본 연구에서는 Monopile 방식 풍력발전기 강구조물의 부식을 방지하기 위하여 S355 steel의 표면 거칠기에 따른 용사 코팅 상태에 관한 연구를 수행했다. 일차적으로는 시편별 서로 다른 표면거칠기를 부여하기 위해 밀링머신에 페이스 커터를 결합하여 시편별로 다른 조건의 Ra값 기준 표면거칠기를 부여했다. 실험 조건으로는 시편 가공 시 4가지의 회전속도(60, 400, 1200, 2000 rpm), feed rate 150(mm/min) 조건을 선정했다. 2차로는 와이어 용융 방식의 아크 용사 코팅을 실시했다. 코팅 조건으로는 분사 거리 200mm, 전압 24V, 전류 120A, 분사 압력 5bar, 와이어 삽입 속도 30g/mm, 와이어 직경 2mm이다. 용사 코팅 후 FE-SEM으로 표면을 관찰한 결과 모든 시편의 S355 면과 코팅층(아연-알루미늄) 사이에 유격이 발생하지 않고 성공적으로 안착이 되었음을 확인할 수 있었다.

This study aimed to assess and determine the optimal model for predicting the full bloom date of ‘Fuji’ apples across South Korea. We evaluated the performance of four distinct models: the Development Rate Model (DVR)1, DVR2, the Chill Days (CD) model, and a sequentially integrated approach that combined the Dynamic model (DM) and the Growing Degree Hours (GDH) model. The full bloom dates and air temperatures were collected over a three-year period from six orchards located in the major apple production regions of South Korea: Pocheon, Hwaseong, Geochang, Cheongsong, Gunwi, and Chungju. Among these models, the one that combined DM for calculating chilling accumulation and the GDH model for estimating heat accumulation in sequence demonstrated the most accurate predictive performance, in contrast to the CD model that exhibited the lowest predictive precision. Furthermore, the DVR1 model exhibited an underestimation error at orchard located in Hwaseong. It projected a faster progression of the full bloom dates than the actual observations. This area is characterized by minimal diurnal temperature ranges, where the daily minimum temperature is high and the daily maximum temperature is relatively low. Therefore, to achieve a comprehensive prediction of the blooming date of ‘Fuji’ apples across South Korea, it is recommended to integrate a DM model for calculating the necessary chilling accumulation to break dormancy with a GDH model for estimating the requisite heat accumulation for flowering after dormancy release. This results in a combined DM+GDH model recognized as the most effective approach. However, further data collection and evaluation from different regions are needed to further refine its accuracy and applicability.

This study was conducted to determine the optimal irrigation starting point by analyzing tree growth, physiological responses, fruit quality, and productivity in peach orchards. Seven-year-old ‘Kawanakajima Hakuto’ peach trees were used in an experimental field (35°49′30.4″N, 127°01′33.2″E) located within the National Institute of Horticultural and Herbal Science located in Wanju-gun, Jeollabuk-do. The irrigation starting point was set with four levels of –20, –40, –60, and –80 kPa from June to September 2022. While there were no significant differences in increase of trunk cross-section area and leaf area among treatments, shoot length and diameter decreased in the –80 kPa and –20 kPa treatments. The photosynthetic rate measured in August was highest for –60 kPa (17.7 μmol·m-2·s-1), followed by –40 kPa (15.6 μmol·m-2·s-1), –20 kPa (14.5 μmol·m-2·s-1) and –80 kPa (14.0 μmol·m-2·s-1). SPAD value measured in May and August was lower in the –80 kPa and –20 kPa treatments than in the –60 kPa and –40 kPa treatments. The harvest date reached three days earlier in the –20 kPa treatment compared to other treatments. The fruit weight was highest in the –60 kPa (379.1 g), followed by –40 kPa (344.0 g), –80 kPa (321.0 g) and –20 kPa (274.9 g). Firmness was the lowest in the –20 kPa treatment. The soluble solid content was highest in the –60 kPa treatment (13.3°Bx).The ratio of marketable fruits was highest in the –60 kPa treatment (50.7%) and lowest in the –80 kPa treatment (23.4%). In conclusion, we suggest that setting the irrigation starting point at –60 kPa could improve the fruit quality and yield in peach orchards.

