EP3409953B1

EP3409953B1 - Method in a compressed air system and related arrangement

Info

Publication number: EP3409953B1
Application number: EP17173684.6A
Authority: EP
Inventors: Santeri Pöyhönen; Tero Ahonen; Jero Ahola
Original assignee: ABB Schweiz AG
Current assignee: ABB Schweiz AG
Priority date: 2017-05-31
Filing date: 2017-05-31
Publication date: 2020-12-23
Anticipated expiration: 2037-05-31
Also published as: EP3409953A1

Description

FIELD OF THE INVENTION

The present invention relates to compressed air systems, and particularly to detection of amount of leakage from a compressed air systems.

BACKGROUND OF THE INVENTION

Compressed air is utilized throughout industries and is the greatest electrical energy consumer in many industrial facilities. In the European Union, air compressors account for approximately 10 % of the industry's total electricity consumption. At the same time, it has been found out that compressed-air systems (CAS) possess notable energy savings potential. Typically, the biggest preventable energy losses in a CAS can be attributed to air leaks. Monitoring the magnitude of leaks in an industrial CAS can show the system operator when often time consuming maintenance measures should be carried out to stop the air leaks. Thus, there is a need for an inexpensive method to quantify the leaks in CAS's.
The twin rotary screw compressor is one of the most common types of compressor used in the industry. Due to its geometry, the volumetric flow rate produced by the compressor is nearly directly proportional to its rotational speed. Thus, if the compressor is equipped with a frequency converter, the rotational speed estimate provided by the frequency converter can be used to estimate the produced volumetric flow rate.
To save energy in compressed air systems the leaks of the compressed air system should be detected. However, there are no automated procedures which would detect the leakage level of the compressed air systems.
Document DE202008013127U1 discloses a method of determining leakage level of a compressed air system. In the disclosure the running times of the compressor are recorded periodically. Based on the recorded running times, the leakage level is determined assuming that on a quiet period the use of the compressed air from the system is minimal.

BRIEF DESCRIPTION OF THE INVENTION

An object of the present invention is to provide a method and an arrangement for implementing the method so as to solve the above problem. The object of the invention is achieved by a method and an arrangement which are characterized by what is stated in the independent claims. The preferred embodiments of the invention are disclosed in the dependent claims.
The invention is based on the idea of using a frequency converter which is also used in producing the compressed air to the compressed air system in detecting the level of leakage of the system. The leakage level is preferably determined periodically, for example, daily or once in a week. Generally the determination of the leakage level can be determined when there is no consumption of the compressed air from the system.
An advantage of the method is that a leakage level of a compressed air system can be determined without any extra components. A frequency converter driving the compressor is utilized to gather the required data from the compressed air system together with a pressure sensor provided in the system. The frequency converter may also carry out the required calculations based on the data that is gathered.

BRIEF DESCRIPTION OF THE DRAWINGS

In the following the invention will be described in greater detail by means of preferred embodiments with reference to the attached drawings, in which

Figure 1 shows a flowchart of an embodiment of the invention;
Figure 2 shows an example of an exponential curve fitted on the system pressure measurement data;
Figure 3 shows an exemplary graph of the monitored development of the pressure decay coefficient k;
Figure 4 shows an exemplary scenario, where the estimate for the compressed air system volume differs between two sets of identification run instances; and
Figure 5 (a) and (b) show examples of the increase of the pressure during increase of the system pressure.

