A perspective from Neil Fricker, former IFRF Director and Investigator
It seems no time at all since I kicked-off this series with an article about some of the work undertaken by IFRF during its ‘Livorno period’ in the 2010s, when I was briefly the IFRF Director. My main link to IFRF is, however, the late-1960s when I worked for almost four years as a Research Investigator, primarily engaged in running the Furnace Trials on the then new fuel (at least for Europe), natural gas, as well as taking responsibility for activities relating to some of the IFRF aerodynamic studies on swirling flows. Like all of the other contributors to the ‘Reigniting the…’ series of MNM articles, this period as a young researcher at IFRF was extremely enriching, not only for the technical insights gained by close proximity to so many 2 MW flames (!), but also for the career-long network of colleagues and friends that it established. In 1968, fresh out of university, I joined the new investigation team set up by Wolfgang Leuckel with co-contributor Klaus Hein and Wolfgang himself, forming the link to earlier teams and earlier work.
From its foundation in 1948 and throughout the 1950s, the various IFRF research teams had been very much a Dutch-Anglo-French affair, although with the formation of both US and Belgian ‘national flame committees’ during the 50s, the collaborative base of IFRF was beginning to expand. While there had been much discussion between IFRF and various groups in the Federal Republic of Germany throughout the 1950s, it was not until the early 1960s that this developed traction – mainly due to the German coal industry becoming increasingly interested in IFRF’s work on pulverised coal combustion – with the German Flame Research Committee formed in 1962. Eager to be fully engaged in IFRF’s research programme, the time was right to add German investigators to the IJmuiden team, the first being Klaus Hemsath (1963), followed by Wolfgang Leuckel (1964), Klaus Hein (1966) and Ludger van Heyden (1969). Following the forming of the Italian Committee in 1965, two Italian Investigators joined the IJmuiden team in 1968 – Ambrogio Milani and Claudio Cozzi. IFRF was becoming truly ‘international’ at last. In the decade to follow, investigators from Australia, India and Japan further widened the ‘geography’ of the IFRF team.
The early-1960s saw the introduction of ‘flame stability’ as a research topic, with studies on pressure jet atomised flames. Flame stability had not been an issue with the high air temperatures used for open-hearth furnace studies in the 1950s, with all of those experiments concerned with oil-fired flames utilising steam or air blast-atomisation.
After some parametric studies on the effectiveness of bluff body flame stabilisers, swirled combustion air was introduced into IFRF research for the first time during the third pressure jet oil performance trial (‘PT-XII’, conducted in 1962 by János Béer and Norman Chigier). Some key reports on this work are available from the IFRF archive:
- Static pressure distribution in the wake of a stabiliser disc (Chigier), IFRF doc. no. G 2/a/7, 1963
- The flow region near the nozzle in double concentric jets (Chigier, Béer), IFRF doc. no. G 2/a/6, 1963
- Swirling jet flames issuing from an annular burner (Chigier, Béer), IFRF doc. no. K 20/a/9, 1963
- Report on the third pressure jet oil performance trial (Béer), IFRF doc. no. F 31/a/39, 1965
The 1960s also saw extensive aerodynamic studies on swirl and comparative studies on the effectiveness of swirl generators. The usual swirl vanes fitted to the periphery of bluff body stabilisers in those days were replaced with full-flow swirler vanes achieving higher swirl numbers, then with tangential air inlets to the burner, allowing swirl to be varied without removing the burner from the furnace. This work culminated with the development of the continuously variable ‘moving block’ swirl generator by Wolfgang Leuckel in the mid-1960s. It also spawned the ‘swirl meter’, which allows the total angular momentum of a swirling flow to be determined with just one measurement (rather than from the integration of point measurements of profiles of axial and tangential velocity/density). Please take a look at the following IFRF reports available from the archive:
- Swirl intensities, swirl types and energy losses of different swirl generating devices (Leuckel), IFRF doc. no. G 2/a/16, 1966
- Further experimental results and comparisons with theory for the aerodynamic characteristics of swirl generators (Leuckel), IFRF doc. no. G 2/a/16 (text) and G 2/a/16/1 (figures), 1967
This increasing activity on the aerodynamics of combustion was assisted greatly with the opening of a new cold model laboratory – referred to as the ‘Aerodynamics Lab’ – near the office building.