Determining the size or area of a plant's leaves is an important factor in predicting plant growth and improving the productivity of indoor farms. In this study, we developed a convolutional neural network (CNN)-based model to accurately predict the length and width of lettuce leaves using photographs of the leaves. A callback function was applied to overcome data limitations and overfitting problems, and K-fold cross-validation was used to improve the generalization ability of the model. In addition, ImageDataGenerator function was used to increase the diversity of training data through data augmentation. To compare model performance, we evaluated pre-trained models such as VGG16, Resnet152, and NASNetMobile. As a result, NASNetMobile showed the highest performance, especially in width prediction, with an R_squared value of 0.9436, and RMSE of 0.5659. In length prediction, the R_squared value was 0.9537, and RMSE of 0.8713. The optimized model adopted the NASNetMobile architecture, the RMSprop optimization tool, the MSE loss functions, and the ELU activation functions. The training time of the model averaged 73 minutes per Epoch, and it took the model an average of 0.29 seconds to process a single lettuce leaf photo. In this study, we developed a CNN-based model to predict the leaf length and leaf width of plants in indoor farms, which is expected to enable rapid and accurate assessment of plant growth status by simply taking images. It is also expected to contribute to increasing the productivity and resource efficiency of farms by taking appropriate agricultural measures such as adjusting nutrient solution in real time.

This study was aimed to determine the changes in CO2 concentration according to the temperatures of daytime and nighttime in the CO2 supplemental greenhouse, and to compare calculated supplementary CO2 concentration during winter and spring cultivation seasons. CO2 concentrations in experimental greenhouses were analyzed by selecting representative days with different average temperatures due to differences in integrated solar radiation at the growth stage of leaf area index (LAI) 2.0 during the winter season of 2022 and 2023 years. The CO2 concentration was 459, 299, 275, and 239 μmol·mol-1, respectively at 1, 2, 3, and 4 p.m. after the CO2 supplementary time (10:00-13:00) under the higher temperature (HT, > 18°C daytime temp. avg. 31.7, 26.8, 23.8, and 22.4°C, respectively), while it was 500, 368, 366, 364 μmol·mol-1, respectively under the lower temperature (LT, < 18°C daytime temp. avg. 22.0, 18.9, 15.0, and 13.7°C, respectively), indicating the CO2 reduction was significantly higher in the HT than that of LT. During the nighttime, the concentration of CO2 gradually increased from 6 p.m. (346 μmol·mol-1) to 3 a.m. (454 μmol·mol-1) in the HT with a rate of 11 μmol·mol-1 per hour (240 tomatoes, leaf area 330m2), while the increase was very lesser under the LT. During the spring season, the CO2 concentration measured just before the start of CO2 fertilization (7:30 a.m.) in the CO2 enrichment greenhouse was 3-4 times higher in the HT (>15°C nighttime temperature avg.) than that of LT (< 15°C nighttime temperature avg.), and the calculated amount of CO2 fertilization on the day was also lower in HT. All the integrated results indicate that CO2 concentrations during the nighttime varies depending on the temperature, and the increased CO2 is a major source of CO2 for photosynthesis after sunrise, and it is necessary to develop a model formula for CO2 supplement considering the nighttime CO2 concentration.

This study was conducted to investigate optimal conditions for cutting propagation of the strawberry cultivar “Sulhyang” through the collection methods of cuttings (runners tips), leaf number of cuttings, and cutting time. Cuttings were collected from the mother plant in the nursery bed (MP) and plants after fruit harvest (HP); the leaf number of cuttings was 0, 1, and 2, and the cutting time was at one-week intervals from June 4 to July 9. The survival rates for MP and HP cuttings were notably high, reaching 99.5% and 98.7%, respectively, but no significant difference was found. The number of roots were higher in MP cuttings, and there was no significant difference in crown and leaf growth. The fruit yields were 419.2 and 428.4 g, for MP and HP cuttings, respectively. The survival rates according to leaf number of cuttings were 98.1% and 98.3% for 1 and 2 remaining leaves, respectively, and remarkably lower at 25.3% for no remaining leaves. The root numbers were 26.0 and 26.3 for 1 and 2 remaining leaves, respectively, compared with 23.5 for no remaining leaves, with no significant differences in crown and leaf growth. The fruit yields were 424.4 and 421.5 g for 1 and 2 remaining leaves, respectively, and 396.7 g for no remaining leaves. The survival rates according to cutting time was over 97.2% in all cutting time without any difference in each treatment. The root, shoot, and crown of the nursery plant before planting showed the best growth in the cuttings on June 4 and 11, resulting in the highest fruit yields of 433.3 and 426.4 g, respectively, with the lowest yields at 384.5 g for cutting time on July 9. Both MP and HP materials proved suitable for strawberry cuttings. The optimal leaf number for cuttings was at least 1, and the optimal cutting time in Gyeongnam area was evaluated as around June 4-11.

