US20220356873A1

US20220356873A1 - Method for identifying damage on a compressor

Info

Publication number: US20220356873A1
Application number: US17/618,545
Authority: US
Inventors: Lorenz Kunz
Original assignee: BASF SE
Current assignee: BASF SE
Priority date: 2019-06-14
Filing date: 2020-06-04
Publication date: 2022-11-10
Also published as: EP3983681A1; CN113906216B; EP3983681B1; WO2020249461A1; CN113906216A

Abstract

Described herein is a method for identifying damage on a compressor having an intake side and a discharge side, including the following steps: (i) detecting measurement data of the intake pressure (p1) and intake temperature (T1) measurement variables on the intake side, as well as end pressure (p2) and end temperature (T2) on the discharge side; (ii) determining a calculated end temperature (T2b), a calculated intake temperature (T1b), a calculated end pressure (p2b) or a calculated intake pressure (p1b) as a target variable, representing a good operating state of the compressor, as a function of the measurement data of max. three of the measurement variables (p1, T1, p2, T2); (iii) determining a comparison variable from at least one of the measurement variables (p1, T1, p2, T2) not used in step (ii); and (iv) comparing the comparison variable and the target variable as a gauge of damage to the compressor.

Description

The invention relates to a method for identifying damage on a compressor having an intake side and a discharge side, in which, starting from measurement data of intake pressure, intake temperature, end pressure and end temperature, a comparison variable is calculated as a measure of damage.
Compressors belong to the fluid energy machines which, as work machines, convert supplied energy into different energy states. Compressors are used in various forms, for example in the form of reciprocating compressors for compressing gases.
Compressors are usually operated continuously for several months or years and are stopped only for maintenance purposes. During this continuous operation, the functionality of components of the compressor may become impaired, for example by wear, deposits or component failure. This can lead to a drop in the efficiency of the compressor up to the complete functional incapability thereof. In order to be able to identify such processes early and optionally take measures which counteract damage such as wear, erosion or deposits, various monitoring and diagnosis methods are known in the prior art. In reciprocating compressors, monitoring of the valves on the intake side and on the discharge side is highly relevant in this respect.
For example, document EP 1 184 570 A2 describes a system for monitoring the valves of a reciprocating compressor, in which piezoelectric vibration sensors are mounted on each cylinder of the compressor. The sensors detect, via vibrations, the noise generated by the valves as they open and close. By means of downstream signal processing, conclusions can be drawn about the current state of the compressor.
In US patent application US 2010/0106458 A1, a different method is used to monitor a reciprocating compressor. Starting from pressure measurements, state information is obtained with the aid of neural networks via wavelet analyses of the signal sequences obtained.
Combinations of the above-mentioned measuring methods are also known. For example, document US 2012/0134850 A1 describes a method and a system for monitoring reciprocating compressors, in which at least one pressure sensor and at least one vibration sensor each acquire data and, from the combination of both sets of data, information is obtained about the state of the compressor.
Specifically for monitoring the valves of a reciprocating compressor, methods are known in which sensors are mounted on all the valves to be monitored. Such sensors are mostly temperature sensors or vibration sensors from which alone or from a combination with further sensors on the compressor, information about the state of the machine is obtained. US patent application US 2017/0030349 A1 describes such a method.
In US patent application US 2013/0115109 A1, a method for monitoring compressors is described, in which process data are recorded by means of pressure and temperature sensors at the inlet and at the outlet of the compressor. By means of an evaluation logic, the pressure loss across the valves of the compressor is determined in order to determine therefrom target values for the outlet temperature of the compressor. By comparing the calculated target value with the actual value of the outlet temperature, the evaluation logic reaches a conclusion about the current operating state of the compressor and optionally emits a warning.
Document JP 2002 147905 A discloses a method for monitoring a compressor in a refrigerating apparatus, in which, via pressure and temperature sensors at the inlet and at the outlet of the compressor, a criterion for the state of the compressor is determined from measured values of the intake temperature, the intake pressure, the outlet temperature and the outlet pressure, for example by calculation of a polytropic exponent.
A disadvantage of the methods and systems known from the prior art is that they require complicated instrumentation, for example in the form of vibration sensors on all the components to be measured, and/or a complicated evaluation logic for supplying the desired information from the measured signals.
The object was to provide a method for monitoring compressors which is able to reliably provide information about possible damage within the compressor and is thereby simple and inexpensive to install and maintain.
This object is achieved in accordance with the invention by a method as claimed in claim 1. Preferred forms of the invention are indicated in dependent claims 2 to 6. Claim 7 and claim 8 describe a computer program and a computer program product which are suitable for carrying out a method according to the invention.
A subject of the invention is a method for identifying damage on a compressor having an intake side and a discharge side, wherein the method comprises the following steps: (i) acquiring measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side and end pressure (p2) and end temperature (T2) on the discharge side;