A key step forward made during this period was the definition of the universal mixing and combustion parameters: ‘aerodynamic mixing factor’; ‘stoichiometric mixing factor’; and ‘degree of oxidation’. These parameters could be calculated directly from in-flame concentration measurements, allowing comparative punctual values of mixing and combustion to be made for flames firing different fuels and at any excess air levels. The aerodynamic mixing factor is also directly comparable with tracer measurements made in flames or measurements made on physical models. Further information concerning these parameters can be found in the following IFRF report:
- Mixing factors and degree of oxidation – definitions and formulas for computation (Hemsath), IFRF doc. no. G 0/a/1/1, 1964
Another important feature of this period was the decision to record the advances made in the 1950s within an IFRF book on ‘Industrial Flames’. Two ‘volumes’ were written: One (in English) on ‘Industrial Scale Flame Measurement Techniques’ was published as a book by Jacques Chédaille and Yves Braud (Edward Arnold (Publishers) Ltd, London, 1972), while the other, on ‘Confined Flame Aerodynamics’, was written in French by Professor Roger Curtet of the University of Grenoble. Prof. Curtet had earlier developed an analytical solution to forecast external recirculation in double-concentric confined flows, giving rise to the ‘Craya-Curtet number’, from which the well-known ‘Thring Newby similarity criterion’ emerges as a particular solution – more on this when we reignite the 1950s! Although it was translated into English by Norman Chigier and me in the early 1970s, as far as I know this excellent account of the fundamentals of combustion aerodynamics was never published as an IFRF Report. I probably still have a copy buried in my filing system somewhere… translating it gave me a wonderful grounding in the topic!
The book by Jacques Chédaille and Yves Braud remains essential reading for those wishing to understand the basic principles of in-flame measurement techniques. It is out of print now, but IFRF has published it again as a soft-back A4 report:
- Measurements in industrial flames (Chédaille, Braud), IFRF doc. no. K 20/a/38, 1966
I arrived on the scene in January 1968, and after acquiring the necessary skills and insights while working as a team leader on the ‘O-17A’ oil burner trials, took the role of Trial Leader (known by the Investigators as the “team leaders’ team leader”) for the subsequent natural gas trials ‘NG-1’ and ‘NG-2’.
Natural gas was a new fuel in Europe in the 1960s. IFRF saw its availability as an opportunity to understand the near-field flow and mixing in swirling coal flames – measurements and visual observations are quite difficult in the region near the burner quarl in such flames. Consequently, the decision was made to base the NG-1 trials on the swirling coal burner used earlier for the ‘C-13’ coal trials. However, natural gas was already increasingly becoming a fuel of interest to IFRF Members in its own right, and this was also the period when computational fluid dynamics (CFD) modelling was emerging as a tool for industrial flame simulation. I can vividly recall two interesting meetings with IFRF Member representatives before we commenced the NG-1 tests: Firstly, a well-known burner manufacturer informed us that the C-13 burner would not give a stable flame on natural gas because of the absence of a flame stabiliser disc; and, secondly, a CFD pioneer announced that no further IJmuiden tests were needed because CFD could replace experiments…!