DETAILED DESCRIPTION OF THE INVENTION

In the present invention information on the volume of the compressed air system is needed in order to provide an estimate for the leakage mass flow rate. As shown in document Jarvisalo et al. 2016, the volume of an compressed air system can be estimated by the following equation, derived from the ideal gas law: $V_{sys} = \frac{\int_{t_{1}}^{t_{2}} q_{m, net} (t) dt \cdot R_{air}}{p}$
where q_m,net is the net mass flow rate into the system, t is time, R_air is the gas-specific constant of the compressed gas (i.e. 287.06 J/kgK for dry air), p is the system pressure, T is the system temperature, and subscripts 2 and 1 denote the end and start moments, respectively.
Using equation (1), a pressure measurement and a measurement or assumption for the temperature (T ₂ = T ₁ = suction-side temperature T _suc), the volume of a compressed air system can be determined with a fill-up identification sequence, which increases the system pressure from p ₁ to p ₂. In a compressed air system with leakage, if the magnitude of the leakage mass flow rate q _m,leak with respect to the mass flow rate produced by the compressor q _m,est into the system is negligible, we can assume that the system's net mass flow rate q _m,net is equal to the mass flow rate produced by the compressor (i.e., q _m,net = q _m,est). The compressor mass flow rate can be estimated with a frequency converter as is explained in the following.
Due to the displacing nature of the twin rotary screw compressor, each full rotation of the rotor delivers a constant volume of air from the suction inlet to the discharge end of the compressor. Thus, when the rotational speed of the compressor is known, the produced flow rate (q _v,est) can be estimated with equation $q_{v, est} = \frac{n_{est}}{n_{n}} * q_{v, n} - q_{v, offset}$
where q _v,n is the nominal volumetric flow rate at the nominal rotational speed n _n, which are both usually available on the compressor nameplate, n _est is the estimated rotational speed of the compressor provided by the frequency converter and q _v,offset denotes an offset resulting from compression losses. The estimated rotational speed corresponds closely to the rotational speed reference if such is provided to the frequency converter.
The resulting volumetric flow rate estimate can be converted into mass flow rate with equation $q_{m, est} = q_{v, est} * \frac{p_{suc}}{R_{specific} * T_{suc}}$
where p _suc and T _suc denote the pressure and temperature of the air at the compressor suction, respectively.
The accuracy of the soft-sensor, i.e. frequency converter based estimation, was evaluated with laboratory tests in Jarvisalo et al. 2016. It was found in the document that soft sensor has a constant offset, but also produces accurate information on the flow rate produced by the compressor driven with a frequency converter.
In the present invention an identification run is conducted to quantify the condition of a compressed air system with respect to leakage. The determined measure of leakage can be monitored over long periods of time to evaluate the condition of the system and its development system condition. In an embodiment estimates on the compressed air system leakage rate at any given system pressure are provided, which enables the economic analysis of compressed air system leakage. Such economic analysis includes the costs related to the leakage and the pay-back time of maintenance measures.
In the invention, the system pressure is measured using a pressure sensor or similar device which produces an output which indicates the pressure of the system. Due to the fact that leakage, i.e. unintentional compressed-air consumption, is determined based on a pressure measurement and the compressed air system volume estimator, the method cannot differentiate between intentional air consumption and unintentional consumption by leakage. Thus, to estimate leakage with the invention, the presented identification run should be carried out at a time of no intentional compressed-air consumption.
To be able to execute the leak estimation sequence at a time of zero compressed-air demand, allowed time frames, during which to run the leak estimation sequence, should be determined. Depending on the facility in which the compressed air system is situated, this may typically occur during night-time and weekends. Once a suitable time frame has been found, during which it can be reliably determined that no compressed-air demand exists, the identification run sequence to determine the system's leakage rate can be initiated.
During periods of time, when the compressed-air demand of a compressed air system is zero, the leakage-wise condition of the system can be determined by conducting an identification run sequence, which first increases the system's pressure to a level p ₂ (later referred to as "fill-up phase") and then lets it leak to a level p ₁ ("leak phase" later on). Then, based on the rate of the pressure decay during the leak phase of the identification run, the condition of the system can be determined. A flow diagram of an embodiment of the method is shown in Figure 1.
