter. Under these circumstances, the operators did not have a holistic view of the operating status of the solar system via a common platform. In addition, there were other system integration issues and it was difficult to determine if the energy produced was being lost. And even if it was found that energy was actually being lost, it was almost impossible to determine the cause of the loss.
When anomalies occurred, the operators of the solar system had to dispatch maintenance personnel to look for the fault. This did not turn out to be easy because the plant operators have a large number of plant locations in Taiwan. As human resources were limited, this meant that problems were recognized and resolved too late. In fact, there were locations where 20 percent of the power was lost because bird droppings covered the panels – a general problem, because ultimately every single dust particle inhibits the light intensity. To resolve this issue, maintenance personnel had to travel to the remote area where the problem occurred to conduct a fault assessment, then return to the facility to collect the necessary equipment and parts. All in all, it took six months from discovery to fix. Mr. Zhang, the chairman of thingnario, realized that this was a significant pain point for the customer and he took steps to improve customer satisfaction.
Artificial intelligence to improve operational efficiency
The designed Photon system includes intelligent monitoring functions and intelligent operating mechanisms to ensure that the systems are effectively monitored. By providing the maintenance staff with the correct information in a timely manner, they were able to increase the overall operational efficiency of the solar station.
The technology produced was designed in such a way that it records all the data from the inverter as well as the weather data from the Sky Camera, which was developed in-house and provides information about cloud cover and the effects on the solar modules.
Photon has five main functions:
- Artificial intelligence (AI) capabilities: The AI ​​engine analyzes large amounts of historical and real-time sensor data to identify patterns and predict how much electricity will be generated in the next 5 to 30 minutes. When there is a large discrepancy between the amount of energy predicted and the amount actually produced, the system sends an alert to notify the solar system operators so that they can carry out preventive maintenance.
- Highly scalable: Photon processes time stamp data and business data separately in the backend database. This ensures that all data is processed quickly for every solar system, regardless of its size.
- O&M Task Management System: It manages all events including maintenance, attendance and expenses. The digitization of operating costs together with the AI ​​system helps new operators to familiarize themselves more quickly and reduces the time and thus the costs of troubleshooting.
- Simple three-step setup for data collection: first: set the solar station information on the software side, second: configure the network settings, and third: carry out field work
- Dashboard for instant data analysis: The single-line diagram, the system configuration map and the real-time data make the operational status clear.
Ensuring durability, reliability and longevity
Before the company was able to accurately capture huge amounts of data, it had to make sure that the industrial computer they chose met all of the project’s requirements. After careful analysis, Moxa’s arm-based industrial computers of the UC series were found to be the most suitable. The fanless industrial computers of the UC series from Moxa not only offer efficient and stable data acquisition functions, but are also equipped with several serial interfaces to simplify the connection with inverters, pyrheliometers and other measuring devices. The low power consumption also reduces maintenance costs. In addition, they withstand temperatures from -40 to 70 ° C and meet all of Thingnario’s expectations.
Moxa has several branches around the world and distributors in over 70 countries. This global as well as local expertise should make a significant contribution to the expansion of global business.
Real-time information to increase productivity
In addition to the AI ​​analysis used to predict the amount of electricity produced by the solar array, the real-time alerts (which were not available in the previous system) helped the operator increase the amount of electricity produced by 10% and reduce labor costs by almost 30% . The EPC (engineering, procurement and construction) contractor responsible for maintaining the facility also recognized the benefits of a good monitoring and operating system that presented useful information on a one-page dashboard.
08/24/2021