Chiun-Hsun Chen’s research while affiliated with Feng Chia University and other places

An increase in renewable energy injected into the power system will directly cause a fluctuation in the overall voltage and frequency of the power system. Thus, renewable energy prediction accuracy becomes vital to maintaining good power dispatch efficiency and power grid operation security. This article compares the one-day-ahead PV power forecasting results of three models paired with three groups of weather data. Since the number, loss, and matching problem of weather data will all influence the training results of the model, a pre-processing data framework is proposed to solve the problem in this study. The models used are a deep learning algorithm-based artificial neural network (ANN), long short-term memory (LSTM), and gated recurrent unit (GRU). The weather data groups are Central Weather Bureau (CWB), local weather station (LWS), and hybrid data (the combination of CWB and LWS data). Compared to the other two groups, hybrid data showed a 5–8% improvement in measurements. In addition, when it comes to different weather conditions, the advantages of the LSTM model were highlighted. After further analysis, the LSTM model combined with hybrid data showed the most accurate measurements, which was proved through forecasting results for one month. Finally, the results indicate that when the amount of data is limited, using hybrid data and the five weather features is helpful for training the model. Accordingly, the proposed model shows better one-day-ahead PV forecasting.

Solar photovoltaic (PV) power generation is prone to drastic changes due to cloud cover. The power is easily affected within a very short period of time. Thus, the accuracy of grasping cloud distribution is important for PV power forecasting. This study proposes a novel sky image method to obtain the cloud coverage rate used for short-term PV power forecasting. The authors developed an image analysis algorithm from the sky images obtained by an on-site whole sky imager (WSI). To verify the effectiveness of cloud coverage rate as the parameter for PV power forecast, four different combinations of weather features were used to compare the accuracy of short-term PV power forecasting. In addition to the artificial neural network (ANN) model, long short-term memory (LSTM) and the gated recurrent unit (GRU) were also introduced to compare their applicability conditions. After a comprehensive analysis, the coverage rate is the key weather feature, which can improve the accuracy by about 2% compared to the case without coverage feature. It also indicates that the LSTM and GRU models revealed better forecast results under different weather conditions, meaning that the cloud coverage rate proposed in this study has a significant benefit for short-term PV power forecasting.

This paper described full-scale tests of water mist fire suppression system under forced desmoking conditions. Pool fire was effectively suppressed in 25.0 and 37.0 s for 1,500.0 (test 1) and 3,000.0 (test 2) mL of oil with no ventilation, respectively. Moreover, under ventilation rates of 50.0 and 75.0 m³/min, the suppression time were lengthened to 54.0 s (test 1) and 202.0 s (test 2), and 102.0 s (test 1) and 222.0 s (test 2), correspondingly. However, the smoke exhaust had a strong effect on the water mist to result in fire could not extinguish, when the ventilation rate increased to 120.0 m³/min. The temperature distribution during the early period in the scenario of no ventilation has a higher magnitude than those scenarios of 50.0, 75.0, and 120.0 m³/min, because combustion proceeds smoothly to result in temperature increased rapidly at no ventilation. Furthermore, similar results presented in smoke concentration distribution at four scenarios (0, 50.0, 75.0, and 120.0 m³/min), and combustion occurred steadily at no ventilation to result in more smoke during the early period. This study employed a computer modeling to verify water mist system in a compartment space and determine how the water mist system could be optimized.

In most of the fault detection methods, the time domain signals collected from the mechanical equipment usually need to be transformed into frequency domain or other high-level data, highly relying on professional knowledge such as signal processing and fault pattern recognition. Contrary to those existing approaches, we proposed a two-stage machine learning analysis architecture which can accurately predict the motor fault modes only by using motor vibration time-domain signals without any complicated preprocessing. In the first stage, the method RNN-based VAE was proposed which is highly suitable for dimension reduction of time series data. In addition to reducing the dimension of sequential data from 150*3 to 25 dimensions, our method furthermore improves the prediction accuracy evaluated by several classification algorithms. While other dimension reduction methods such as Autoencoder and Variational Autoencoder cannot improve the classification accuracy effectively or even decreased. It indicates that the sequential data after dimension reduction via the RNN-based VAE still can maintain the high-dimensional data information. Furthermore, the experimental results demonstrate that it can be well applied to time series data dimension reduction and shows a significant improvement of the prediction performance, even with a simple double-layer Neural Network can reach over 99% of accuracy. In the second stage, Principal Components Analysis (PCA) and Linear Discriminant Analysis (LDA) are used to further perform the second dimension reduction, such that the different or unknown fault modes can be clearly visualized and detected.

In Taiwan, the rapid development of the petrochemical industry over the past four decades has resulted in economic progress and technological advancement. However, this success has been accompanied by pollution and sporadic accidents. Numerous accidents related to crude oil and its derivatives have resulted in a loss of life, employee injuries, environmental and property damage, economic decay, social outcry, and even political turmoil. This study investigated a caprolactam storage tank accident through root cause analysis with a computational approach. The previously described approaches returned findings that indicated causes similar to the real cause of this caprolactam accident. The four summarized descriptions of false conditions of the caprolactam explosion can serve as a reference for other investigations into the causes of petrochemical accidents. Reason of the caprolactam storage tank accident was determined by root cause analysis accompanying with a computational approach.

