Evaluation of an Integrated Low-Cost Pyranometer System for Application in Household Installations ^†

Abstract

The climatic conditions of a region are a constant object of study, especially now that climate change is clearly affecting quality of life and the way we live. The study of the climatic conditions of a region is conducted through meteorological data. Meteorological installations include a set of sensors to monitor the meteorological and climatic conditions of an area. Meteorological data parameters include measurements of temperature, humidity, precipitation, wind speed, and direction, as well as tools such as an oratometer and a pyranometer, etc. Specifically, the pyranometer is a high-cost instrument, which has the ability to measure the intensity of the sunshine on the surface of the earth, expressing the measurement in Watt/m². Pyranometers have many applications. They can be used to monitor solar energy in a given area, in automated systems such as photovoltaic system management, or in automatic building shading systems. In this research, both the implementation and the evaluation of an integrated low-cost pyranometer system is presented. The proposed pyranometer device consists of affordable modules, both microprocessor and sensor. In addition, a central server, as the information system, was created for data collection and visualization. The data from the measuring system is transmitted via a wireless network (Wi-Fi) over the Internet to an information system (central server), which includes a database for collecting and storing the measurements, and visualization software. The end user can retrieve the information through a web page. The results are encouraging, as they show a satisfactory degree of determination of the measurements of the proposed low-cost device in relation to the reference measurements. Finally, a correction function is presented, aiming at more reliable measurements.

1. Introduction

According to meteorological data, solar radiation on the Earth’s surface is an important parameter to be measured, as it concerns both atmospheric science, in terms of studying the variation of solar radiation [1], and the design of renewable energy sources such as solar photovoltaic systems [2]. A large number of pyranometers have been proposed according to [3,4], which are costly in terms of both purchase and maintenance. Research has [5] discussed photodiode pyranometers and outlined their advantages in measuring solar irradiance with quick response times compared to traditional thermopile pyranometers. These advantages make photodiode sensors a popular option for tracking solar energy, particularly in remote locations. However, the issues of aging and performance degradation need to be addressed in long-term installations, underscoring the necessity of ongoing evaluations of sensor performance. Another study explores a method to calculate the uncertainty of measuring shortwave solar irradiance using thermopile and semiconductor radiometers. This research highlights the persistent challenges in achieving measurement accuracy despite technological advancements, emphasizing the importance of refining models to enhance data reliability [6]. This foundational knowledge is crucial given the reliance on such accuracy in real-world solar applications. The work in [7] emphasizes that thermopile pyranometers achieve the lowest uncertainty under most conditions, although they are more expensive and require frequent maintenance in dusty environments, particularly arid regions. On the other hand, silicon photodiodes are a cost-effective option that require less maintenance but have a limited spectral response, leading to higher uncertainty. This comparison underscores the practical trade-offs associated with pyranometer selection, considering factors like cost and long-term reliability. The research conducted in [8] discusses the stability and accuracy of pyranometers, specifically focusing on the drift bias and exposure errors that can affect solar radiation data. Their findings reveal that Eppley PSP pyranometers experience a decrease in responsivity over time, correlating with changes in the spectral response due to the aging of the sensing material. This is pertinent given that the efficiency of some materials can similarly suffer from aging effects, thus necessitating careful monitoring and maintenance to ensure accurate solar irradiance measurements. The study [9] focuses on the measurement and analysis of thermal offset in a common type of pyranometer using a capping-based method and compares different evaluation criteria. Measurements under various weather and environmental conditions revealed significant differences in thermal offset between day and night, as well as between cloudy and clear skies. The findings suggest that ambient temperature, radiation levels, and the ratio of direct to diffuse irradiance are key factors influencing daytime thermal offset. In another study [10], the authors explore methods for measuring solar radiation and highlight their importance in site-specific testing in Portugal. They examine various modeling approaches—empirical, time-series, AI-based, and hybrid models—each with strengths and limitations shaped by atmospheric, geographic, and climatic factors. The study stresses the importance of selecting models suited to regional conditions to ensure accurate solar radiation estimates, which are essential for improving energy efficiency and supporting sustainable building practices. The development of technology, and in particular of low-cost sensors, provides the opportunity for the construction of a low-cost pyranometer. Several works have presented the construction of a low-cost pyranometer using a thermoelectric sensor (Peltier) [11,12], photodiode [13,14,15], phototransistor [16,17,18], light-dependent resistor (LDR) [19,20], and photovoltaic panel [21,22]. According to the aforementioned works, data is either recorded locally (SD card) or uploaded to cloud applications via the Internet. With the evolution of microcontrollers and the integration of data transmission protocols (Wi-Fi, LoRa, XBee), it is easy to build Internet-of-Things (IoT) devices such as in the work [23] presented here, the study of a LoRa network coverage radius in urban and rural tissue and a low-cost IoT arrangement, using a LoRa network, for measuring a water station in a village water tank. Works [24,25] present a low-cost device based on a wireless network for studying battery quality, and a low-cost device using a wireless network for UV radiation measurements, respectively. There are many applications for such a system (pyranometer) besides meteorology. In photovoltaic park facilities, the implementation of low-cost solar radiation sensors is essential for effective energy management and optimization. Study [26] highlights the importance of using inexpensive materials, such as green light-emitting diodes (LEDs), which can complement traditional methods while reducing costs. The research indicates that a mathematical model can be employed to account for local geographical variations in solar energy absorption, an essential factor for optimizing the efficiency of solar installations. Another work [27] has detailed the design and implementation of a solar radiation sensor using an Arduino Uno paired with a reference cell. The authors highlight the cost-effectiveness of the device, which is suitable for both off-grid and grid-connected solar installations, emphasizing the utility of the Proteus design software for efficient validation and testing of the hardware. The research conducted in [28] illustrates a microcontroller-driven design that demonstrates efficient tracking performance, with potential applications in large solar parks to enhance energy yield. The integration of such systems with shading devices is critical for optimizing indoor climates and enhancing energy efficiency. Studies show that manual solar shading can reduce glare and enhance thermal comfort, which directly correlates with lower energy demands for cooling [29,30]. As the authors of [30] point out, well-designed shading devices can significantly mitigate solar heat gain, leading to a reduction in overall cooling load and operational costs. Furthermore, [31] provided evidence that occupant behavior can influence shading operations, underscoring the importance of interactive design in shading systems to adapt to indoor light and temperature conditions.