- (ii) determining a calculated end temperature (T2b), a calculated intake temperature (T1b), a calculated end pressure (p2b) or a calculated intake pressure (p1b) as a target variable which represents a good state of the compressor, as a function of the measurement data of a maximum of three of the measurement variables (p1, T1, p2, T2);
- (iii) determining a comparison variable from at least one of the measurement variables (p1, T1, p2, T2) not used in step (ii);
- (iv) comparing the comparison variable and the target variable as a measure of damage on the compressor.

In accordance with the invention, the target variable determined in step (ii) is determined according to an isentropic compression model including the isentropic exponent (κK) of the gas to be compressed and a correction factor (η), and the correction factor (η) is adjusted on the basis of measurement data.
The isentropic exponent (κ) of the gas to be compressed required for the calculation is known to the person skilled in the art and can be found, for example, in publicly accessible or commercially available databases or tables.
In the case of gases to be compressed which, owing to their thermodynamic properties, behave in a similar manner to an ideal gas, the isentropic compression model is sufficient to adequately describe the actual processes. The correction factor (η) can in this case be omitted or be set at a neutral value.
By contrast, in the case of gases to be compressed which, owing to their thermodynamic properties, differ from the behavior of an ideal gas, the correction factor (η) is to be included in the calculation of the target variables, the correction factor taking account of effects of the actual compression, for example owing to heating of the gas during the intake stroke by thermal conduction at inner walls of the compressor, in the intake valve or by mixing of the gas drawn in with hot residual gas in the compression chamber. This correction factor (η) is adjusted on the basis of measurement data.
Method steps (i) to (iv) can be carried out in the indicated order. However, different orders of the method steps are also covered by the invention. In particular, steps (ii) and (iii) can also be carried out in the reverse order or also simultaneously.
In a first advantageous embodiment of the method in accordance with the invention, in step (ii) a calculated end temperature (T2b) is determined as the target variable as a function of the measurement data of the end pressure (p2), the intake pressure (p1) and the intake temperature (T1), and in step (iii) the measured end temperature (T2) is determined as the comparison variable. The method accordingly comprises the steps:

- (i) acquiring measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side and end pressure (p2) and end temperature (T2) on the discharge side;
- (ii) determining a calculated end temperature (T2b) as the target variable which represents a good state of the compressor, as a function of the measurement data of the end pressure (p2), the intake pressure (p1) and the intake temperature (T1);
- (iii) determining the measurement data of the end temperature (T2) as the comparison variable;
- (iv) comparing the comparison variable (T2) and the target variable (T2b) as a measure of damage on the compressor;
- wherein the end temperature (T2b) calculated in step (ii) is determined according to an isentropic compression model including the isentropic exponent (κ) of the gas to be compressed and a correction factor (η), and the correction factor (η) is adjusted on the basis of measurement data.

