Solar radiation zoning and daily global radiation models for regions with only surface meteorological measurements in China

Qimeng Cao,Yan Liu,Kailin Lyu,Ying Yu,Danny H.W. Li,Liu Yang

DOI: https://doi.org/10.1016/j.enconman.2020.113447

IF: 10.4

2020-12-01

Energy Conversion and Management

Abstract:<p>Daily global solar radiation is an indispensable variable in developing typical meteorological year and its data are important for constructing energy-efficient buildings. Solar radiation zoning and empirical models are essential for estimating the daily global solar radiation in regions that only have surface meteorological measurements. However, empirical models have become increasingly complicated without significant improvements in their accuracy. Thus, developing elaborate solar radiation zoning with minimal inputs and high accuracy daily global solar radiation models are key issues. In the present study, we develop an optimal criterion for solar radiation zoning by k-means clustering based on data obtained between 1970 and 2017 from 112 solar radiation measuring stations in China. According to the optimal criterion, solar radiation zoning is conducted based on data acquired by 1042 surface meteorological stations between 1970 and 2017. Zone models are established with linear relationships between the clearness index and relative sunshine duration based on data in each zone. Cross-validation is conducted to evaluate the performance of models by comparing individual models based on single station. The results show that the relative sunshine duration is the optimal criterion for solar radiation zoning and nine zones are identified. The zone models obtain good accuracy. Solar radiation zoning with zone models established in present study can be reliable and generalized to other regions with available sunshine duration data.</p>

energy & fuels,thermodynamics,mechanics

What problem does this paper attempt to address?

The problems that this paper attempts to solve are: in areas in China where only surface meteorological measurement data are available, how to develop fine solar radiation zoning and high - precision daily global solar radiation models. Specifically, the paper focuses on the following points:

1. **Solar Radiation Zoning**: Classify the data of 1,042 surface meteorological stations across the country through cluster analysis (k - means clustering) to determine the best solar radiation zoning criteria.
2. **Daily Global Solar Radiation Model**: Based on the data within each zone, establish a linear relationship model to estimate the daily global solar radiation \( H \), and use the relative sunshine duration \( S/S_0 \) as an input variable.
3. **Model Verification**: Evaluate the performance of regional models and single - station models at each station through cross - validation to ensure the accuracy and reliability of the models.

### Specific Problem Description

- **Background and Requirements**:
- The daily global solar radiation \( H \) is an indispensable variable in building energy consumption estimation and indoor thermal environment evaluation.
- In China, solar radiation measurement stations are sparsely and unevenly distributed, resulting in a lack of direct solar radiation data in many areas.
- Therefore, it is necessary to develop indirect methods (such as empirical models, machine learning methods, and satellite remote sensing methods) to estimate solar radiation.

- **Limitations of Existing Methods**:
- Although empirical models are simple and easy to use, their accuracy does not improve significantly as the complexity increases.
- Many empirical models rely on data from specific regions and are difficult to generalize to other regions.

- **Research Objectives**:
- Develop an optimized solar radiation zoning standard and classify 1,042 surface meteorological stations in China using the k - means clustering method.
- Establish a daily global solar radiation model based on the relative sunshine duration \( S/S_0 \) and evaluate its performance through cross - validation.
- Ensure that the established model can be reliably applied to other regions with similar meteorological data.

### Solutions

1. **Data Processing and Quality Control**:
- Collect and process data on daily global solar radiation \( H \), sunshine duration \( S \), maximum temperature \( T_{\text{max}} \), minimum temperature \( T_{\text{min}} \) and relative humidity \( RH \) from 1,042 surface meteorological stations.
- Conduct quality control on the data and exclude invalid and abnormal values.

2. **Solar Radiation Zoning**:
- Use the k - means clustering method to divide 1,042 meteorological stations into 9 zones according to the relative sunshine duration \( S/S_0 \).
- Determine that the optimal number of clusters is 9, so that the error sum of squares (ESS) reaches a stable point.

3. **Model Establishment and Verification**:
- In each zone, establish a linear model of \( K = a + b\cdot\frac{S}{S_0} \), where \( K \) is the daily clear - sky index and \( S/S_0 \) is the relative sunshine duration.
- Compare the performance of regional models and single - station models through cross - validation, and use indicators such as root - mean - square error (RMSE), mean absolute bias error (MABE) and Nash - Sutcliffe efficiency coefficient (NSEC) for evaluation.

### Conclusion

Through the above methods, the paper has successfully established fine solar radiation zoning and developed high - precision daily global solar radiation models. These models can be reliably applied to other regions with similar meteorological data, providing strong support for the design and evaluation of energy - efficient buildings.