In this paper, an integrated solar radiation monitoring system is presented. The innovation of the system is due to the presentation of an affordable integrated solar radiation monitoring system, which includes a solar radiation measuring station and an information system. The monitoring station uses an affordable light sensor, photodiode technology, and a very affordable and common microcontroller that meets the functional requirements of the device. The system consists of two parts: Firstly, the IoT (low-cost pyranometer) device, consisting of a microcontroller and photodiode sensor. Secondly, the information system for collecting, storing, and displaying the data to the end user via the Internet. All the software used is open source, aiming at the free circulation of ideas and community participation. In addition, the system was evaluated to demonstrate the reliability of the measurements. The innovation of the system is based on five pillars:

• increased accuracy;
• open-source software infrastructure;
• integrated system;
• low energy consumption;
• outdoor durability.

All this makes the proposed low-cost system affordable and unique, as the aforementioned research works do not combine the characteristics of the proposed low-cost integrated solar radiation monitoring system.

2. Materials and Methods

2.1. Experimental Setup

For the implementation of the low-cost pyranometer, the WeMos D1 mini [32] microcontroller (Figure 1a) was used. The WeMos D1 mini is a very compact 802.11 (Wi-Fi) wireless microcontroller development board. It is based on the popular ESP8266 wireless microcontroller in an integrated development board. This microcontroller is affordable and presents low power consumption and satisfactory computational power according to the requirements of this project. More specifically, it has eleven (11) Digital I/O Ports and one (1) Analog Input Port, while working at a clock speed of 80 MHz and supported by a memory of 4 M bytes. It also supports communication protocols such as I2C, SPI, UART interfaces, and wireless (Wi-Fi) networks. For the solar irradiance, the digital light intensity sensor module BH1750 [33] (Figure 1b) was used, which presents a satisfactory range of brightness while it uses the I2C communication protocol. A plastic box (IP67) accommodates all of the above, where the upper part is transparent for the inflow of solar energy. The final construction of the low-cost pyranometer is presented in Figure 1c. The Wemos D1 mini was programmed with the user-friendly Arduino IDE software (v1.8.19). Data transmission is done over the Wi-Fi protocol. Since energy efficiency is a key issue in low-cost systems, this is achievable via affordable microcontrollers through a set of events that must be evaluated during implementation [34].