In a second advantageous embodiment of the method in accordance with the invention, in step (ii) a calculated intake temperature (T1b) is determined as the target variable as a function of the measurement data of the intake pressure (p1), the end pressure (p2) and the end temperature (T2), and in step (iii) the measured intake temperature (T1) is determined as the comparison variable. The method accordingly comprises the steps:

- (i) acquiring measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side and end pressure (p2) and end temperature (T2) on the discharge side;
- (ii) determining a calculated intake temperature (T1b) as the target variable which represents a good state of the compressor, as a function of the measurement data of the intake pressure (p1), the end pressure (p2) and the end temperature (T2);
- (iii) determining the measurement data of the intake temperature (T1) as the comparison variable;
- (iv) comparing the comparison variable (T1) and the target variable (T1b) as a measure of damage on the compressor;

wherein the intake temperature (T1b) calculated in step (ii) is determined according to an isentropic compression model including the isentropic exponent (κ) of the gas to be compressed and a correction factor (η), and the correction factor (η) is adjusted on the basis of measurement data.
In a third advantageous embodiment of the method in accordance with the invention, in step (ii) a calculated end pressure (p2b) is determined as the target variable as a function of the measurement data of the end temperature (T2), the intake pressure (p1) and the intake temperature (T1), and in step (iii) the measured end pressure (p2) is determined as the comparison variable. The method accordingly comprises the steps:

- (i) acquiring measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side and end pressure (p2) and end temperature (T2) on the discharge side;
- (ii) determining a calculated end pressure (p2b) as the target variable which represents a good state of the compressor, as a function of the measurement data of the end temperature (T2), the intake pressure (p1) and the intake temperature (T1);
- (iii) determining the measurement data of the end pressure (p2) as the comparison variable;
- (iv) comparing the comparison variable (p2) and the target variable (p2b) as a measure of damage on the compressor;

wherein the end pressure (p2b) calculated in step (ii) is determined according to an isentropic compression model including the isentropic exponent (κ) of the gas to be compressed and a correction factor (η), and the correction factor (η) is adjusted on the basis of measurement data.
In a fourth advantageous embodiment of the method in accordance with the invention, in step (ii) a calculated intake pressure (p1b) is determined as the target variable as a function of the measurement data of the intake temperature (T1), the end pressure (p2) and the end temperature (T2), and in step (iii) the measured intake pressure (p1) is determined as the comparison variable. The method accordingly comprises the steps:

- (i) acquiring measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side and end pressure (p2) and end temperature (T2) on the discharge side;
- (ii) determining a calculated intake pressure (p1b) as the target variable which represents a good state of the compressor, as a function of the measurement data of the intake temperature (T1), the end pressure (p2) and the end temperature (T2);
- (iii) determining the measurement data of the intake pressure (p1) as the comparison variable;
- (iv) comparing the comparison variable (p1) and the target variable (p1b) as a measure of damage on the compressor;

