報告書番号

T02032

タイトル（和文）

低煙突を対象としたダウンドラフト時の大気拡散予測手法の開発

タイトル（英文）

Development of a method for predicting atmospheric dispersion of exhaust gas from a low stack during downdraft

概要 （図表や脚注は「報告書全文」に掲載しております）

近年，景観への配慮，コスト削減，航空法の規制などの理由で，煙突の高さを低く計画する火力発電所がある。このような場合，煙突風上の山によるダウンドラフト（下降気流）にともなう大気汚染物質の高濃度が，環境影響評価の予測項目の一つとして重要である。現在，火力発電所の環境影響評価では，電力中央研究所が提案した排ガス拡散数値モデルによる地形影響評価手法TOPADSが標準となっている。本手法は高所，大容量煙源に対して風洞実験や野外トレーサ実験をもとに妥当性の確認をしているが，ダウンドラフト時の大気拡散予測への適用性については検討を行っていない。また，大気拡散数値モデルをダウンドラフトに着目して検証した事例はない。本報告では，野外気象観測とトレーサ実験によりダウンドラフトの実態を把握し，その結果にもとづいて大気拡散予測手法の適用性を検討し，以下の結果を得た。(1)赤城山麓（平均斜面勾配10度）において，ダウンドラフト時の乱流と拡散特性を把握した。平均風速分布と乱流強度分布は地形の影響を受けて，それぞれ高さの1/3乗，-1/4乗に比例する。下降流は，少なくとも地表面から300ｍの高さまで卓越し，煙源高さが100ｍのとき，最大濃度の出現距離は平地と比べて0.3倍以下，最大濃度は1.5～３倍であった。(2)TOPADSは，全体に過大予測する傾向があるが，最大濃度についてみると野外トレーサ実験結果と1.5倍程度の差にすぎない。TOPADSはダウンドラフト時にも比較的よい予測結果を示すといえる。(3)乱流計算結果を用いる替わりに，気象観測結果から乱流量を求め，それらをもとに拡散計算する手法を提案した。本手法は，ケースによりTOPADSの過大予測傾向を改善した。(4)ヨーロッパで汎用的に用いられているMERCUREによる感度解析結果は，乱流モデルで格子系，水平方向の物質拡散係数，気象データの取り扱いを工夫することにより，野外トレーサ実験の再現が可能なことを示した。(5)気象観測から得た乱流量をもとに拡散計算する方法は，高速パソコンで通常２日ほどかかる乱流計算を行う必要がないため，使い勝手が非常によい。

概要 （英文）

In recent years, some thermal power plants have adopted plans to limit the height of stacks for scenic considerations, cost reduction and compliance with aviation regulations. In these cases, it is important to consider the high concentration of air pollutants caused by downdrafts due to mountains located windward of stacks in the environmental impact assessment. Currently, the method of evaluating the topographical effect by means of a numerical model for exhaust gas dispersion proposed by the Central Research Institute of Electric Power Industry (CRIEPI) is adopted as the standard method in the environmental impact assessment of thermal power plants. This method is called TOPADS (Sytem for evaluating TOPographical effect on Atmospheric Dispersion). The validity of TOPADS is confirmed experimentally using a wind tunnel and field tracer experiment for large-capacity elevated sources; however, the applicability of the method for evaluating atmospheric dispersion during downdraft conditions has not been tested. Furthermore, there is no study on a numerical model for atmospheric dispersion which takes into consideration downdrafts. In this study, we examined the applicability of the atmospheric dispersion estimation method for downdrafts through understanding the actual conditions of atmospheric dispersion on the basis of meteorological observations and field tracer experiments. The following results were obtained.(1) Turbulence and dispersion characteristics during a downdraft were studied at the base of Mt. Akagi (average inclination: 10 degrees). The vertical distributions of average wind velocity and intensity of turbulence are affected by topology and are proportional to the 1/3rd and -1/4th power of the height from the ground surface, respectively. The downdraft is predominant up to the height of 300 m from the ground. When the stack is 100 m high, the location and value of the maximum concentration of air pollutants and were 0.3 times closer and 1.5-3 times greater, respectively, than the corresponding values estimated for flat ground. (2) TOPADS tends to overestimate the effect of topology; however, the difference in the maximum concentration of air pollutants between results obtained by the model and those obtained by field tracer experiments is slight. The model overestimated only by a factor of approximately 1.5. TOPADS provides reasonable estimates of atmospheric dispersion for downdraft periods.(3) A method to calculate dispersion in which turbulence statistics are parameterized using the results obtained by meteorological observations was proposed. This approach reduced the overestimation tendency of TOPADS in some simulation cases.(4) Results of sensitivity analysis by MERCURE which is generally used in Europe showed that the turbulence model could reproduce field tracer experimental results if the grid system, the horizontal diffusivity of the pollutant and meteorological data are treated appropriately.(5) Calculation of turbulence by a high-speed personal computer takes two days; however, the parameterization of the turbulence statistics does not require such calculations, thus the parameterization method presented in this report is very easy to use.

報告者

キーワード