A solution combustion process for the synthesis of hollandite (BaAl2Ti6O16) powders is described. SYNROC (synthetic rock) consists of four main titanate phases: perovskite, zirconolite, hollandite and rutile. Hollandite is one of the crystalline host matrices used for the disposal of high-level radioactive wastes because it immobilizes Sr and Lns elements by forming solid solutions. The solution combustion synthesis, which is a self-sustaining oxi-reduction reaction between a nitrate and organic fuel, generates an exothermic reaction and that heat converts the precursors into their corresponding oxide products in air. The process has high energy efficiency, fast heating rates, short reaction times, and high compositional homogeneity. To confirm the combustion synthesis reaction, FT-IR analysis was conducted using glycine with a carboxyl group and an amine as fuel to observe its bonding with metal element in the nitrate. TG-DTA, X-ray diffraction analysis, SEM and EDS were performed to confirm the formed phases and morphology. Powders with an uncontrolled shape were obtained through a general oxide-route process, confirming hollandite powders with micro-sized soft agglomerates consisting of nano-sized primary particles can be prepared using these methods.

The objective of this study was to investigate the distribution of functional compounds in perilla leaves of various genetic resources and their antioxidant activities. A comprehensive analysis of functional compounds was conducted for 90 genetic resources, focusing on total polyphenol content (TPC), total flavonoid content (TFC), individual phenolic content (IPC), and lutein. Their antioxidant activities were then analyzed based on their radical scavenging capacity using ABTS (2,2’-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) and DPPH (2,2-diphenyl-1-picrylhydrazyl). The TPC content exhibited a range of 13.19 to 35.85 mg gallic acid equivalent/g, whereas the TFC content varied from 11.74 to 46.51 mg catechin equivalent/g. Total IPC was detected in a range of 6,310.98 to 40,491.82 μg/g. Lutein was detected at levels between 70.97 and 597.97 μg/g. ABTS and DPPH radical scavenging activities of perilla leaves ranged from 30.39 to 58.58 mg trolox equivalent (TE)/g and from 7.74 to 46.56 mg TE/g, respectively. Furthermore, correlation analysis demonstrated that rosmarinic acid, a phenolic acid, exhibited a significantly positive correlation with antioxidant activity. These findings suggest that various genetic resource of perilla leaves could effectively mediate antioxidant capacity. Results of this study provide valuable information for use of perilla leaves in Korea as functional food materials.

The major active components of Astragalus membranaceus (AM) are isoflavones, which exist in the form of various glycosides. Nuruk is a traditional fermentation starter in Korea, and is suitable for the biotransformation of isoflavone glycosides because it contains various microorganisms and enzymes. This study was performed to determine changes in the isoflavones and antioxidant properties of AM fermented (AF) with nuruk over 24 hours. AF was sampled after 0, 2, 4, 6, 12, 18, and 24 h of fermentation, and calycosin 7-glucoside, ononin, calycosin, and formononetin content, and the antioxidant properties of AF were analyzed. The total phenolic content increased with fermentation time, and the ABTS radical scavenging activity increased until 6 h of fermentation and then decreased. During fermentation, the isoflavone glycosides decreased significantly as fermentation time increased. The contents of calycosin and formononetin, which are aglycons of calycosin-7-glucoside and ononin, increased from 100.54 μg/g to 276.84 μg/g and from 56.29 μg/g to 123.04 μg/g, respectively, at 18 h of fermentation. Significant correlations were observed between fermentation time, isoflavone content, and antioxidant properties. The results of this study showed that fermentation with nuruk is suitable for the biotransformation of isoflavones in AM.