wherein the intake pressure (p1b) calculated in step (ii) is determined according to an isentropic compression model including the isentropic exponent (κ) of the gas to be compressed and a correction factor (η), and the correction factor (η) is adjusted on the basis of measurement data.
When acquiring the measurement data, it is not absolutely necessary that the measurement variables intake pressure, intake temperature, end pressure and end temperature are acquired as separate measurement variables. It is possible in accordance with the invention, in dependence on the specific embodiment, also to acquire combined or derived measurement variables. For example, in an embodiment in which the determination of the target variable in step (ii) is dependent on a ratio of the end pressure (p2) to the intake pressure (p1), it is possible, instead of acquiring the measurement variables intake pressure (p1) and end pressure (p2), also to acquire the ratio (p2/p1 or p1/p2) thereof directly as the measurement variable.
In the first embodiment, in which measurement data of the end temperature (T2) are determined as the comparison variable, it is further preferred that the calculated end temperature (T2b) is determined in accordance with the equation
T2b=T1/η·(p2/p1){circumflex over ( )}(1−1/κ),
wherein κ is the isentropic exponent of the gas to be compressed. The correction factor η can be constant or can be adjusted in dependence on the measurement variables. In a variant, the correction factor η is determined as a function of the intake temperature (T1), the intake pressure (p1) and the end pressure (p2).
In the second embodiment, in which measurement data of the intake temperature (T1) are determined as the comparison variable, it is further preferred that the calculated intake temperature (T1b) is determined in accordance with the equation
T1b=T2·η·(p1/p2){circumflex over ( )}(1−1/κ),
wherein κ is the isentropic exponent of the gas to be compressed. The correction factor η can be constant or can be adjusted in dependence on the measured variables. In a variant, the correction factor η is determined as a function of the end temperature (T2), the intake pressure (p1) and the end pressure (p2).
In the third embodiment, in which measurement data of the end pressure (p2) are determined as the comparison variable, it is further preferred that the calculated end pressure (p2b) is determined in accordance with the equation
p2b=p1·(η·T2/T1){circumflex over ( )}(κ/κ−1)),
wherein κ is the isentropic exponent of the gas to be compressed. The correction factor η can be constant or can be adjusted in dependence on the measurement variables. In a variant, the correction factor η is determined as a function of the intake temperature (T1), the intake pressure (p1) and the end temperature (T2).
In the fourth embodiment, in which measurement data of the intake pressure (p1) are determined as the comparison variable, it is further preferred that the calculated intake pressure (p1b) is determined in accordance with the equation
p1b=p2·(T1/T2/η){circumflex over ( )}(κ/(κ−1)),
wherein κ is the isentropic exponent of the gas to be compressed. The correction factor η can be constant or can be adjusted in dependence on the measurement variables. In a variant, the correction factor η is determined as a function of the intake temperature (T1), the end temperature (T2) and the end pressure (p2).
The adjustment of the correction factor η can be carried out in different ways. In an advantageous form, the correction factor η is determined by regression from historical measurement data. In a further advantageous form, the adjustment of the correction factor (η) is carried out on the basis of measurement data in that, following an overhaul of a compressor, the measurement values acquired after restarting are defined as good and used for adjusting the correction factor. The compressor can thereby purposively be operated by predefined operating states in order to define the good state.
In the first embodiment, in which measurement data of the end temperature (T2) are determined as the comparison variable, it is accordingly preferred that the correction factor η is calculated in accordance with the equation
η=a·T1+b·p2/p1+c
and the factors a, b and c are determined by regression from measurement data of intake temperature (T1), intake pressure (p1) and end pressure (p2).
In the second embodiment, in which measurement data of the intake temperature (T1) are determined as the comparison variable, it is accordingly preferred that the correction factor η is calculated in accordance with the equation
η=a·T2+b·p1/p2+c
and the factors a, b and c are determined by regression from measurement data of end temperature (T2), intake pressure (p1) and end pressure (p2).
In the third embodiment, in which measurement data of the end pressure (p2) are determined as the comparison variable, it is accordingly preferred that the correction factor η is calculated in accordance with the equation
η=a·p1+b·T2/T1+c
and the factors a, b and c are determined by regression from measurement data of intake temperature (T1), intake pressure (p1) and end temperature (T2).
In the fourth embodiment, in which measurement data of the intake pressure (p1) are determined as the comparison variable, it is accordingly preferred that the correction factor η is calculated in accordance with the equation
η=a·p2+b·T1/T2+c
and the factors a, b and c are determined by regression from measurement data of intake temperature (T1), end temperature (T2) and end pressure (p2).
When adjusting the correction factor 17, further measurable variables can also be used, for example a speed of the compressor (N), control signals of the intake valve lifter (s), a clearance volume (k) or the gas composition (w₁, w₂, w₃, etc.).
In the first embodiment, in which measurement data of the end temperature (T2) are determined as the comparison variable, the correction factor η can be calculated, for example, in accordance with the equation
η=a·T1+b·p2/p1+c+d·N+e·s+f·k+g ₁ ·w ₁ +g ₂ ·w ₂+ . . .
and the factors (a, b, c, d, e, f, g₁, g₂, . . . ) can be determined by regression from the corresponding measurement data.
In the second embodiment, in which measurement data of the intake temperature (T1) are determined as the comparison variable, the correction factor η can be calculated, for example, in accordance with the equation
η=a·T2+b·p1/p2+c+d·N+e·s+f·k+g ₁ ·w ₁ +g ₂ ·w ₂+ . . .
and the factors (a, b, c, d, e, f, g₁, g₂, . . . ) are determined by regression from the corresponding measured data.
In the third embodiment, in which measurement data of the end pressure (p2) are determined as the comparison variable, the correction factor η can be calculated, for example, in accordance with the equation
η=a·p1+b·T2/T1+c+d·N+e·s+f·k+g ₁ ·w ₁ +g ₂ ·w ₂+ . . .
and the factors (a, b, c, d, e, f, g₁, g₂, . . . ) can be determined by regression from the corresponding measurement data.
In the fourth embodiment, in which measurement data of the intake pressure (p1) are determined as the comparison variable, the correction factor η can be calculated, for example, in accordance with the equation
η=a·p2+b·T1/T2+c+d·N+e·s+f·k+g ₁ ·w ₁ +g ₂ ·w ₂+ . . .
and the factors (a, b, c, d, e, f, g₁, g₂, . . . ) are determined by regression from the corresponding measured data.
The mentioned equations and calculation rules are generally to be interpreted such that additions to the respective terms are possible and included in the invention, for example with regard to normalization or scaling. Thus, for example, the term (p2/p1) is synonymous in terms of the content of the invention with the term (p2+1)/(p1+1).
The method in accordance with the invention can be used both in compressors with only one compressor stage and in compressors with a plurality of compressor stages. In compressors which have a plurality of compressor stages, method steps (i) to (iv) are preferably carried out for at least two compressor stages, particularly preferably for all the compressor stages. It is thereby possible to localize damage in the sense that it can be associated with the respective compressor stage.
A further subject of the invention is an apparatus for identifying damage on a compressor having an intake side and a discharge side, wherein the apparatus comprises the following:

