Flash Furnace - Sensors
Dr.Nick.Sokolov CRD
PhD(Wavelets).MS.BE.(ME,EE), Owner CRD P/L, Father of the Australian Worsted Comb Fibre processing Technology
Instrumenting the Flash Furnace
Monitoring the structure integrity of the furnace wall.
Under this topic the discussion will be around monitoring the furnace containment walls, structural support, spring strategy and analysis ( including background in operation ) of the Flash Furnace. The graphics above shows a side view of the Flash Furnace and structural control springs used to vary the reaction pressure on the furnace wall.
As the copper smelting environment is full of radio frequency interference (RFI), the idea was to instrument the furnace structure with a wireless sensor system. The problem was 2 fold; RF interference and placement of the sensors. In 2007-8 I have carried out radio frequency sweeps of the smelter finding a vacant RF band width that could be utilized for that purpose.
Benefits to the Smelter
The structural monitoring would improve the control over the furnace wall integrity and extend the time between furnace rebuild cycles. In each smelter shutdown the production stops - if rebuild time can be delayed through good data and its implementation of spring strategy, we can improve the return in the smelting process.
Flash Furnace Background
Flash furnace smelting - Flash furnace is used to produce Matte. Matte is molten Cu-Fe-S solution ( perhaps easier to refer to as a mixture ) with a slag Fe-O-SiO2 from dried concentrate powder. The powder is blown into a flash furnace where Fe and S undergo partial oxidation generating heat. The result is ~65% Cu and Cu barren slag ~1%Cu.[1]
The Flash furnace gets its name from a very efficient way of reacting the powder as it falls to the bottom of the furnace. Under gravity Matte is separated from the slag. Slag is skimmed and Matte is tapped producing metallic ~65% Cu - the first stage in recovery process of Cu and associated metals ( such as Au, Ag, Pt.. depending on the ore body. That will be covered in another post as most products of the Flash furnace are used, for instance slag is used to produce Ni from rich Ni-Fe-S concentrate or powder ).
The structure of the Flash furnace
The Flash Furnace is an unusual shape for a furnace - it is rectangular with a convex floor, cooling elements, cylindrical reaction shaft and cylindrical offgas shaft ( uptake shaft ). It also has observation and tapping portals. The entire structure is supported on columns known as "buckstays" as shown in next figure. ( The structure is green but for sake of clarity not all the elements were shown ).
The top of the flash furnace also supports the hangers for the furnace bricks and structure for the reaction and uptake shaft, non of which is shown to reduce the confusion and only concentrate on important elements of the structure. The top of the furnace or the "crown" is controlled by 48 control rods and same number of associated springs. In reality there are only 6 controlling forces. Here is picture of end crown control rods and springs.
In each corner of the Flash Furnace there are eight edge controlling buckstays that hold the "crown" in place. Each side of the Flash Furnace is independently controlled by its own set of springs usually referred to North- South and East -West.
The entire structure of the Flash Furnace is located on concrete plinths topped by steel pates. The plates are lubricated to provide movement with thermal expansion during the life of the Flash Furnace operation.
There are many interesting problems in Flash furnace operation and they are : structural integrity, operational strategy, cooling - heating, concentrate injection, energy optimization and chemical balance matrix to produce Matte.
This publication is devoted to "buckstays"; control elements ( springs ) ,support slides ( bearing surface ) with the objective to keep the walls from collapsing or "restraining" the thermal "walk". The first graphic shows Flash Furnace and indicates the location of buckstays and springs of only one side of the furnace. The springs attached to buckstays control the walls of the furnace to stay vertical. The spring adjustment and maintenance is a major part of a "spring strategy" to maintain the furnace structural integrity.
The furnace sits on large steel bearing pads and is free to move. As the result of thermal operation of the furnace the furnace buckstays will move. The aim is to keep the buckstays vertical ( z - direction ). Undesirable motion would be tilting ( rotation Rx Ry) at each end of individual buckstays, however buckstays are made to translate ( x,y ) along the level bearing floor surface. Springs are used to control the tilting of the walls through buckstays. Each wall of the furnace is independent and retained by the buckstays. A breach in the wall would result in slag and matte spilling from a great height, endangering people and the operation of the smelter. The Flash furnace is usually located at the highest point in a smelter allowing for gravity feed to electric and anode furnaces. ( I will cover those in separate publications ).
What was done..
Each Furnace floor has different RF noise characteristics. RF noise is dependent on many factors. The evidence has always been there. Take an ordinary site communication device near the furnace, the signal degrades, due to RF interference . There are also other competing sources of RF, some equipment made and some as by-products of the physics of smelting. Bottom line, each smelter needs to have its own audit of RF sources - something I have not seen in my visits to various smelters.
