圖1 實驗通道的橫截面，顯示緊貼左側“欄桿”的可燃物。（1）固定欄桿，由9.5mm厚的膠合板構成；（2）可移動欄桿（也由膠合板構成）；（3）1.3mm厚的鋁片；（4）石蠟浸過的硬紙板；（5）實驗期間用于固定硬紙板的角鋼。

The fuel

可燃物

As the obiective of these experiments was to examine the behaviour of the flames rather than to model the fire initiation and growth, the fuel was chosen to give established burning in as short a period of time as possible, with sustained burning for at least 1 min. Cardboard (2 mm thick,98 kg/m') soaked in paraffin (c.3g/m) was found to be suitable: the cardboard acted as a stable wick, allowing the liquid to burn off fairly uniformly. Paraffin is perhaps too volatile for this purpose, but other less volatile fuels which were tried caused the cardboard to char and buckle during flaming, which could give irregular behaviour. With paraffin, the card remained in place until the liquid fuel had been largely consumed, thus allowing the behaviour of the flames to be observed.

由于此實驗的目的是研究火焰行為，而不是模擬火焰發生發展的，因此選擇的可燃物要在短時間內進行穩定燃燒，至少持續燃燒1分鐘。浸有石蠟（c.3g/m2）的硬紙板（2mm厚，98kg/m2）是較為合適的可燃物。硬紙板作為穩定的燈芯，可使液體均勻穩定的燃燒。根據此實驗目的，石蠟可能揮發性過強，但是如果使用其他揮發性較差可燃物，會導致燃燒過程中硬紙板炭化和變形，繼而產生不規律的行為。直到液體可燃物幾乎被完全燃盡前，浸有石蠟的紙片將一直保持原位，以便觀察火焰行為。

Sections of paraffin-soaked card, 20 cm in length, were held against the sides and fioor of the channel by lengths of thin-section, mild-steel angle. Ignition was achieved by using a gas torch to set alight pieces of paraffin-soaked string lying along the foot of the "balustrades'. Card on the floor' was ienited by direct fiame contact along its lower edge.Complete involvement of the fuel-bed surfaces was normally achieved within 20-30s, and quasi-steady burning lasted for another 60-75 s.

將長20cm浸有石蠟的紙片，鋪設在通道的側面和底面上，并用角鋼和薄片固定。使用氣體點火器點燃放置在“扶梯”底部浸有石蠟的細繩，實現引燃。“地板”上的紙片被來自下方的火焰直接點燃。通常在20~30s內，可燃物表面將完全處于燃燒狀態，準穩態燃燒可再持續60-75s。

Several series of experiments were carried out, but the most important involved a section of channel which was in the same proportion as the King's Cross escalator. The behaviour of the fames that were observed characterise the so-called 'trench effect' very well. The 20-cm lengths of paraffin-soaked card were located with their lower edges 50 mm from the foot of the channel. Figures 2-7 compare three situations in which only one side ('balustrade') is burning (Figs 2 and 3), (i) in which both sides are burning (Figs 4 and 5), and (ii) in which both sides and the "floor' are burning (Figs 6 and 7). It was noted that when only a section of "floor' was burning, the flames remained in the channel, flowing upwards at an angle of 30, hugging the floor above the burning section: the implications for fiame spread are discussed elsewhere.

雖然進行了多組實驗，但最重要是在與國王十字站扶梯比例相同通道內進行的實驗。所觀察到的火焰行為，很好地表征了所謂的“溝槽效應”。浸有石蠟的長20cm紙片從距離底邊50mm的位置開始鋪設。圖2-7對比了三種情況：（1）只有一側（“欄桿”）燃燒（圖2和3）；（2）兩側都燃燒（圖4和5）；（3）兩側和“地板”都燃燒（圖6和7）。可以看到，當只有“地板”一面燃燒時，火焰被限制在通道內，緊貼著燃燒區域的上部地板，沿30°傾斜面向上蔓延燃燒。此種火焰蔓延情況將在他處論述。

When only one side is burning, the fames are observed to rise vertically (Figs 2 and 3). When both sides are alight, the flames are deflected towards the centre-line of the channel, their tips merging (as shown in Fig. 4), and also towards the slope of the channel, effectively reducing the height of the flames as seen from the side (Fig. 5). When both sides and the floor are alight, the flames interact and are deflected o such an extent that the confinement to the channel is almost complete (Figs 6 and 7).

當只有一側燃燒時，觀察到的火焰是垂直向上（如圖2和圖3）。當兩側都在燃燒時，火焰將偏向通道的中心線，其火焰前鋒發生合并（如圖4），也會偏向通道斜坡，從側面看顯著降低了火焰高度（如圖5）。當兩側和地板都燃燒時，火焰相互作用，并會傾斜到完成限制在通道燃燒的程度（如圖6和圖7）。

The same type of behaviour was observed when the burning surfaces were located near the top (within 200 mm) of the channel, and when the length of the burning fuel bed was reduced from 20 to 5 cm. The effect of channel width was also investigated, albeit superficially. It was more pronounced for 'a 200-mm-wide channel, but there was still some fame deflection when one side was completely removed (infinite width), provided the floor was alight.