The then Head of Station, Wolfgang Leuckel, listened carefully to these views and, after each meeting, told me to carry on with the NG-1 trails using the C-13 burner with no flame stabiliser and to make detailed measurements. In the event, the NG-1 and NG-2 trials totally justified his opinions. They proved that natural gas could, indeed, be burned in swirling combustion air without a bluff body flame stabiliser. They also provided the insights needed to identify and describe four distinct flame types (‘Types 1, 2, 3 and 4’ – no prizes for originality in the names, but then Klaus Hein and I were working seven day per week, 12-hour shifts at the IJmuiden furnace to identify them!), of which ‘Type 1’ and ‘Type 2’ became the classical flame types used in subsequent years to design low-NOx oil, coal and gas flames (at the time we described these flame types, NOx emission concerns were not even on the horizon). Further information on this work on flame types and flow patterns can be obtained from the IFRF archive:
- A study of swirl stabilised gas flames (Fricker), IFRF doc. no. F 35/a/2, 1969
In the intervening 50 years, CFD has made great strides forward. It is now an extremely powerful design tool for combustion systems, but, in my opinion, the need for experiments and validation of CFD codes remains even to this day. In the meantime, in 1968, empirical rules for optimising burner designs to create stable Type 1 and Type 2 flames were already available (!) – see:
- Flow and mixing patterns in gas flames with swirl in the annular air stream (Leuckel, Fricker), IFRF doc. no. G 2/a/18/2, 1968
Two further ‘firsts’ – based on the NG-1 and NG-2 work – occurred during my time at IJmuiden:
Firstly, the C-13 based gas burner was taken off the IFRF station and tested on a vertical, water-cooled cylindrical furnace at Delft University of Technology. What had been extremely well-behaved flames on the square cross-sectioned IJmuiden furnace became, at times, totally violent flames in the Delft cylindrical furnace. I concluded that, for the high external entrainment appetite of swirling flames, the availability of ‘corners’ in the IJmuiden furnace was a key factor in explaining these differences. This work is reported in the following reports in the IFRF archive:
- Swirl stabilisation of high jet momentum natural gas flames in narrow cylindrical furnace: Volume 1 – text; Volume 2 – figures (Leuckel, Fricker), IFRF doc. no. F 35/a/4/2, 1971
- An investigation of the behaviour of swirling jet flames in a narrow cylindrical furnace (Wu, Fricker), IFRF doc. no. K 20/a/61, 1972
Secondly, many industrial furnaces in the metals and petrochemical industries used multiple burners. It was believed that applying swirl to such systems would result in unstable flow patterns and flame movement caused by interference of neighbouring flames. In the first multiple burner trials (‘MJ-1’, conducted in 1971), interactions between nine swirl burners arranged in various configurations within a 3×3 diamond matrix were tested, and it was shown that the swirling flames remained well-behaved even when in close proximity. In subsequent work after leaving IFRF, I was able to understand that the instabilities observed on petrochemical furnaces were a result of the 2-D (wide and narrow) firebox shapes, rather than solely the effect of swirl. For more information, please see:
- Investigations into the combustion of natural gas in multiple burner systems (van Heyden, Michelfelder, Fricker), IFRF doc. no. F 35/a/5, 1970
The late-1960s also saw an extension of the swirler studies started by János Béer’s team and continued by Wolfgang Leuckel in Jacques Chédaille’s team. A small triumph of mechanical engineering design was Wolfgang Leuckel’s development of the three-channel moving block swirl generator. This device allowed the pattern of angular momentum on three concentric, annular air ducts to be varied at will. Cold flow tests with this system were used to investigate radial turbulent exchange of mass and momentum within solid-body, free-vortex and beyond-free-vortex swirling flows. The results confirmed the ability of a solid body rotation to suppress mixing, and a beyond-free-vortex flow to accelerate mixing when compared to a free-vortex (neutral) or a non-rotating flow. The results were quantified as turbulent exchange coefficients – see the following report for more details:
- Further experiments on turbulent exchange in swirling flows through an annular channel (Leuckel, Fricker), IFRF doc. no. G 2/a/19/2, 1969
Finally, the concept of ‘unmixedness’ was added to the mixing and oxidation factors developed earlier by Klaus Hemsath (see above). This allowed a normalised value of this parameter to be calculated from the difference between mixing factor (an indication of macro-mixing) and degree of oxidation (an indication of micro-mixing), thereby providing a coarse indication of the local turbulence from time average measurements of flame compositions.
Like all my fellow contributors to this ‘reigniting the…’ series of blogs, my three years at IJmuiden were to form the solid basis for the rest of my professional career in the steel and gas industries, in addition to being the source of lifelong friendships.
Vive the IFRF!
[Ed. – thanks for this your second excellent contribution to our series, Neil. What an innovative decade the 60s proved to be for IFRF – it must have been very exciting to have been a part of it!]