8.0 Test Systems for Steam Traps
8.1 Introduction
The following factors are particularly important when testing the operation of steam traps:
- Live steam leakage
- Condensate back-up
8.2 Live steam leakage
Testing steam traps for loss of live steam by measuring the temperature at leaking traps is not as easy as it is often assumed to be. Unfortunately, a temperature measurement alone does not provide any indication of whether a trap passes steam or condensate at saturation temperature. With the exception of thermostatic traps, the temperature of the condensate directly upstream of the trap is always the same as the saturated steam temperature. If a float, inverted bucket or thermodynamic trap is used, the condensate temperature upstream of the trap is exactly identical to the saturation temperature of 10 bar saturated steam, in other words 180 ºC, for a heat exchanger with 10 bar live steam.
Temperature measurements on thermostatic steam traps are somewhat more complicated owing to their cyclic method of operation. A value close to saturation temperature is measured on the condensate side (return) immediately after the trap shuts. Slight subcooling can be determined as soon as it opens. If the back pressure in the condensate system is 5 bar, the saturation temperature is 152 ºC. If only steam (with a pressure of less than 30 bar) is allowed to pass by the trap, it can superheat in the condensate system. On the other hand, if just 4 % condensate is present in the steam, the amount of heat required to revaporise this condensate is so great that no further superheating is measured downstream of the trap.
8.3 Condensate back-up
If condensate backs up in a heat exchanger, part of the heat exchange surface is flooded. The area available for heat transfer is reduced and the condensate temperature upstream of the trap falls below the saturation temperature. If no condensate is discharged at all, the temperature drops to that of the product being heated. If condensate is discharged at less than saturation temperature, we refer to it as "sub-cooled" condensate.
Backing-up can occur in any type of steam trap as a result of:
- Pressure differential across the trap too small
- Strainer screens clogged
- Insufficient trap capacity
In a float-type steam trap, backing-up can also be caused by the ball float filling with condensate due to a leak, so that the valve remains closed. Inverted bucket traps are prone to backing-up if the bleed hole is obstructed. In a thermodynamic trap, the problem may arise if there is air above the valve disc. Thermostatic traps can be affected if the cooling leg is too short or the trap is incorrectly adjusted. Condensate back-up can be determined by measuring the temperature. Since there are not normally any test points upstream of the trap, the measurement is usually based on the surface temperature. These measurements have a tendency to be unreliable and errors are not uncommon. To restrict them to a minimum, it is a good idea to measure not only the surface temperature upstream of the trap but, if possible, also that of the steam pipe directly downstream of the steam control valve. By comparing the two measurements, you can then establish whether or not condensate has backed up upstream of the trap. If you use an infrared thermometer, the two test points must be made of the same material because measurements on stainless steel cannot be compared with the results for carbon steel, for instance.
If no condensate has backed up, the temperature upstream of a float, inverted bucket or thermodynamic trap is identical to the saturated steam temperature (with a tolerance of approximately 5 °C). Backing-up occurs if the temperature upstream of the steam trap is less than that downstream of the steam control valve and simultaneously less than the saturation temperature.
With thermostatic traps (bimetallic or balanced pressure), the temperature upstream of the trap varies from approximately saturation temperature on closing to about 15 Kelvin below saturation at the moment of opening (bimetallic). The temperature difference for a balanced pressure steam trap should be a maximum of 5 Kelvin. If the difference exceeds this value, we speak of "condensate back-up".
It should be noted that the sub-cooling degree of bimetallic traps increases with the back pressure in the condensate pipe. If a bimetallic trap is set to higher sub-cooling or a balanced pressure trap fitted with a sub-cooling capsule, the temperature profile will be completely different, of course.