The information system is based on a Linux operating system, and the data is collected and stored in an InfluxDB [35] database, which is ideal for time-series analysis. To visualize the data, the Grafana Lab software [36] (ver. 9.5.2) was used, where the user can retrieve the data through a web page. The communication among the low-cost pyranometer, the database, and the visualization software is facilitated through an API key.

2.2. Experimental Results

The experimental results from the integrated solar radiation monitoring system are presented below. For this study, a two-week period (25 March 2024 to 9 April 2024) was used. The two graphs shown in Figure 2 concern first, the luminance received by the sensor, as the sensor measures luminance (Lux), and second, the conversion of luminance into solar energy, given that for the solar light spectrum, 1 Lux is equal to 0.0079 Watt/m²; also, given the response curve of the sensor [37], the code of the microcontroller was modified to present the measurement in Watt/m².

3. Results and Discussion

To evaluate the results, a correlation method was applied between the measurements from the low-cost pyranometer and the official measurements from a nearby meteorological data monitoring station [38]. Figure 3 shows the time series of raw measurements of the low-cost pyranometer and reference measurements for the two-week period (25 March 2024 to 9 April 2024).

When investigating the sensor response, it was deemed appropriate to study in-depth the repeatability and seasonality of the low-cost sensor. For this reason, three days (26 March 2024, 28 March 2024, and 1 April 2024) are presented and analyzed separately. The days selected include one day when it was cloudy and two days when there was sufficient sunshine. In this way, it is possible to identify the variation in the measurements returned by the sensor in terms of both repeatability and seasonality, in the sense of the time periods of a day, such as morning, noon, and afternoon. More specifically, Figure 4a shows the time series of measurements between the proposed system and the reference for 26 March 2024 (a cloudy day), while Figure 4b shows the percentage error (%) between the two aforementioned measurements.

The next two figures present the measurements of full sunlight days (28 March 2024 and 1 April 2024). Figure 5a and Figure 6a show the time series of measurements between the proposed system and the reference for 28 March 2024 and 1 April 2024, respectively, while Figure 5b and Figure 6b show the percentage error (%) between the two aforementioned measurements of each day, respectively.

The following conclusions can be drawn from Figure 4, Figure 5 and Figure 6: The error appears to be greater on days when there is shade, such as cloudy days. The sensor exhibited similar behavior during the morning and evening hours, when the sun does not shine directly on the sensor, resulting in an increase in error. During the day, it is observed that, especially on the cloudy day (26 March 2024), the average error approaches 80% throughout the entire period of sunshine, while on the other two days, when the sky was clear, the average error was limited to a maximum of 15%. It should be noted that both repeatability and seasonality (daily) are symmetrical, especially on days with clear skies.

For data analysis purposes, Figure 7 shows the time series and correlation of raw measurements of the low-cost pyranometer and reference measurements for the two-week period from 25 March 2024 to 9 April 2024

The low-cost pyranometer arrangement gives satisfactory results, given that the sensor has a limited spectrum in the wavelength it receives. The reliability of the system is shown by the determination of the R² coefficient, which is very high (between the low-cost and reference measurements). Although the determination of the R² coefficient is satisfactory, there is a variation between the measurements (low-cost and reference pyranometers). In particular, it is observed that if the measurement values are higher, then the greater the variation in the measurements between the low-cost and reference measurements. This identifies a linear relationship between the low-cost and reference measurements. The variations in the measurements can be corrected by applying a linear equation (Equation (1)) as a correction factor, in accordance with the linear relation of the correlation graph (Figure 7b).

$$y=ax+b$$

(1)

Τhe application of the linear equation as a correction factor shows an improvement in the measurements using the linear coefficients of the equation as correction factors [38,39,40]. Using the linear equation as the correction factor of the measurements, after continuous experiments and tests, Equation (2) was derived as the correction equation of the measurements.

$${LC}_{corr}=(1.263\times {LC}_{raw})-24$$

(2)

where ${LC}_{corr}$ is the corrected measurements of the low-cost pyranometer, and ${LC}_{raw}$ is the raw measurements of the low-cost pyranometer.

