/ Dokumentnamn/Name of document
status rEport / Sida/Page
21(21)
Utfärdare (avd nr, namn, tfn, geo placering, sign)/Issuer (dept, name, phone, signature)
6110, Timothy Benham, 332 09 85, PVÖA 201 / Datum/Date
2003-05-23 / Bilaga/Appendix / Reg nr/Reg. No
6700-03-151
Ärende/Subject
Volvo Measurements within the Particulates project
Receiver
Assoc. Professor Zissis Samaras.
Status-report on Volvo’s particle measurements
within project ’Particulates’
Introduction 2
Construction of Particulates diluter 2
Mass-flow controller 2
Control and data collection 2
Experimental set-up and operation 2
Mounting on the CVS rig 2
Extra ’Nose’ to prevent turbulent problems 2
Dilution air 2
Primary dilution control 2
Dilution air temperature control 2
CPC data 2
DGI flow control 2
Test program 2
Euro III engine 2
Euro I engine 2
Euro III + CRT engine 2
Preliminary results and discussion 2
Comparison between technologies 2
Effect of fuel 2
Dilution ratio under SS points 2
Conclusions 2
Introduction
Much of the interest by Volvo in the Particulates project was to gain more experience and knowledge in particulate measurement. The project assumed a high degree of competence in operation of particle equipment and CVS systems. Volvo Technology (VTEC) has a short history of particle measurement and thus limited experience in the use of such equipment, which has meant a very fast learning process for all involved.
Since we had access to competence from Volvo Powertrain Corporation (VPT), the experience with CVS systems was available as response to questioning but not held by the individuals planning and carrying out the work from VTEC’s side. The CVS rigs at VPT are always heavily booked with certification work and thus access to these rigs by research projects is extremely limited and liable to re-planning at short notice. Due to this limitation in access it was not possible to test the Particulates system fully prior to measurement.
Many problems with building and operating the particulates system were avoided or identified at an early stage due to the help and advice offered to VTEC by VTT.
Construction of Particulates diluter
It is regrettable that the parts of the dilution system were supplied without any drawings, instructions or specifications. This lack of information caused several problems with the construction of the system and thus delayed its commissioning. This manifest itself in several ways. One was during the initial trial of the system, it was discovered that an extension to the aging chamber was missing so the required aging time could not be achieved. When the necessary tube was sent from Dekati the system had to be re-built to accommodate it, and thus would no longer fit into the available space on the CVS rig. During engine testing another rig was used which had more space for the diluter. This rig also had to be modified to mount the diluter therefore.
The two most difficult problems to solve were a: the DGI flow control since the DGI requires control of a large volume of air at low pressure, and b: the overall system control and data collection since Dekati recommended the construction of a control unit rather than purchase of the bespoke unit.
Mass-flow controller
Since the DGI required a mass-flow controller which had a very low pressure-drop and was capable of operating in partial vacuum, this had to be identified and purchased for the project. The model finally chosen was a Bronkhorst MFC F-112AC-FAC-44-V coupled to a regulating valve F-004BC-IV-44-V.
The controller was coupled to an Edwardes vacuum-pump with a capacity of 12m3/hr.
This combination proved successful for the operation of the DGI.
Control and data collection
The choice for control of the hardware, and data collection was the modular National Instruments Field-Point system. This has the benefit of using standard Ethernet communication with the possibility of placing the slave-unit a long distance from the master computer and operator. Being a modular system, it can be easily extended with units for any type of data collection or transmission.
It was therefore necessary to construct a control-program and user-interface. This service was bought in from a consultant programmer. The resulting PMP program (Particle Measurement Program) controlled and collected data from the mass-flow controllers, thermocouples, NOx analysers, the CPC, and communicated with the ASMO via its serial port. Dilution ratios were calculated on-line which simplified the set-up process. Only 2 computers were needed for data collection, one for the ELPI, and the system control computer. All data was saved at 1-second intervals in a common file as ASCII.
Figure 1 shows the graphical user interface of the PMP program constructed for control of the Particulates dilution system. The features seen are as follows:
Three main fields for display of data both graphically and numerically. The graphs can be configured to display any channel of data verses time with axes configurable accordingly. The numerical display shows both control-data and collected data, with the respective units.
Below the main data fields are 4 control-fields. The one to the far right is remote control of the ASMO. The one to the middle-right (Mass-flow control) shows the set points of the flow-controllers, in practice only 2 were used. The field to the middle-left (Dilution air temperature control) shows the dilution air set-point and duty-cycle control of the cooling, plus indication of when cooling is active. The final field to the left is the main system control. The system is started/stopped with the [Start system] toggle-button. To create a data-file and start saving data the [Start test] toggle-button is used. On start of a test the time in hh:mm:ss is displayed to the right of the button. The third button starts/stops flow through the DGI and an indicator shows when this is active.
The field at the bottom of the picture shows the current name of the data file being used plus status and communications information from the Field-Point unit.
The input-output channels and test-set-up data are input in a ‘Set-up’ window from the ‘Set-up’ menu button.
Figure 1. Example of the GUI for the control and data collection program PMP.
Experimental set-up and operation
Mounting on the CVS rig
There was no suitable access into the exhaust-pipe which could be used for mounting the primary diluter. To permit mounting a flange was made to fit that of the diluter and it was then welded into the exhaust-pipe. The diluter could then be directly bolted to the flange, see Figure 2.
The only available access for the diluter in the rig used for engine testing was in a bend immediately in front of the CVS mixing chamber. The flange was mounted at an angle so that the instruments would be at a suitable height for operation. A second, smaller, flange was mounted for measurement of exhaust pressure and temperature at the diluter position, see Figure 2.
Figure 2. Flange mounting of the primary diluter.