8.4 Test methods
The following methods are used to test or monitor a steam trap's operation:
- Visual inspection
- Noise measurement / ultrasonic testing
- Conductivity measurement
- Calorimetric testing
Visual inspection
The operating behaviour of a trap that discharges into the atmosphere can usually be tested relatively easily by carrying out a visual inspection. Apart from float traps, all other trap types work intermittently, in other words they remain open for a while and are subsequently closed for a time. When a trap with a cyclic method of operation opens, the condensate / flash steam mixture is discharged. The trap shuts at the instant that hot condensate reaches a few degrees below saturation temperature. If the trap does not have a tight seal, a steam cloud will be visible at the outlet, giving the impression that steam is leaking. In this case, the trap must be replaced or cleaned. With the exception of float-type steam traps, almost all traps that drain into the atmosphere can be monitored very effectively in this way.
If the trap drains into a closed system, a drain device must be installed downstream of the trap to enable it to be tested in the manner described above. The drain valve opens and the condensate is intermittently discharged into the atmosphere. With an inverted bucket steam trap, the method of operation can change to continuous drainage, for example if the load is too low. A thermodynamic trap that opens and shuts more than five times a minute is not working correctly and is passing steam. A method for determining steam losses is described at the end of this chapter. Note that all live steam leakage of just 1 kg/h represents a financial loss of €200.- over the course of a year.
Visual inspections are certainly not the most reliable way to assess a steam trap's operation. In many cases, especially when pipes are drained into the atmosphere, flashing is incorrectly interpreted. Every time the trap opens, the flash steam appears at the outlet as a steam cloud. This condensate, which is flashed to atmospheric pressure and subsequently revaporised, are erroneously thought to be leaked live steam. Steam traps are often set to a higher sub-cooling degree in order to reduce flashing. This is not recommended, however, because sub-cooling the condensate excessively can lead to water hammer damage.
Sight glasses
Sight glasses for observing fluid flow should normally be installed upstream of the steam trap. If the trap is operating correctly, the amount of condensate will be indicated by the water level in the sight glass (Fig. 8-1 a). If backing-up occurs inside the trap, the sight glass will be completely filled with condensate (Fig. 8-1 b). In a trap that is leaking live steam, the condensate level will be forced below the level of the sight glass inlet lug (Fig. 8-1 c).
Fig. 8-1 a: Normal operation
Fig. 8-1 b: Condensate back-up
Fig. 8-1 c: Live steam leakage
If the sight glass is mounted downstream of the trap on the condensate side, it is not always easy to distinguish between live steam loss and flashing. In both cases, a mixture of steam and condensate will be visible in the viewing window. The principal disadvantage of sight glasses and water level indicators is that they constitute an additional source of leakage and that the window may become clouded after a long period in operation or under the influence of chemicals in the condensate. The arrangement of the steam traps is another problem. They are normally installed at the lowest point or possibly on a pipe bridge, which means the sight glasses are comparatively inaccessible and difficult to observe.
Noise measurement / ultrasonic testing
Ultrasonic test devices measure the high-frequency vibrations produced by steam, gases or liquids flowing through the traps using a contact transducer. These vibrations have a frequency that is inaudible to the human ear. Fig. 8-3 shows a simple ultrasonic testing device.
This detector allows the opening and closing movement of an intermittently operating trap to be precisely reconstructed. With continuously operating float traps, the sound level can be taken as a benchmark. The assessment of the sound level is subjective and a certain level of experience is essential.
The trap can be assumed to be operating correctly if the ultrasonic signal is close to a limit value and modulates slightly. The limit value varies according to the steam pressure and the trap type (Fig. 8-2).
Fig. 8-2: Limit value curve
If the limit value is exceeded, the trap is probably leaking steam. The measured value is influenced by ambient noise and extraneous sound sources (steam pressure reducing valves, control valves, steam turbines). Ambient noise can usually be filtered out, except in the immediate vicinity of the trap, where it may be too loud.
In this case, the globe valve downstream of the trap can be shut for a while. Almost all of the ultrasonic level that is measured now can be attributed to the extraneous sound sources because there is nothing flowing through the trap. If a higher overall sound level is ascertained when the globe valve is opened, the difference compared to the previous level must be due to the trap.
