Open Access

Die BAM plant Untersuchungen zur unfallbedingten Freisetzung brennbarer Gase in Umschließungen. Anlass ist die prognostizierte Zunahme verflüssigter Gase als Antriebsquelle im Lkw-Schwerlastverkehr aufgrund der strenger werdenden Abgasnormen. Im Hinblick auf die Bewertung von Gefahren zukünftiger Kraftstoffe für die Sicherheit von Infrastrukturen soll als Versuchsszenario verflüssigtes Erdgas (LNG) in einer verkehrstunnelartigen Einhausung auf dem Testgelände Technische Sicherheit der BAM freigesetzt werden. In diesem Beitrag werden die sicherheitstechnischen Vorbetrachtungen geschildert, die den Anlass zur Festlegung auf dieses Szenario gaben sowie die Relevanz einer notwendigen vertieften Betrachtung illustrieren. Dabei wird auf die allgemeinen Entwicklungen im Güterverkehr, die sicherheitstechnischen Eigenschaften von LNG sowie den baulichen Anforderungen an und für die Gefährdungsbetrachtung relevanten Bedingungen in Verkehrstunneln eingegangen. Vereinfacht soll die prinzipielle Herangehensweise bei der Gefährdungsbeurteilung dargestellt werden sowie im Zuge dessen anhand einer überschlägigen Berechnung der Dimension der explosionsfähigen Wolke im Tunnel die Gefährdung für Verkehrsteilnehmer und die Bauwerksstruktur abgeschätzt werden. Hilfsmittel, die zum Einsatz kommen, sind physikalische Modellvorstellungen, empirische sowie CFD-Methoden.

Die BAM untersucht auf ihrem Testgelände TTS die Auswirkungen der Freisetzung bestehender und zukünftig erwarteter, in Fahrzeug zum Einsatz kommender gasförmiger Kraftstoffe in besonders schützenswerten, neuralgischen Verkehrsräumen wie Tunnelbauwerken

In combination with new types of drive technologies, more and more flammable gases could be introduced into the traffic area “road tunnel” in future. If these vehicles have an accident and the gases are subsequently released, there is the possibility that the mentioned substances show a heavy gas behavior due to their storage conditions. From a safety point of view, this scenario has to be assessed with regard to its effects on the tunnel structure and the people involved. The aim of the experimental investigations carried out in this work is to create the basis for such an analysis by investigating the spreading behavior of the gases after they have been released in the tunnel and determining the influence of individual geometric and operational parameters on this behavior. Specifically, this work is based on the scenario of a continuous, momentum-free release of propane within a vaulted road tunnel with active longitudinal ventilation. The tunnel-specific boundary conditions that are relevant to this scenario are first worked out by a theoretical analysis. The basis of the subsequent experimental part is formed by two test rigs built as part of the work, which are similar in their essential features, but differ in scale. In both facilities, the aforementioned release scenario was simulated under various boundary conditions based on an idealized tunnel segment. The dispersion behavior is primarily assessed by the averaged concentration distribution near the ground, which results from a large number of detectors arranged in the experimental area. The detectors operate on the principle of weakening infrared light in the presence of hydrocarbons. Measurements of the flow field properties accompany the concentration measurements. The conception of the test rig on a scale of 1:12 is based on dimensional analysis. The actual release campaign comprises a large number of individual experiments in which selected parameters were varied over the tunnel-relevant range. The predicted heavy gas behavior for propane emerged clearly in the experiments. In cases in which the cloud laterally reaches the tunnel walls, the gases are channeled, which is accompanied by reduced longitudinal dilution. It was possible to identify the release rate and the flow velocity as the factors that have the greatest influence. The former basically increased the concentration, while the latter decreased it. All other tested parameters resulted in more complex propagation situations, which force a differentiated consideration of the influence. Lifting the source from the ground also reduced the concentration. In the case of an eccentrically arranged source and a transverse slope of the roadway, the influence is largely limited to the lateral concentration distribution in the immediate vicinity of the source. Nevertheless, both parameters plus a possible longitudinal slope of the roadway only showed a slight effect on the area far-downstream from the source. In addition, the phenomenon of backlayering, which is known from the spreading of fire smoke in the tunnel, could be demonstrated with a steep longitudinal slope. Obstacles increased the complexity of the situation. While globally the dilution of the cloud is partly invariant to obstacles, locally an increase as well as a decrease in concentration can be observed under certain conditions. For the unobstructed tunnel, it was finally possible to define a dimensionless parameter that describes the curve of the longitudinal dilution on the ground within the heavy gas cloud. From this, a simple graphical nomogram is derived for the continuous release of heavy gases in an unobstructed tunnel environment, which can be used to estimate the concentration in relation of a dimensionless source distance. For exploring the real, undistorted behavior experiments were also carried out in original scale. Due to the similarity of both test rigs in terms of scale, the test results can also be used to check the scalability of the spreading situations. For that, two configurations that have already been examined in small scale were selected. The main limitation for test execution and regarding the scalability comparison was the dependence of the flow conditions within the test rig from external wind conditions which occurred despite of taken countermeasures. The large-scale release was associated with pronounced fog formation. The near-ground spreading corresponding to the heavy gas behavior could be confirmed. However, in detail the spread was far more unsteady. Looking at the time-averaged concentrations, the processes already known from the small-scale test were qualitatively well approximated. Remaining quantitative differences, however, require critical consideration. This discrepancy is more likely a consequence of the experimental compromises have to be made in the specific case. From the observed gas behavior, the development of a surface fire initiated by ignition of the re-leased gases is considered to be the most likely subsequent scenario for the release of heavy, flam-mable gases in tunnel-like enclosures. The thesis concludes with recommendations on the methodological approach to be favored in the future.