According to the invention the method comprises setting a first pressure limit p ₁ and a second pressure limit p ₂, the second pressure limit being higher than the first pressure limit and increasing the pressure of the compressed air system with the screw compressor until the pressure of the compressed air system reaches the second pressure limit. Preferably the pressure of the compressed air system is increased with the nominal rotational speed of the compressor.
Further in the invention during the increase of the pressure, i.e. when the screw compressor is rotated to push air to the compressed air system, method comprises estimating the rotational speed of the screw compressor with the frequency converter at multiple of time instants and measuring the system pressure of the compressed air system at multiple of time instants. The estimate of the rotational speed of the screw compressor can be obtained directly from the frequency converter and the system pressure is obtained using a pressure sensor. The pressure readings from the pressure sensor are fed to the frequency converter either directly or via some device operating between the frequency converter and the pressure sensor. The estimates of the rotational speed and values of pressure are determined multiple of times during the increase of the pressure. Typically the estimates are provided and the pressure is read in intervals ranging from 0,2 seconds to 5 seconds.
The determined rotational speed estimates and the measured pressures are stored together with timestamps. The values are stored preferably in the frequency converter.
When the pressure of the compressed air system has reached the second pressure value, the volume of the compressed air system is calculated from the determined rotational speed estimates and measured system pressures together with the stored timestamps and temperatures and pressures in the system at start and end of increase of the pressure and known constants. The calculation of the volume is carried out preferably with equation (1) and the calculation is preferably carried out in the frequency converter. As known, frequency converters have a certain amount of processing capacity together with readable memory which allow to perform specific calculations which are programmed to the frequency converter.
After the second pressure limit is reached, the method comprises letting the pressure to decrease from the compressed air system until the pressure falls to the first pressure limit. During the decrease of the pressure the system pressure again measured at multiple of time instants during the decrease of the pressure. Thus the frequency converter stops to drive the compressor when the pressure limit is reached and the pressure measurement is preferably carried in continuous manner from the increase phase of the pressure to the decrease phase. The measured values of pressure are stored together with timestamps similarly as during the increase of the pressure.
Once the pressure of the compressed air system has decreased to the first pressure level, an exponential curve is fitted on the pressure measurement data. It is known that the decay of pressure from a compressed air system follows an exponentially decaying curve. The curve is fitted preferably in the frequency converter. The exponential curve has a pressure decay coefficient which indicates the decay time of the pressure. The exponential fit can be produced with any known manner such as using a method of least squares.
Once the exponential curve has been fitted to the measured pressure data and the parameters of the fitted curve are known, the pressure decay coefficient is used as a measure indicating the leaking condition of the compressed air system as will be shown below in greater detail.
When leakage is the only cause of mass exiting the compressed air system, its pressure should decay exponentially. An exponential curve can be fitted on the pressure measurement data of a leak phase as seen in Figure 2. Thus, the behaviour of system pressure (p _sys) during the leak phase can be expressed as $p_{sys} (t) = a \cdot e^{- kt} .$
In equation (4) k is the pressure decay coefficient. According to the ideal gas law, the mass of the air contained in the compressed air system is expressed as $m_{sys} = \frac{p_{sys} \cdot V_{sys}}{R_{air} \cdot T_{sys}}$
where V_sys is the volume of the system and T_sys is the temperature in the system. When V_sys, R _air and T _sys are assumed constant, the following equation for the leakage mass flow rate can be derived from equation (5): $\frac{m_{sys}}{dt} = \frac{{dp}_{sys}}{dt} \cdot \frac{V_{sys}}{R_{air} \cdot T_{sys}} = q_{m, leak}$
With (4), the time derivative of system pressure, dp_sys /dt takes the form $\frac{p_{sys} (t)}{dt} = \frac{d}{dt} (a \cdot e^{- kt}) = - k \cdot a \cdot e^{- kt}$
In order to express dp_sys /dt as a function of the system pressure p_sys we substitute the time t in (7) with the inverse function of (4): $t (p) = \frac{\ln (\frac{a}{p})}{k}$
Thus, for the leakage mass flow rate, we get $q_{m} (p_{sys}) = - k \cdot a \cdot e^{- k \cdot \frac{\ln (\frac{a}{p_{sys}})}{k}} \cdot \frac{V_{sys}}{R \cdot T_{sys}} = - k \cdot p_{sys} \cdot \frac{V_{sys}}{R \cdot T_{sys}}$