This research describes a novel system of four Savonius wind rotors, aligned in parallel. A solar panel system, generating power by solar energy, was positioned in front of the parallel system to guide the airflow impinging on the rotors. This array was installed in rural areas to generate electric power. The effect of the solar panel deflector arrangements, used to guide the air stream, was investigated. This type of plant development has not previously been examined. We employed a computational fluid dynamics software, Fluent, to analyze the flow fields and system performance prior to experimentation, then compared these simulations to our experimental data. The parameters studied include wind velocity, wind direction (with/without solar panel deflector), and the rotational speed of the rotors to identify the relationship between the tip-speed ratio (TSR) and power coefficient (Cp). For the numerical simulation results at TSR 0.8, the maximum Cp value of the parallel system without a solar panel deflector was 0.289, whereas at the optimal spacing between the parallel systems with a deflector (50 cm), the Cp was 0.389. This represents a difference factor of 1.35 between the two Cp values. The velocity vector distribution showed that the deflector could guide the airflow to impinge on the rotors from below, gaining extra wind power. The experimental results show that the wind velocity and rotational speed of the wind rotors exhibit large fluctuations in open fields. To combat this, experiments were repeated in both day and night conditions, in different seasons, to gather a range of Cp and TSR values. The average measured wind speed was 6.99 ± 1.52 m/s. Four Savonius wind rotors in a parallel system can generate 8.25 kW h of energy per day, with an optimal power generation efficiency of 20.7%. Our 3-kW hybrid wind and solar system, which used optimal simulation conditions to determine its experimental design, can generate 14.55 kW h of power per day, with a corresponding optimal power generation efficiency of 21.7%. Our measured Cp curve shows that a deflector can improve the system performance by up to 10.1%.

This work describes a novel system of four Savonius wind rotors in a parallel matrix, installed in rural areas to generate electric power. The effect of wind deflector arrangements, used to concentrate the air stream, is investigated. This type of plant development has not previously been examined. We employ a computational fluid dynamics software, Fluent, to analyze the flow fields and system performance in advance, then compare these simulations to experimental data. The parameters studied include wind velocity, wind direction (with/without deflector) and the rotational speed of the rotors, with the aim of identifying the relationship between the tip-speed ratio (TSR) and power coefficient (Cp). In simulation results, at a TSR of 0.6 the system with a wind deflector performs 1.16 times higher power coefficient than the system without a deflector. Generally, the addition of a wind deflector enhances the performance by 1.09 times, especially within the lower TSR regime (0.2-0.8). The experimental results show that the system with a wind deflector performs 1.23 times higher than the one without deflector, with this maximal difference occurring at a TSR of 0.7. However, when we plot experimental Cp against TSR, we find that the wind deflector enhances the performance by 1.15 times in the higher TSR regime (0.6-1.0). The large fluctuations that occur in measurements are attributable to open field tests where the parameters cannot be controlled as precisely as those in simulations.

This study conducted 30 kW-generator experiments at a small biogas plant on a swine farm to serve as a preliminary study for the construction of a 300-kW power plant for a larger scale biogas plant in the near future. The experimental results showed that the optimum biogas flow rate for the engine was around 240 to 260 L/min, and the maximum power generation, the corresponding thermal efficiency, and the percentage of consumed CH4 were 26.8 kW, 28.7%, and 95.4%, respectively, at a biogas supply rate of 260 L/min. With 3% oxygen-enriched air, the maximum power generation, thermal efficiency, and percentage of consumed CH4 increased up to 28.2 kW, 30.2%, and approximately 100%, respectively, for a biogas supply rate of 260 L/min, and the engine can operate normally at a lower limited fuel supply rate of 220 L/min. The heat exchanger could recover 923 kJ/min of heat from the flue gas, leading to an overall efficiency of 47.3%, at a biogas supply rate of 240 L/min. With the total swine population (around 4.3 million heads) from the farm scale of over 1,000 heads of pigs in Taiwan, the estimation, based on the data obtained from this study, shows the economic benefits of using biogas, including an annual electricity generation of 2.67 × 108 kWhe (corresponding to electricity charge savings of US$26.7 million), where natural gas charge savings total US$ 8.7 million, and the carbon dioxide reduction totals 180,000 tons.

A 30 kW generator at the Taiwan Sugar Swine Farm in Taichung was used to collect data intended for long-term electricity generation. This three-part experimental study utilized a biogas with a 73% methane concentration. First, the effect of the biogas supply rate on generator performance at different excess air ratios is investigated. Second, the results from this study are compared with results obtained from work performed on a biogas with a 60% methane concentration. Finally, for a 73% methane concentration biogas, the effect on generator performance of preheating the inlet gas to different temperatures is investigated by applying a waste-heat recovery system.Power generation increases with increasing biogas methane concentration, except when λ (excess air ratio) < ∼0.85. However, thermal efficiency increases with increasing methane concentration only when λ > ∼0.95, although on the relatively rich side (λ < 0.95), there is no benefit. The improved generator performance obtained by preheating the inlet gas is apparent when the excess air ratio is relatively high, such as when λ > 1.3.