Final decision is that all communication from the sensors must be digital in nature preferably in a packet form as used. Digital sensors are less impervious to RF interference but still require clear band width to operate.
RF sweep and equipment required will be a subject of another post related to Flash smelting operation. Another topic will be a Matlab driven simulation of an Excel sheet strategy failure. In that Topic Matlab will provide furnace wall simulation and Excel sheet provides spring correction strategy - in each simulation depending on the starting points the furnace had a structural failure.[2,4]
What is the Problem ...
Each buckstay moves from vertical alignment with the crown. Here is the picture of North- East corner of the Flash Furnace buck stay alignment - it is clear that the buckstays are no longer vertical
One can see a clear gap developing between the furnace wall and its support structure - clearly a safety issue. One can also see the gap in the bearing separation in the plinth support of the Flush Furnace.
In the case of the Flash Furnace bearing plinths one needs to see clear distribution of load through pressure of the bearing. Clearly this is not the case.
How did operation of Flash Furnace arrive at this point.
In 2006 an engineering report and a strategy was commissioned by the smelter engineers [2]. This was to analyse and check for the calculations used in the Excel spread sheet used to determine the spring compression and hence the structural integrity of the furnace.
"Simplified Spring Equilibrium Calculations for the Flash Furnace" made assumptions about the structural operation of the buckstays. Biggest assumption being that there is no rotation and moments experienced by the buckstay elements. It is evident from number of years of operation that the assumptions made are incomplete and that is clearly demonstrated by Buckstays moving away from vertical alignment as shown in previous photo.
The ability to test the structural equilibrium of the model a set of starting points were given to an excel sheet [3] and were looped through a Matlab program that kept varying the spring behaviour. In each case the Excel strategy using VB programs failed to keep the Buckstays upright resulting in potential instability of the Flash Furnace walls.
This is my 2008 suggestion during a meeting on how the analysis should be approached. Flush Furnace structural model must be improved similar to the following picture, rather then simplification presented in 2006 [2].
The structural analysis sketch suggests that one can not neglect moments as proposed by Marx G.F [2]. The moments and rotation of the walls of the Flash Furnace are significant as shown in previous photographs. The Flash Furnace is not a simple structure but a dynamic machine requiring serious consideration.
CONCLUSION:
1. The Flash Furnace structural integrity model needs to be revisited for a detail analysis using the "crown" of the furnace as a reference point.
2. Any model of Flash Furnace and spring strategy must depend on a reliable feedback from a sensor.
3. The sensors are to be installed on columns to measured the departure by vertical position. Rotation on [Rx with constant z vector ]. Tilt sensors are now cheap.
4. That sensors are integrated with a micro processor and digital signal. The signal is to be communicated using a vacant smelter RF interface channel.
5. The vertical data is to be integrated into a smelter control center and maintance adjustments sent out in a simplified form rather then current practice.
RECOMMENDATION.
Following recommendations are made:
1. remove existing Excel based spring strategy
2. Instrument the vertical columns known as buckstays running tilt sensors.
3. Wireless communication to the smelter operations centre.
4. New model of spring strategy based on Flash Furnace crown
5. Computer generated spring strategy based on nut rotation resulting in spring compression rather then spring compression length.
6. RF audit of the smelter.
References.
1. Devenport W.G. et.al "Flash Furnace Smelting: analysis, control and optimization" 2ed. ISBN 0-87339-577-8. ( copy was purchased for BHPB ) 2003
2. Marx G.F. "Simplified Spring Equilibrium Calculations for the Olympic Dam Flash Furnace" 15 June 2006
3. McKague, A, Norman, G, 1984. Operation of Falconbridge's new smelting process, CIM Bulletin, June 1984, 77:86-92.
4. Excel "Spring Strategy" Author unknown. Currently used by BHPB to carry out spring check on Flash Furnace.
***The Tilt sensors and the method of data collection, wireless digital communication and software is subject to a copyright application.
Reliability Engineering Consultant, M.Sc.
7 个月Great publication. Thanks for sharing!
PhD(Wavelets).MS.BE.(ME,EE), Owner CRD P/L, Father of the Australian Worsted Comb Fibre processing Technology
9 年Local Radio Frequency (RF) audit must a serious concern if one is choosing the "internet of things" which is wireless or radio frequency allocation. RF in sub meter and millimeter wavelength interacts to produce harmonics and interference by interacting with various objects in local area. Basic experience of not working properly is wifi signal loss in house operating a microwave food heater. Source of microwave oven is usually not constructed to the standard of communication and has leakages in primary band of 2.4Gh. ONE NEEDS A LOCAL RF AUDIT not having one will result in slow wifi connections. Note wifi operated near 2.4Gh or 5Gh :-).