With equation (9), once the system volume V_sys and the exponential coefficient k have been determined, the leakage mass flow rate from the compressed air system into the atmosphere can be calculated for any given system pressure. According to an embodiment of the invention the leakage mass flow rate as a function of system pressure is determined. Here, the coefficient k serves as a measure of the leak-wise condition of the compressed air system, and its value can be monitored over a long period of time to determine when maintenance measures to prevent leakage are necessary. Long-term data on the coefficient k can help reveal erroneous results, which should be filtered out, as seen in the exemplary graph shown in Figure 3. Figure 3 shows an example in which the calculation of the coefficient has been carried out during a period of multiple hundred days including 24 values of coefficient on different days. As seen in Figure 3, four of the values of k deviate clearly from the other values of k. Such deviating values can be can simply be ignored.
The comparability of the results for k acquired from identification runs conducted at different times can be assessed by comparing the identification runs' estimated values for V_sys. The k values of identification runs with significantly different values for V_sys. should not be compared. Difference in the V_sys provided by two identification runs suggests that the identified compressed air system network was not the same at the times when the identification runs were carried out. Figure 4 shows an example in which the system volumes between different identification runs are compared. The measurements correspond to that of Figure 3. It is seen from Figure 4 that the estimated system volumes are in line except for four identification runs in which deviating results were obtained. These deviating values correspond to those that are deviating in Figure 3.
Measures should be taken, if possible, to ensure that the compressed air system network is always the same (i.e. a certain part of the compressed air system is accessible by the compressed air produced by the compressor) when the leakage identification run is conducted. If the compressed air system network cannot be adjusted in such a way, monitoring the value for V_sys. as mentioned before can help overcome the issue of changes in the identified system. Furthermore, if V_sys. varies with multiple executions of the leakage identification run over time, and it is known which parts of the compressed air system were accessible by the air produced by the compressor during the identification run instances, it may even be possible to locate parts of the estimated leakage. This can be caused by differences in the compressed air system piping network configuration, i.e. a different combination of shut-off valves being closed or open during the identification run sets.
Ultimately, when a compressed air system's working pressure over time is known, the estimated leakage mass flow rate and the performance details provided by the compressor manufacturer or e.g. the compressor specific energy map, as determined in Järvisalo et al. 2016, can be used to calculate the cost of compressed air leakage.
Because the method cannot differentiate between unintended leakage and intentional compressed-air consumption, the results it provides are useful only if no consumption occurs during the fill-up phase or the leak phase of the identification run. Thus, to ensure the validity of the results obtained with the method, it is important that compressed air consumption occurring during the fill-up and leak phases of the sequence run can be detected. In the following, procedures to detect compressed air consumption and measures to take upon its detection are described.
As long as leaks are the only source of consumption in the compressed air system, the time derivative of the system pressure should remain nearly constant during the fill-up phase, as seen in (a) of Figure 5. If varying plant-operation-based consumption occurs during the fill-up phase, the occurring consumption should be visible through the deviation of the time derivative of the system pressure, the dp/dtvalue as seen in (b) of Figure 6 approximately between 420 and 450 seconds. A linear fit can be fitted on the pressure data of the fill-up phase and its goodness of fit can be analysed to decide whether consumption occurred during the fill-up phase. If consumption was detected, the system pressure will be allowed to decay back to p₁, and the fill-up phase repeated until a consumption-free iteration is achieved.
Similarly, consumption during the leak phase can be detected by analysing the goodness of the exponential fit, which was fitted on the pressure measurement data of the leak phase of the identification run. Again, if consumption is detected the leak phase will be repeated until a run with no consumption is carried out.
In the flowchart of Figure 1, the identification run is initiated at a predetermined time in 11 and it is checked if the system pressure is below the first pressure limit in 12. If the pressure is not below the limit, then the checking is carried out until the pressure falls below the limit. Once the pressure is below the limit, the compressor is started in 13 and the values of rotational speed estimates and pressure values together with timestamps are stored. The compressor is stopped when the second pressure limit is reached.