Figure 8 shows the time series and scatter plot, corrected by the Equation (2) measurements of the low-cost pyranometer and reference measurements for two weeks from 25 March 2024 to 9 April 2024. Figure 8a shows the time series of the corrected measurements of the low-cost pyranometer and reference measurements, and Figure 8b shows the scatter plot (correlation and the determination of correlation) between the corrected measurements of the low-cost pyranometer and reference measurements.

According to Figure 8, for both the time-series plot (Figure 8a) and the scatter plot (Figure 4b), the correction is evident. In addition, confirmation of the measurements’ correction is shown by the linear coefficient $a$ and $b$ of the linear Equation (1) from Figure 7b and Figure 8b, where in Figure 8b, coefficient $a$ is close to 1, and coefficient $b$ is close to 0, which means more realistic measurements.

Finally, for the evaluation of the reliability of measurements, the methods of mean square error (MSE) and root mean square error are applied to the set of measurements for the duration of two weeks.

Table 1. Evaluation of measurements via the MSE and RMSE methods.

Method	MSE	RMSE
Primary (raw)—reference measurements	2798.66	0.00034
Corrected—reference measurements	2090.00	0.0000001

shows the MSE and RMSE values.

The results of the MSE and RMSE methods in Table 1 show the importance of the correction, as well as the way the correction was made. Through the results of the MSE and RMSE methods, it appears that the correction is decisive, and the improvement is evident to a high degree.

The findings of this study show that a low-cost sensor can be used to measure solar radiation. Although it shows satisfactory results and good reliability, it should be noted that after analysis, it was observed that on days with minimal sunshine, the average deviation is around 15%, while on cloudy days, this percentage approaches 80%.

Another important finding is the proposal of a linear correction function for the measurements, which acts on the linear coefficients with the aim of achieving more realistic measurement values. The reliability of the correction equation is confirmed by the MSE and RMSE methods, where in both methods the corrected values show a better degree in relation to the raw values.

The use of low-cost sensors can be efficient and immediately applicable thanks to their affordability and responsiveness. Obviously, low cost can result in low responsiveness and low measurement reliability. This work presents a comprehensive solar energy monitoring system along with a calibration method, which, according to the results, shows a correction of measurements with the aim of achieving more realistic measurement values. It should be noted that the aging of sensors during operation is a reality and has a direct impact on measurements. Their exposure over a long period of time can better reflect both the behavior of the sensor and its performance throughout the seasons and under different environmental conditions. The study of both long-term durability and aging is a field of research for all types of sensors, and for our research team, the performance of the proposed system over a long period of time is a future task. Finally, the feasibility of the proposed system for interconnection with other systems (with real-time applications) should be mentioned, such as for the management of photovoltaic parks, sun tracking systems, automatic shading systems in building installations, etc. The wireless networking (Wi-Fi) capability offered by the microcontroller makes it possible to send both commands and server messages using various methods via the Internet (such as POST, GET, etc.). Of course, the proposed system can also be connected to other applications via the information system that supports the online monitoring system, using appropriate software such as Python and JavaScript.

4. Conclusions

In this work, an integrated low-cost solar radiation monitoring system was presented. This system has applications both in atmospheric science and in renewable energy systems (photovoltaic systems). The proposed low-cost solar radiation monitoring system consists of a microcontroller and photodiode sensor. The construction materials are affordable, with a total cost of approximately EUR 30.

The proposed low-cost device provides excellent results, according to the evaluation procedure, and presents a high degree of reliability, as the correlation coefficient determination shows a value of R² = 0.97. For the optimization of the measurements, the application of a simple linear equation as a correction factor using the linear coefficients of the equation was investigated with satisfactory results, as the corrected values from the low-cost pyranometer are very close to the reference values.

Τhe reliability of the measurements is also demonstrated by the MSE and RMSE methods, where the results show both the importance of correcting the measurements and the improvement of the measurements. In general, low-cost systems (e.g., microcontrollers, sensors) can operate satisfactorily, especially in field conditions, and are increasingly the subject of research and study.