This study aimed to investigate the effects of Centella asiatica powder on the quality characteristics and storage stability of goat meat Tteokgalbi. The goat meat Tteokgalbi samples were prepared with four different quantities (0, 0.5, 1, and 1.5%) of Centella asiatica powder. The moisture content of the samples containing Centella asiatica powder was significantly lower than that of the control (p<0.05). The protein content of the samples containing Centella asiatica powder was significantly higher than that of the control (p<0.05). As the proportion of Centella asiatica powder increased, the pH, L*, and a* values of the uncooked and cooked samples decreased (p<0.05). The water holding capacity and the cooking yield of the uncooked and cooked samples were higher in the samples containing Centella asiatica powder than in the control (p<0.05). Also, with the increasing amounts of the Centella asiatica powder, the thiobarbituric acid reactive substance (TBARS) levels decreased. The 1-1.5% Centella asiatica powder group showed lower levels of volatile basic nitrogen (VBN) than the other groups as per the assay (p<0.05). The results indicate that Centella asiatica powder could enhance the quality characteristics and storage stability of goat meat Tteokgabi.

For appropriate nutrient management and enhanced plant growth, soil sensors which reflect soil nutrient levels are required. Because there is no available sensor for nutrient monitoring, electrical conductivity (EC) sensor can be used to evaluate soil nutrient levels. Soil nutrient management using EC sensors would be possible by understanding the relationship between sensor EC values and soil temperature, moisture, and nutrient content. However, the relationship between soil sensor EC values and plant available nutrients was not investigated. Therefore, the objectives of the study were to evaluate effect of different amount of urea on soil EC monitored by sensors during pepper and broccoli cultivation and to predict the plant available nutrient contents in soil. During the cultivation period, soil was collected periodically for analyzing pH and EC, and the available nutrient contents. The sensor EC value increased as the moisture content increased, and low fertilizer treated soil showed the lowest EC value. Principal component analysis was performed to determine the relationship between sensor EC and available nutrients in soil. Sensor EC showed a strong positive correlation with nitrate nitrogen and available Ca. In addition, sum of available nutrients such as Ca, Mg, K, P, S and N was positively related to the sensor EC values. Therefore, EC sensors in open field can be used to predict plant available nutrient levels for proper management of the soil.

This study aims to analyze the sensory characteristic of cooked rice made with 10 types of rice varieties produced in South Korea using descriptive analysis. Twelve subjects who cook rice regularly, at least once a week, were recruited for the study and participated as descriptive panelists. A total of 29 descriptive terms were developed to describe the sensory characteristics of cooked rice. The findings showed that there were significant differences in the sensory attributes, specifically the appearance and texture attributes among the rice varieties. Among the varieties, Haedeul, Ilpum, and Koshihikari were observed to have a large quantity of rice germ and a relatively yellowish appearance. In contrast, Samkwang, Chamdream, and Yeonghojinmi were relatively whiter than the other samples. Odae, Saecheongmoo, and Sindongjin were rated high in the attributes of roughness, hardness, and chewiness, and Jinsang was rated high in moisture, bitterness, stickiness, and wholeness of rice grains. The differences in the appearance and texture of these rice varieties are likely to significantly impact consumer preferences. Furthermore, the development of customized menus based on the sensory characteristics of the different varieties of rice could encourage consumers to purchase these varieties.

This study was conducted to determine the optimal dipping time and concentration of gibberellin for improving the growth and quality of domestic cultivar 'Seolhyang' strawberry when using runner plants. Strawberry runner plants were collected on November 10th and soaked in GA3 concentrations of 50, 100, and 150 mg·L-1 for 30 and 60 minutes, respectively. After 75 days of planting, the growth results showed that in the 30-minute, 50 mg·L-1 treatment, the crown diameter was thicker and the T/R ratio was lower, indicating better plant vitality. Runner length increased with lower gibberellin concentrations, particularly promoting vegetative growth. Photosynthetic efficiency was more influenced by gibberellin concentration than dipping time, and using concentrations above a certain threshold acted as a stress factor for runner plants, leading to decreased photosynthetic efficiency. For enhancing seedling growth, soaking with 50 mg·L-1 of gibberellin for 30 minutes was found to be optimal. This study verified the effects of gibberellin treatment on strawberry runner plants to improve plant growth and quality, providing useful basic data for using gibberellin.