The ultrasonic test itself often does not provide any indication of whether a steam trap is passing steam or condensate at a specific moment in time. It should be combined with a surface temperature measurement for this reason. A combination test device is shown in Fig. 8-3. In addition to the ultrasonic level, this tester also measures the surface temperature of the trap. The measured ultrasound and temperature values are visualised on the tester's display.
Fig. 8-3: Combined temperature / ultrasonic testing device, Type Multifunction tester
The measured values are compared with a limit value that varies according to the steam pressure and trap type. If they remain below this limit throughout the measurement phase, the trap is operating correctly and there is no leakage of live steam. If the limit value is exceeded, the trap is passing steam. The amount by which it is exceeded corresponds to the quantity of steam lost. All the values measured can be stored in the device and later processed on a PC. The operation of the various traps can thus be monitored over a long period of time. The temperature sensor compares the measured temperature with the saturation temperature to determine whether or not condensate has backed up upstream of the trap.
If the steam pressure downstream of the pressure control valve is known, the saturation temperature can be read off from the steam table. If it is not known, we recommend calculating the saturation temperature using the surface temperature measurement method described earlier in this chapter.
Assessment of the measured values for float traps
Assuming the trap is functioning correctly, only a float-type steam trap will produce a continuous ultrasonic signal because it discharges condensate continuously. If the values measured for a float trap (with a specific steam pressure) remain below the limit value, only condensate is being discharged. Fig. 8-4 shows a printout for a trap that is leaking live steam. To make sure that the high sound value is not due to pressure reducing valves or other apparatus in the vicinity, it is advisable to eliminate ambient noise using the method described above.
Fig. 8-4: Printout for a float trap that is leaking live steam
This signal exceeds the limit value throughout the measurement phase. There is no ambient noise that could distort the measured values. The surface temperature is identical to the saturation temperature. Conclusion: The trap is passing steam.
Assessment of the measured values for thermostatic traps
Thermostatic steam traps do not drain condensate continuously, so the measured ultrasonic signal has gaps (the same is also true when it comes to thermodynamic and inverted bucket traps). The curve recorded for a thermostatic trap during drainage is likely to be partly above and partly below the limit value. The fact that a measured value is higher than the limit value does not necessarily mean the trap is passing steam. Only if the measured value does not return to zero at the end of the trap cycle is it accurate to speak of steam loss. "Zero" means simply ambient noise, in other words the value that is determined when the trap is shut. The zero value based on ambient noise can sometimes be greater than zero. If the sound value measured for a thermostatic trap does not return to zero again at the end of the cycle, a comparative temperature measurement upstream of the trap can provide an indication of whether it is working correctly. If the temperature is close to or greater than the saturation temperature, the trap is passing steam. If a thermostatic trap has a thermodynamically amplified conical movement, the ultrasonic image is often extremely sharp. Fig. 8-5 and Fig. 8-6 respectively show the signal printout for a thermostatic trap that is operating correctly without and with thermodynamic amplification.
Fig. 8-5:Printout for an ultrasonic test of a thermic (bimetallic-) steam trap without thermodynamic amplification (continousliy operation)
Fig. 8-6:Printout for an ultrasonic test of a thermic (bimetallic-) steam trap with thermodynamic amplification (intermittent operation)
Assessment of the measured values for inverted bucket traps
The cycle of an inverted bucket steam trap is very simple to reconstruct. The condensate is discharged in bursts and the measured value rises to 100 %. If steam is present in the trap, the ultrasonic value is quickly reduced to zero. If a constant, high signal is measured, the trap is passing live steam.
Assessment of the measured values for thermodynamic traps
The working cycle of thermodynamic traps can be reconstructed equally easily. A full-scale deflection is indicated when the trap opens. The sound value reaches its maximum. At the moment of closing, this value returns almost to zero. If a constant, high value is measured, the trap is passing steam. More than five cycles (opening / closing) a minute are a sign that the valve disc and seat are no longer able to shut sufficiently tightly and the trap is approaching the end of its life.
Conductivity measurement
The conductivity is measured in a test chamber containing a water seal that is located upstream of the trap.
