ANNUAL WEATHER DATA GENERATING

H. Yang, Y. Li, L. Lu, R. Qi (2011). First Order Multivariate Markov Chain Model for Generating Annual Weather Data for Hong Kong. Energy and Buildings. 43(9): 2371-2377.

https://www.sciencedirect.com/science/article/pii/S0378778811002556

In Yang et al. (2011), the authors considered the problem of energy system simulation. They introduced a stochastic approach, which is called the first-order multivariate Markov chain model, to generate the annual weather data for better evaluating and sizing different energy systems.

It is found that the Markov chain is a very effective way to describe weather variables, such as solar radiation, temperature and wind speed. However, the previous studies were limited to one single weather variable. The inter-relationship between different climate phenomena is omitted in these models. Yang et al. (2011) then employed the first-order multivariable Markov chain model proposed in Ching et al. (2002) to generate a series of weather data which involves solar radiation, air dry bulb temperature and absolute humidity. The categorical sequence is a sequence of weather variable with several possible states. The parsimonious first-order multivariable Markov chain model in Ching et al. (2002) provides an efficient and easy parameter estimation method, in which the estimation of the parameters is just a linear programming problem. Following this model, the state probability distribution of weather variable thus being able to be predicted without sophisticated and time-consuming calculation.

To validate the weather data generating model, Yang et al. (2011) analyzed 15-years actual hourly weather data of Hong Kong for generating the annual hourly data. The generated weather data was compared with the existing developed Typical Meteorological Year (TMY) and Test Reference Year (TRY) data of Hong Kong and long-term actual weather data. The comparison considered several important criteria, including general statistical characteristics, distribution probabilities and persistence probabilities. Although there are some differences between the actual weather data and generated one when the distribution probabilities and persistence probabilities are considered, the generated weather data is still acceptable when the data is compared with the TMY and TRY. While having comparable predicting power, the proposed weather data generating model has several advantages than TMY and TRY, i.e., no requiring weighting factors, better statistical characteristics and persistent structure.

W. Ching, E. Fung, M. Ng, A multivariate Markov chain model for categorical data sequences and its applications in demand predictions, IMA Journal of Management Mathematics 13 (2002) 87–199.

https://ieeexplore.ieee.org/document/8142556