- sensors for acquiring measurement data of intake pressure (p1) and intake temperature (T1) on the intake side and end pressure (p2) and end temperature (T2) on the discharge side,
- a calculation unit which is adapted (a) to receive a predefined target variable as the input variable and/or to determine a target variable as a function of the measurement data, (b) to determine a comparison variable as a function of the measurement data, and (c) to carry out a comparison between the target variable and the comparison variable, and
- an output unit for outputting a signal which represents a measure of damage on the compressor.

The apparatus in accordance with the invention can be used both in compressors with only one compressor stage and in compressors with a plurality of compressor stages. In compressors which have a plurality of compressor stages, the apparatus in accordance with the invention preferably comprises sensors for acquiring measurement data on at least two compressor stages, particularly preferably on all the compressor stages, and the calculation unit is preferably adapted to carry out the calculation steps (a), (b) and (c) for at least two compressor stages, particularly preferably for all the compressor stages.
Further subjects of the invention are a computer program and a computer program product which are suitable for carrying out a method according to the invention.
The computer program in accordance with the invention contains program code which, when the computer program is executed on a suitable computer system, is suitable for carrying out the method in accordance with the invention.
The computer program product in accordance with the invention comprises a computer-readable medium and a computer program, stored on the computer-readable medium, with program code means which are suitable, when the computer program is run on a suitable computer system, for carrying out the method in accordance with the invention.
The subjects in accordance with the invention are suitable for detecting many kinds of damage on different machine elements of compressors. Examples are damage on valves, piston rings, packing glands, defective control devices, for example on intake valve lifters. The only requirement is that the damage manifests itself in the thermodynamic behavior of the compressor stage under consideration.
Unlike in the prior art, it is sufficient in the method in accordance with the invention and the apparatus in accordance with the invention to acquire and process only pressure and temperature data. Compared to vibration sensors, for example, the sensors required in accordance with the invention are inexpensive and in most cases are already provided as standard equipment of the compressors. Cost-intensive retrofitting with vibration sensors, for example, is not necessary. In addition, the determination of the target variable and comparison variable and the comparison thereof in steps (ii) to (iv) of the method in accordance with the invention require only the evaluation of a small number of mathematical equations and can be implemented at low expense.
The method in accordance with the invention and the apparatus in accordance with the invention allow possible damage such as wear, erosion or deposits to be identified early during operation of the compressor, so that measures which prevent component failure and unplanned machine downtime can be taken in good time.
The examples which follow illustrate the advantages of the method in accordance with the invention by means of actual measurement data from the operation of various compressors, but without limiting the invention to these examples.