The condensate continues to flow under the separating rib. A small hole for pressure compensation is provided in the top of the rib, so that a constant condensate level exists in both chambers and the test chamber is prevented from running dry. If the trap is operating correctly, the water seal remains intact. If it is passing steam, the latter forces out the seal completely.
A conductivity electrode is mounted in the test chamber and a measuring instrument connected to the electrode contacts. If the test chamber is filled with condensate, the conductivity is detected by the electrode and a green signal lamp lights up. If the test chamber has been completely blown out due to the leakage of live steam, the electrode will no longer be surrounded by condensate and the circuit is interrupted. The live steam loss is indicated by a red signal lamp.
In addition to manual conductivity test devices, systems for monitoring the operation of several traps on a central instrument board are also available.
The conductivity electrode can optionally be installed directly in the steam trap (Fig. 8-7). In this case, there is no need for a test chamber upstream of the trap and the measurement is much more straightforward.
Fig. 8-7: Conductivity measurement in a steam trap
The conductivity measurement method is suitable for monitoring both continuously and intermittently operating steam traps. However, account must also be taken of the following factors and side-effects:
- The conductivity of high-purity condensate is so low that live steam loss is indicated (e.g. if fully demineralised water is used as feedwater for the steam boiler).
- Normal condensate contains impurities, and deposits on the electrodes convey false information.
- The conductivity measurement is meaningless if amines that form a protective layer are used to eliminate corrosion, because in this case protective layers will also build up on the electrodes.
- Condensate back-up is only indicated if a temperature sensor is installed in the test chamber.
- If the functional test upstream of the traps is carried out manually, there is a risk that
- the outlet will become dirty in the course of time, especially if the process frequently involves solids.
- The closing torque of some trap types is so near to the steam limit that the water seal may be completely forced out, even though there is no live steam.
Calorimetric testing
The calorimetric test method is based on the physical temperatures of gas (air) and steam that trigger a discharge function in the trap (Fig. 8-8).
Fig. 8-8: Calorimetric test method, ARI Type CONA®control
The chamber is installed in front of the steam trap being monitored. The chamber has two temperature sensors, one of which is continually heated. The information provided by the sensors is processed and output to LED indicators ("Power“, "Inspect Trap“, "Blockage“ and "Steam leakage“) The chamber remains full of condensate (Fig. 8-9).
Fig. 8-9: Correctly operating trap, ARI Type CONA®control
The sensor is flooded. The condensate in the test chamber cools down slightly. The heated temperature sensor transfers part of its heat to the condensate. A defined temperature difference between the two sensors is not exceeded. Only the "Power" light is lit and the steam trap is functioning correctly. If the trap does not shut steam-tight, the sensor will be surrounded by steam (Fig. 8-10).
Fig. 8-10: Leaking trap, ARI Type CONA®control
The temperature gradient between the heated and unheated sensors changes and the setpoint specified for the temperature difference is exceeded. This indicates that the trap is passing steam, and the "Steam Leakage" LED is lit.
Fig. 8-11: Blocked trap, ARI Type CONA®control
If condensate back-up occurs upstream of the trap, the test chamber is flooded (Fig. 8-11).
The condensate in the test chamber cools down. If no condensate is discharged because a trap is blocked, the temperature falls below the set limit value (default: 95 °C) and the sensor reports cold condensate. The "Blockage" LED lights up.
When the system is shut down, the steam pressure in the system is reduced and the hot condensate in the test chamber and the pipe vaporises. The sensor is surrounded by steam as a result and the temperature falls below the set limit value. It reports cold condensate because the minimum limit was exceeded as well as steam leakage because both probes are surrounded by steam / gas and the temperature difference is too large. Both the "Blockage" and the "Steam Leakage" lamps light up.
The calorimetric system described here with external or bolt-on test chambers for steam traps is suitable for indicating states centrally. Deviations are readily visible, a defective trap can be identified easily thanks to bus technology and costly steam losses are eliminated without any wasted time. The calorimetric sensor detects not only steam leakage but also blocked traps. The system is independent of condensate conductivity as well as deposits on the sensors that are liable to distort the measurements (e.g. protective amine layers or magnetite).