By mounting the diluter at this angle it was possible to get all instruments into the limited space available, see Figure 3. It will be seen that the system is very compact and thus the length of transport tubing is as short as possible, and thus the particle transport losses are low.
Figure 3. The whole dilution system.
Extra ’Nose’ to prevent turbulent problems
On the first day of testing it was found that the primary dilution ratio could not be set as required. The ratio of 12.5 could not be achieved even with extra dilution air from a secondary mass-flow controller, and was extremely unstable. After many system checks it was concluded that there must be a fault with the primary diluter and it was therefore dismounted and inspected.
It was found that the diluter was not constructed with a porous-tube as was thought, but a with a perforated-tube with the main perforations at the tip of the diluter. The dilution-ratio fault was therefore believed to be due to turbulence in the exhaust-pipe which was drawing the dilution air out of the diluter since it was delivered so near to the entrance of the probe. The volume being drawn by the measurement instruments was therefore being made up by raw exhaust and thus the correct dilution ratio could not be set.
Dekati was contacted and confirmed that the diluter was of the perforated type, and sent a specification for a nose for the diluter which would reduce the turbulence at the tip. This is shown mounted on the diluter in Figure 4. It is clamped in place and has a short perforated tube at the inlet followed by a diffuser up to the internal diameter of the diluter.
Figure 4. Anti-turbulence nose mounted on diluter.
The extra length caused by the anti-turbulence nose caused the sample point of the diluter to be from the inside of the bend of the exhaust pipe rather than in the centre as had been intended.
After the nose had become heated in the exhaust flow, the screws clamping it to the diluter could not be loosened again. The nose could therefore not be removed without damage, which meant that a small leakage into the system could not be avoided during the system tightness test since the nose did not have an air-tight seal to the probe.
Dilution air
To ensure that the dilution air was sufficiently dry for the tests (<5% RH) a double column dehumidifier was used. This dried the air to a level of 3 – 4% RH. It is constructed as 2 columns of drying agent which are used alternately. While one column is being used the other is being regenerated.
During the first day of testing it became apparent that the switching from one column to the other caused a sudden pressure-drop at the outlet of the dryer. This caused a temporary drop in the dilution ratio and therefore a spike in the analysed data every 2 minutes. These spikes have been manually edited out of the steady-state data since they are known to be false. It is not possible to edit them out of the transient data however since this is in itself a series of spikes.
As soon as was possible a pressure equalizing tank was fitted after the dryer to prevent these fluctuations. Since this was working at 7bar and the diluters were using pressure-reduced air the dilution ratio was no longer affected and the spikes did not appear. The large majority of the tests were conducted with this tank mounted.
Primary dilution control
The primary dilution control was affected by a mass-flow controller which was manually adjusted to give a dilution ratio of 12.5. At each steady-state point the dilution ratio was adjusted after an initial stabilisation time defined by the normal parameters used by Volvo Powertrain during engine testing. For the cycles, the dilution ratio was set with the engine running at its B50 (engine speed B, 50% load) point. This air-flow to the diluter was fixed during the whole cycle.
The primary dilution ratio was calculated from the raw and diluted NOx concentrations and displayed in real-time by the PMP program. NOx was chosen rather than CO2 since it avoided the necessity to re-calculate the concentrations for wet gas. This is possible as the measurement is made hot so that the gas does not need to be dried prior to the instrument. CO2 instruments require that the gas is dried prior to measurement which requires the concentration to be corrected for the removed moisture which would have been necessary for the raw gases.
The raw concentration of NOx was measured by that channel of the rig’s gas-analyser while the dilute concentration was measured by an extra analyser. For daily calibration of the second and third dilution stages the rig’s ‘dilute’ analyser was used but this was otherwise coupled to the CVS. Due to the high dilution factor on the wet-branch calibration was done by injecting a high concentration NOx calibration-gas into the dilution system in-front of the extra dilution and measuring the concentration after dilution.
In the initial period of testing the extra analyser used was an Eco Physics unit, in the second and third measurement periods the extra analyser was a Horiba. All NOx analysers were calibrated at least once per day and in most cases many times per day. The analogue NOx signals (0 – 10V) were collected by the PMP program.
Dilution air temperature control
During the construction of the dilution control unit it was found by VTT that the cooler supplied for the diluter was incapable of providing sufficient cooling when the system was used in the exhaust flow of a heavy-duty engine. For this reason water was used as the cooling agent.
The system had already been constructed to operate with a simple on/off valve for compressed air and since air has a low heat capacity, large fluctuations in temperature were not envisaged. However water has a specific heat capacity which is 4 times that of air and a density of 1000 that of air, so is capable of absorbing 4000 times the amount of heat by a given volume. This initially caused extremely large swings of the dilution-air temperature when the cooling valve operated.
To avoid this a more sophisticated temperature control would have been desirable, but at that stage in the project this was not practical as it would have required a re-design of the equipment and control unit. The solution chosen was to build in duty-cycle control of the cooling valve so that it could be trimmed by the operator. By doing this the dilution air temperature was held at 32°C within the required tolerance.
CPC data
The CPC used by VTEC was a TSI model 3010 which has a maximum for individual particle counting of 10000 #/cm3. Since it was intended that the CPC should be used only in the individual particle counting mode the analogue output, which is linear within this range, was used and read by the PMP program.
In deliverable 3 describing the operation of the CPC it is stated that it shall be set to measure D50 at 7nm. Unfortunately it does not explain how this is done. The TSI manual for the instrument does not describe this procedure either. TSI was therefore contacted and asked how this setting is made. The reply was that this is company confidential information and use of the CPC outside its default settings was not recommended as it was not characterised, and would invalidate the guarantee of the instrument.