EXAMPLE 1

The method in accordance with the invention was applied to the third stage of a compressor in order to identify possible damage occurring there. The compressor was a six-stage, two-crank reciprocating compressor which compresses carbon monoxide from 100 mbarg at about 5° C. to 35° C. to about 325 barg. The first stage of the compressor is equipped with backflow control, with which the delivery rate of the compressor can be set between about 70% and 100% of the maximum delivery rate. The third compressor stage comprises a double-acting piston inside a cylinder. The cylinder of the third stage is so constructed that the two compression chambers on the cover side and on the crank side draw in their gas from a common intake chamber and deliver into a common discharge chamber. The machine is equipped in each compression chamber with two plate valves on each of the intake side and the discharge side. Each stage of the machine is equipped with temperature sensors and pressure sensors on the piping on the intake side and on the discharge side.
FIG. 1 shows a extract from the operating data information system of the compressor for the third compressor stage in the period from September 2015 to October 2016. The following variables are shown in the diagram, wherein the left-hand scale indicates temperatures in degrees Celsius and the right-hand scale indicates pressures in barg:


top curve (solid)	end temperature (T2)
second curve from the top (dot-dashed)	calculated end temperature (T2b)
middle curve (dashed)	end pressure (p2)
second curve from the bottom (dashed)	I ntake pressure (p1)
bottom curve (dotted)	intake temperature (T1)

Measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side of the third stage and end pressure (p2) and end temperature (T2) on the discharge side of the third stage were continuously acquired and recorded. One measurement point per six hours was used to prepare the diagram. For better clarity, the measurement points have been faded out and joined together by interpolation.
As the target variable which represents a good state of the compressor, a calculated end temperature (T2b) of the third stage was determined as a function of the measurement data of the end pressure (p2), the intake pressure (p1) and the intake temperature (T1). The calculated end temperature (T2b) was determined according to an isentropic compression model including the isentropic exponent (κ) of the gas to be compressed and a correction factor (η). The isentropic exponent for carbon monoxide was set in the relevant pressure and temperature range at 1.4.
The calculated end temperature (T2b) was determined in accordance with the equation T2b=T1/η·(p2/p1){circumflex over ( )}(1−1/κ). The correction factor (η) was determined from historical data at the value 0.972.
For the comparison with the target variable, the measured end temperature (T2) was used as the comparison variable.
In the periods from the middle of April to the middle of August 2016, the end of August 2016 and from the beginning of October 2016, the compressor was not in operation. In the period from the beginning of September to the middle of November 2015, the calculated end temperature (target variable) and the measured end temperature (comparison variable) were almost identical. This led to the conclusion that the thermodynamic machine elements were completely intact. From the end of November 2015, first differences between the measured and the calculated end temperature appeared. On the basis of experience of earlier damage, a difference of about 5° C. with a pressure ratio (p2/p1) of about 2.5 gave reason to expect slight damage to the working valves which did not yet require an immediate response.
From January 2016, the difference became greater and at the end of January reached a level of 10° C. From experience, this difference indicated valve damage with spalling at the valve plates. Once the differences had again become significantly worse from the start of February 2016, it was decided to change the valves at the next opportunity. A planned stoppage for motor maintenance at the end of March was used to check the valves. Considerable spalling was thereby found at several intake valve plates. Once the valves had been changed, no difference between the calculated and the measured end temperature was observed following start-up in the middle of August 2016.
The method in accordance with the invention, on the basis of the apparatus in accordance with the invention, thus identified damage reliably and early during operation of the compressor. Comparing the comparison variable and the target variable not only provided information on whether damage was present, but also gave a measure of the severity of the damage. On the basis thereof, it was possible to make decisions about measures for preventing a potential component failure and unplanned machine downtime.