After the compressor is stopped, it is checked whether consumption of compressed air was detected during the fill-up phase in 14. If consumption is detected, the process returns to 12 and begins the fill-up phase again. If consumption is not detected the volume of the compressed air system is calculated using equation (1) in 15. Thus the volume is calculated only with such measurements in which consumption is not detected.
The compressor was stopped in 13 and the pressure starts to decrease. The values of the pressure are continuously measured and stored, and in 16 it is checked if the pressures has decreased to first pressure value. If not, then it is waited until it has while the pressure measurements are carried out in the system. Once the pressure has decreased to the first pressure limit, it is checked in 17 whether consumption was detected between the time interval from the stopping of the compressor to the time instant the pressure decreases to the first limit, i.e. from t₂ to t₃. If consumption of compressed air is detected, the compressor is started and run until pressure is again at the second pressure limit in 18, and the process continues to 16.
If consumption was not detected in 17, the exponential decay coefficient k is determined in 19, the control of the compressor drive is released in 20 and the identification run is ended in 21.
The invention relates also to an arrangement for determining leakage rate of a compressed air system and the arrangement comprises a pressure sensor and a screw compressor driven by a frequency converter. The arrangement further comprises means for setting a first pressure limit and a second pressure limit, the second pressure limit being higher than the first pressure limit. The means for setting the pressure limits is preferably implemented with the frequency converter. For example, the pressure limits may be set directly to the frequency converter by the operator of the arrangement.
The arrangement further comprises means for increasing the pressure of the compressed air system with the screw compressor until the pressure of the compressed air system reaches the second pressure limit. Such means for increasing the pressure of the system are the frequency converter and the screw compressor.
Further in the invention during the increase of the pressure the frequency converter is adapted to estimate the rotational speed of the screw compressor at multiple of time instants and the system pressure of the compressed air system is adapted to be measured at multiple of time instants. The system pressure is measured with the pressure sensor and preferably the pressure measurements are initiated by the frequency converter. That is, the frequency converter has a certain program according to which the pressure sensor is read. Further, the determined rotational speed estimates and the measured system pressures together with timestamps are adapted to be stored. The measured values with timestamps are stored preferably in the frequency converter.
The arrangement comprises further means for calculating the volume of the compressed air system from the determined rotational speed estimates and measured system pressures together with the stored timestamps and temperatures and pressures in the system at start and end of increase of the pressure and known constants. The means for calculating the volume of the compressed air system are preferably implemented using the frequency converter which has a certain amount of processing capacity for carrying out the calculations.
The arrangement further comprises means for letting the pressure to decrease from the compressed air system until the pressure falls to the first pressure limit, and the system pressure is adapted to be measured at multiple of time instants during the decrease of the pressure. The means for letting the pressure to decrease are preferably implemented using the frequency converter which may, for example, control the screw compressor in such a manner that pressure is not increased with the screw compressor and the pressure of the system falls due to the leaks.
Further, the arrangement comprises means for fitting an exponential curve on the pressure measurement data the exponential curve having a pressure decay coefficient. Such means are preferably implemented using the frequency converter which is programmed to carry out exponential curve fitting to the measurement data. Further, the arrangement comprises means for using the pressure decay coefficient as a measure indicating the leaking condition of the compressed air system. Such means is preferably implemented using the frequency converter which uses the decay coefficient for example for producing a series of decay coefficients from multiple of identification runs and displays the decay coefficient to the user. The frequency converter may also provide an alert if the leaking condition is higher than a set alert limit.
It will be obvious to a person skilled in the art that, as the technology advances, the inventive concept can be implemented in various ways. The invention and its embodiments are not limited to the examples described above but may vary within the scope of the claims.