Summary of test methods
Visual inspection
Suitable as a method for determining condensate back-up. Can also be used with restrictions to establish live steam loss in thermostatic traps.
Noise measurement
Efficient way to monitor for live steam loss from intermittently operating traps. Considerable experience is necessary to interpret and assess the operation of float-type steam traps.
Conductivity measurement
Only recommended for detecting live steam leakage under ideal conditions. Sensitive to magnetite deposits, protective amine layers and other boiler chemicals. Does not work with high-purity condensate. Only suitable for determining condensate back-up in combination with a temperature sensor.
Calorimetric testing
Suitable for all trap types without the drawbacks of other methods. The calorimetric sensor detects not only steam leakage but also blocked traps. The system measures the functional state at a given instant independently of condensate conductivity or deposits and may additionally communicate with sophisticated computer monitoring and management systems.
8.5 Determination of leakage losses
The amount of live steam that is lost due to leaking traps tends to be underestimated. The arguments put forward for investments in a permanent steam trap monitoring system are not always taken seriously. After drawing up a loss calculation, however, the figures speak for themselves and there is no doubt that money spent on system of this kind or on regular inspections is justified. Assuming a steam price of €22.- /tonne and a leakage rate of 10 kg/h, a total of €2,000.- is wasted annually, resulting in unnecessary emissions of 13000 kg of CO2.
A decision was then taken to catalogue all traps and monitor their operation regularly by means of ultrasonic tests:
- Condensate back-up was clearly established in eight traps, which were subsequently replaced.
- Steam leakage was discovered in 16 traps. A constant, high ultrasonic signal was recorded for some of these traps, while in other cases the signal failed to return to the initial value. The next step was to gauge the losses from four traps with a high signal and four with a signal that did not return to zero.
It was agreed with the management of the plant that the highest (revealed later to be 60 kg/h) and lowest (revealed as 4 kg/h) measured values should be ignored for the purpose of the subsequent calculations. The average loss from the six remaining traps was then determined. This value was taken as representative for all 16 traps. The average loss per trap was 14 kg/h, making a total of 1900 tonnes of steam per year for the 16 traps together. Based on the present level of energy prices, this represents an annual loss of €42,000.- .
After being presented with the results of the survey, the management approved a budget for a permanent steam trap monitoring system with centralised evaluation. A handheld test unit was purchased to carry out temperature and ultrasonic tests on those traps that are too remote to be integrated in the central system. Special training in steam trap technology was provided to fitters and operators.
The traps that were found to be leaking steam were replaced within a week. The steam cloud on the condensate vessel is now visibly smaller. The reduction in the amount of make-up water matched the measured leakage quantity. Although gauging the leakage losses precisely upfront of this project was an arduous and time-consuming task, the outcome was extremely rewarding.
Gauging calculation (example):
Steam pressure in the heat exchanger 10 bar (saturation temperature = 180 ºC)
Enthalpy of 10 bar saturated steam hg = 2776 kJ
Amount of condensate during the measurement 5 kg
Amount of water at the start of the measurement 20 litres at 10 ºC
Measuring time 3 minutes
Condensate / water mixture temperature 62 ºC
Heat content at the end of the measurement 25 x 62 x 4,2 = 6510 kJ
Heat content at the start of the measurement 20 x 10 x 4.2 = 
Increase in heat content
Without live steam loss the mixture temperature should be: {(20 x 10) + (5 x 180)}/25 = 44 ºC with a heat content of 25 x 44 x 4.2 = 4620 kJ
The temperature rise due to live steam leakage is thus 5670 - 4620 = 1050 kJ. 1050/2776 = 0.38 kg of live steam is entrained with the condensate during the three minutes of the measurement, equivalent to 7.6 kg/h. This represents an annual loss of 65 tonnes of steam, which based on an energy price of €7.50 /GJ works out at a total financial loss of €1,355.- . Conclusion: The investment in new steam traps is recuperated in only a few months!