PURPOSES : This study aims to analyze the causes of pedestrian traffic accidents on community roads. METHODS : This study collected variables affecting pedestrian traffic accidents on community roads based on field surveys and analyzed them using negative binomial regression and zero-inflated negative binomial regression models. RESULTS : Model analysis results showed that the negative binomial regression model is more suitable than the zero-inflation negative binomial regression model. Additionally, the segment length (m), pedestrian volume (persons/15 min), traffic volume (numbers/15 min.), the extent of illegal parking, pedestrian-vehicle conflict ratio, and one-way traffic (one: residential, two: commercial) were found to influence pedestrian traffic accidents on community roads. Model fitness indicators, comparing actual values with predicted values, showed an MPB of 1.54, MAD of 2.57, and RMSE of 7.03. CONCLUSIONS : This study quantified the factors contributing to pedestrian traffic accidents on community roads by considering both static and dynamic elements. Instead of uniformly implementing measures, such as pedestrian priority zones and facility improvements on community roads, developing diverse strategies that consider various dynamic factors should be considered.

이 연구는 잠재프로파일 분석을 활용하여 청소년의 그릿 유형에 따른 진로적응력을 분석하고자 하였다. 한국아동청소년패널조사 2018((KCYPS 2018)의 4차년도 자료를 활용했으며, 전국에 걸쳐 고등학교 1학년에 재 학 중인 2,218명의 학생이 본 연구의 대상이었다. 자료 분석 결과는 다 음과 같다. 첫째, 적합도 지수와 잠재프로파일 분류에 따라 4집단으로 구분되었다. 1집단은 ‘목표·흥미형’, 2집단은 ‘그릿취약형’ 3집단은 ‘단기 노력추구형’ 4집단은 ‘그릿안정형’으로 명명하였다. 둘째, ‘목표·흥미형’, ‘그릿취약형’ ‘그릿안정형’은 진로적응력을 예측하였다. 이러한 연구결과 를 근거로 하여 진로적응력을 예측하는 그릿을 높이기 위한 다각적인 논 의가 활발히 이루어질 필요가 있다.

본 연구는 낙엽활엽수종인 이팝나무(Chionanthus retusa) 잎의 흡음 성능을 평가하고자 하였다. 잎을 주재료로 하여 재료의 크기와 두께 및 건조 조건을 달리해 흡음재를 제조하였다. 중간형 측정관(100㎐-3200㎐)을 적용한 관내법(Impedance tube method)을 이용하여 흡음율을 측정하였 다. 총 18개의 조건에서 측정된 흡음율을 분석하였다. 기건엽 조건(MC 13.92%)과 생엽 조건(MC 162.04%)의 흡음율은 두께가 증가함에 따라 크게 향상되었다. 흡음율은 밀도의 증가에 비례하는데, 부피당 밀도가 더 높은 생엽조건의 두께별 흡음율이 더 높게 나타났다. 생엽 조건군에서 크기 0.5 × 0.5㎠의 2.50㎝ 두께에서 평균 흡음율은 0.643으로 가장 높게 나타났으며, 각 시험처리 조건별 흡음재의 주파수 대역별 흡음율은 기건엽 조건의 두께 1.75㎝에서 가장 높게 나타났다. 이상의 연구결과로부터 전반적인 흡음율은 1000㎐ 이상의 대역에서 더 높은 성능을 보였다.

본 연구는 수분함량과 미생물 첨가제가 알팔파 사일리지의 발효특성과 사료가치에 미치는 영향을 알아보기 위해 수행되었다. 알팔파는 개화 10% 시기에서 수확되었으며 수분 함량(M60, M50, M40 및 M30)이 60, 50, 40 및 30%일 때 각각 이용하였다. 1500g을 샘플링한 후, 증류수 10mL을 첨가한 미첨가구(NAD)와 Lactococcus lactis 와 Pediococcus pentosaceus의 혼합물을 증류수(0.1g/10mL)에 1.5 x 1010cfu/g 농도로 희석한 후 접종한 미생물 균주 첨가구(ADD)를 3개월 및 6개월 발효하였다. 수분 및 미생물 첨가제에 따른 연평균 알팔파의 조단백, 중성세제 불용성 섬유 및 산성세제 불용성 섬유에서 차이가 없었다(p>0.05). 모든 발효기간에서 pH는 ADD의 M40에서 가장 낮았다(p<0.05). NAD 처리구에서는 젖산이 M50에서 가장 높았고(p<0.05), ADD 처리군에서는 M40 시험구에서 젖산이 가장 높았다(p<0.05). NAD와 ADD의 M60은 젖산 함량이 다른 수분 함량들에 비해 가장 낮으며(p<0.05) 낙산이 유일하게 검출되었다. 미생물군집의 상대적 풍부도는 ADD 처리구의 M40과 M50에서 Homo LAB (Enterococcus, Lactiplantibacillus, Lacticaseibacillus, Lactococcus, Pediococcus)의 비율이 가장 높았고, Clostridium은 M60에서 가장 높았다.