Reference:

Jarvisalo, M., Ahonen, T., Ahola, J., Kosonen, A., Niemela, M. Soft-Sensor-Based Flow Rate and Specific Energy Estimation of Industrial Variable-Speed-Driven Twin Rotary Screw Compressor, IEE Transactions on Industrial Electronics, vol. 63, no. 6, pp. 3282-3289, May 2016.

Claims

A method of determining leakage rate of a compressed air system having a pressure sensor and a screw compressor driven by a frequency converter, wherein the method comprises
setting a first pressure limit and a second pressure limit, the second pressure limit being higher than the first pressure limit,
increasing (13) the pressure of the compressed air system with the screw compressor until the pressure of the compressed air system reaches the second pressure limit, and during the increase of the pressure
estimating (13) the rotational speed of the screw compressor with the frequency converter at multiple of time instants,
measuring (13) the system pressure of the compressed air system at multiple of time instants, characterized in that the method comprises
storing (13) the determined rotational speed estimates and the measured system pressures together with timestamps,
calculating (15) the volume of the compressed air system from the determined rotational speed estimates and measured system pressures together with the stored timestamps and temperatures and pressures in the system at start and end of increase of the pressure and known constants, and
letting the pressure to decrease from the compressed air system until the pressure falls to the first pressure limit, and measuring the system pressure at multiple of time instants during the decrease of the pressure,
fitting (19) an exponential curve on the pressure measurement data the exponential curve having a pressure decay coefficient, and
using the pressure decay coefficient as a measure indicating the leaking condition of the compressed air system.
A method according to claim 1, wherein the method comprises after the increase of the pressure
determining the consumption of compressed air from the compressed air system during increase of the pressure from the linearity of the increase of the pressure using the stored system pressures, and if the increase of the pressure indicates consumption of compressed air, then
decreasing the pressure and repeating the step of increasing the pressure.
A method according to claim 1 or 2, wherein the method comprises after the decrease of the pressure
determining the consumption of compressed air from the compressed air system during the decrease of the pressure from the goodness of the exponential fit, and if the exponential fit indicates consumption of compressed air, then
increasing the pressure and repeating the step of letting the pressure to decrease.
A method according to claim 1, 2 or 3, wherein the method comprises repeating the process for obtaining a series of pressure decay coefficients, and using the timely increase of the obtained pressure decay coefficients as an indication of increased leakage level.
A method according to claim 4, wherein the method further comprises comparing the calculated volumes of the compressed air system which are calculated while repeating the process for indicating validity of the pressure decay coefficients.
A method according to any of the previous claims 1 to 5, wherein the method further comprises
forming a function defining leakage mass flow as a function of the pressure of the compressed air system using the determined pressure decay coefficient.
An arrangement for determining leakage rate of a compressed air system, the arrangement comprising a pressure sensor and a screw compressor driven by a frequency converter, wherein the arrangement comprises
means for setting a first pressure limit and a second pressure limit, the second pressure limit being higher than the first pressure limit,
means for increasing the pressure of the compressed air system with the screw compressor until the pressure of the compressed air system reaches the second pressure limit, and during the increase of the pressure
the frequency converter is adapted to estimate the rotational speed of the screw compressor at multiple of time instants,
the system pressure of the compressed air system is adapted to be measured at multiple of time instants, and
the determined rotational speed estimates and the measured system pressures together with timestamps are adapted to be stored, characterized in that
the arrangement comprises further means for calculating the volume of the compressed air system from the determined rotational speed estimates and measured system pressures together with the stored timestamps and temperatures and pressures in the system at start and end of increase of the pressure and known constants, and
means for letting the pressure to decrease from the compressed air system until the pressure falls to the first pressure limit, and the system pressure is adapted to be measured at multiple of time instants during the decrease of the pressure,
means for fitting an exponential curve on the pressure measurement data the exponential curve having a pressure decay coefficient, and
means for using the pressure decay coefficient as a measure indicating the leaking condition of the compressed air system.