當燃燒表面位于通道頂部200mm附近，且當可燃物長度由20cm減至5cm是，觀察到了相同火焰行為。同時，研究了通道寬度對火焰特征的影響，但研究不夠深入。通道寬為200mm時，現象更為明顯，但只要地板被引燃后，即使是一側墻被完全拆除后，有些火焰仍會出現偏向通道的情況。

The scale experiments described in this paper clearly reveal a type of flame behaviour which had not previously been identified. If the internal surfaces of an inclined channel are burning, the flames are deflected into the channel if the angle of inclination is 30 ° . The effect is greatest if the floor is burning, but there is evidence that it will occur to a significant extent if only the sides are involved. The effect is enhanced in narrow channels, although the limits have not been ascertained. No attempt was made to model either the confining shaft or the combustible side wall that were present in the King's Cross fire. The fact that the channel itself was capable of confining the flames and keeping them below the 'handrail' level is consistent in a qualitative sense with the dynamics of natural diffusion flames, which satisfy their demand for oxygen by entraining air from the surrounding atmosphere.

本文所述的1：3比例實驗清晰地揭示了一種以前未發現的火焰行為。如果通道傾斜30°，通道內側發生燃燒，火焰將向通道傾斜。雖然地板燃燒時，現象最為明顯，但也有證據表明，只要一側壁面燃燒，此現象也會出現。雖然邊界條件尚未明確，但狹窄通道將強化此種現象。實驗沒有模擬國王十字站火災中的豎井和易燃側面。通道自身能夠限制火焰并將其保持在“扶手”水平高度以下，這一事實在本質上與自然擴散火焰動力學是一致的，后者通過從周圍大氣中吸入空氣來滿足其對氧氣的需求。

Moreover, such behaviour can account for two critical features of the original fire-why it was not perceived much earlier that a life threatening situation was developing, and why the fire spread so rapidly into the booking hall. The former is consistent with flame deflection into the escalator channel (cf. Fig. 7): none of the survivors reported seeing "large flames' before the fire erupted into the ticket hall-not even Fireman Bell, who descended an adjacent escalator (No. 6) only 2 or 3 min previously.5 The rapid spread of fiame up the escalator is consistent with high rates of heat transfer generated by fiames and hot gases deflected upwards between the wooden balustrades. This hypothesis was subsequently confirmed in the t-scale tests carried out at the Health and Safety Executive Laboratories at Buxton (Derbyshire, UK).

此外，這種火焰行為可以解釋火災初期的兩個關鍵問題，即：為什么人們沒有在初期火災發展到危及人的生命之前，更早地發現火災？為什么火災會如此迅速地蔓延至售票大廳？第一個問題與實驗中火焰向自動扶梯通道內傾斜的研究結果一致（如圖7）：在火燒到售票大廳前，沒有一名幸存者表示看到過“大火焰”，甚至消防員Bell也沒有發現，他在火災前2～3分鐘乘相鄰的（6號）自動扶梯下來。實驗中在木質壁面之間，向上的火焰和熱煙氣造成的快速熱量傳遞與實際火災中火焰在自動扶梯上的迅速蔓延是一致的。隨后在Buxton（英國德比郡）的健康與安全實驗室（Health and Safety Executive Laboratories）進行的1：3比例的實驗證實了這種假設。

It is likely that this behaviour will not be unique to escalators and may be observed in other inclined combustible surfaces with an appropriate degree of confinement.' A better understanding of the controlling factors is required so that other configurations which are capable of rapid fire growth can be identified. Such work is currently in hand at Edinburgh University.

此種火焰行為并非只自動扶梯處發生，當限制條件合適時，在其它切斜可燃平面也可能發生。這就需要更深入的了解控制因素，以便可以識別出其它可能造成火勢快速結構。目前愛丁堡大學正在進行此項研究。

1. Simcox, S, Wilkes, N. S. & Jones, I. P, Fire at King's Cross Underground Station, 18th November 1987: Numerical simulation of the buoyant flow and heat transfer. Report No. AERE-G 4677, Harwell Laboratory, Harwell, UK, 1988.

2. Markstein, G. H. & De Ris, J., Upward fire spread over textiles, l~h Symposium (International) on Combustion. The Combustion Institute, Pittsburgh, PA, USA, pp. 1085-97.

3. Drysdale, D. D, Introduction to Fire Dynamics. John Wiley & Sons, Chichester, UK, 1985.

4. Drysdale, D. D. & Macmillan, A. J. R. Flame spread on inclined surfaces. Fire Safety J. (in press).

5. Fennell, D, Investigation into the King's Cross Underground Fire. HMSO, London, UK, 1988, Ch. 12.

6. Moodie, K. & Jagger, S. F., The King's Cross Fire: Results and analysis from the scale model tests. Fire Safety J, 18(1) (1992) 83-103.