Recently, the government has provided support such as entering new markets, expanding sales channels, and supporting manpower, not just in the form of funding, to efficiently and effectively support limited national resources to improve corporate performance. In this study, we tried to find out the effect of government support for companies that have benefited from the Excellent Technology Research Center Project (ATC Project) and the World Class 300 project using propensity score matching. As a result of the analysis, the effect of government support for the ATC project became visible after the appointment period, while the effect of the World Class 300 project was insignificant. This means that when the size of the company is small, the effect of government support is more pronounced. This suggests that in order to maximize the effectiveness of government support, appropriate national policy interventions such as government innovation funding are needed when the size of the company is small. In this study, differences in the timing, performance indicators, and company size of policy support effects were found in the growth stage of a company from a mid- to long-term time series perspective, suggesting that support policies based on this need to be adjusted and redesigned.

최근 대기오염으로 인한 환경오염을 줄이고자 국제 산업계의 노력의 일환으로 국제해사기구(IMO)의 규제 발효등으로 이어지 고 있다. IMO는 EEXI,EEDI,CII 등 선박에서 나오는 대기오염을 줄이기 위해 각종 규제를 발효시키고 선박에서 소모되는 전력을 줄여 에너 지를 절약하는 방안을 추진하고 있다. 선박에서 사용되는 전력의 대부분은 전동기가 차지한다. 선박에 설치된 전동기 중 큰 부하를 차지 하는 기관실 송풍기는 수요와 관계없이 정속운전으로 운전하기 때문에 주파수제어를 통한 에너지절감을 기대할 수 있다. 본 연구에서는 발전기의 과급기에 연소공기를 공급하는 발전기 송풍기의 전동기 주파수를 제어하여 에너지 절감에 대한 실효성을 입증하였다. 송풍기 주파수 입력에 따른 과급기출구 온도의 출력 데이터를 기반으로 시스템을 모델링하고, 과급기 출구온도를 목표값으로 하여 주파수를 제 어하는 PI 제어계를 형성하여 과급기 설계기준 출구온도를 유지하면서 송풍기의 주파수 제어를 통해 연간 15,552kW 전력소모량을 절감하 였다. 송풍기 팬 주파수 제어를 통한 에너지절감액의 유효성은 하계(4월~9월) 및 동계(3월~10월) 기간동안 검증하였으며 이를 토대로 실습 선의 연간 6,091천원의 유류비 절약과 이산화탄소 8.5Ton, SOx 2.4kg, NOx 7.8kg의 대기오염물질 저감을 달성하였다.

기후변화 영향으로 이상고수온, 태풍, 홍수, 가뭄 등 재난 및 안전 관리기술은 지속적으로 고도화를 요구받고 있으며, 특히 해 수면 온도는 한반도 주변에서 발생되는 여름철 적조 발생과 동해안 냉수대 출현, 소멸 등에 영향을 신속하게 분석할 수 있는 중요한 인자 이다. 따라서, 본 연구에서는 해수면 온도 자료를 해양 이상현상 및 연구에 적극 활용되기 위해 통계적 방법과 딥러닝 알고리즘을 적용하 여 예측성능을 평가하였다. 예측에 사용된 해수면 수온자료는 흑산도 조위관측소의 2018년부터 2022년까지 자료이며, 기존 통계적 ARIMA 방법과 Long Short-Term Memory(LSTM), Gated Recurrent Unit(GRU)을 사용하였고, LSTM의 성능을 더욱 향상할 수 있는 Sequence-to-Sequence(s2s) 구조에 Attention 기법을 추가한 Attention Long Short-Term Memory (LSTM)기법을 사용하여 예측 성능 평가를 진행하 였다. 평가 결과 Attention LSTM 모델이 타 모델과 비교하여 더 좋은 성능을 보였으며, Hyper parameter 튜닝을 통해 해수면 수온 성능을 개 선할 수 있었다.