EXAMPLE 2

The method in accordance with the invention was applied to the first stage of a compressor in order to identify possible damage occurring there. The compressor was a seven-stage, two-crank reciprocating compressor which compresses carbon monoxide from 100 mbarg at about 5° C. to 35° C. to about 325 barg. The first stage of the compressor is equipped with backflow control, with which the delivery rate of the compressor can be set between about 70% and 100% of the maximum delivery rate. The first compressor stage comprises a double-acting piston inside a cylinder. The cylinder is so constructed that the two compression chambers on the cover side and on the crank side draw in their gas from a common intake chamber and deliver into a common discharge chamber. The machine is equipped in each compression chamber with three plate valves on each of the intake side and the discharge side. Each stage of the machine is equipped with temperature sensors and pressure sensors on the piping on the intake side and on the discharge side.
FIG. 2 shows a extract from the operating data information system of the compressor for the first compressor stage in the period from December 2017 to May 2018. The following variables are shown in the diagram, wherein the left-hand scale indicates temperatures in degrees Celsius and the right-hand scale indicates pressures in barg:


top curve (solid)	end temperature (T2)
second curve from the top (dot-dashed)	calculated end temperature (T2b)
third curve from the top (dashed)	end pressure (p2)
second curve from the bottom (dotted)	intake temperature (T1)
bottom curve (dashed)	intake pressure (p1)

Measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side of the first stage and end pressure (p2) and end temperature (T2) on the discharge side of the first stage were continuously acquired and recorded. One measurement point per six hours was used to prepare the diagram. For better clarity, the measurement points have been faded out and joined together by interpolation.
As the target variable which represents a good state of the compressor, a calculated end temperature (T2b) of the first stage was determined as a function of the measurement data of the end pressure (p2), the intake pressure (p1) and the intake temperature (T1). The calculated end temperature (T2b) was determined according to an isentropic compression model including the isentropic exponent (κ) of the gas to be compressed and a correction factor (η). The isentropic exponent for carbon monoxide was set in the relevant pressure and temperature range at 1.4.
The calculated end temperature (T2b) was determined in accordance with the equation T2b=T1/η·(p2/p1){circumflex over ( )}(1−1/κ). The correction factor (η) was determined in accordance with the equation η=a·T1+b·p2/p1+c, wherein the factors of the correction factor were adjusted by regression from historical measurement data of p2, p1 and T1 to a=0.0004702, b=0.06183 and c=0.644289.
For the comparison with the target variable, the measured end temperature (T2) was used as the comparison variable.
In the periods from the beginning to the middle of January 2018, the end of January 2018, the beginning of February 2018, the beginning of March 2018, the end of March 2018 to the beginning of April 2018 and the beginning to the middle of May 2018, the compressor was not in operation. In the periods in which the compressor was in operation, the calculated end temperature (target variable) and the measured end temperature (comparison variable) were almost identical. This led to the conclusion that the thermodynamic machine elements were completely intact. In the period under consideration, no damage was actually found on the compressor.
In this case too, the method in accordance with the invention, on the basis of the apparatus in accordance with the invention, is capable of identifying damage reliably and early during operation of the compressor.

EXAMPLE 3 AND COMPARATIVE EXAMPLE

The method in accordance with the invention was applied to a one-stage, double-acting, two-crank reciprocating compressor which compresses hydrogen from 25 barg at about 5° C. to 35° C. to about 40 barg. Both cylinders are each equipped with an intake line and a discharge line. The compression chambers on the cover side and on the crank side obtain their gas from a common intake chamber and deliver into a common discharge chamber. The machine is equipped in each compression chamber with an annular valve on each of the intake side and the discharge side. The valves on the intake side are each equipped with hydraulic backflow control for regulating the delivery rate. The machine is equipped with temperature sensors and pressure sensors on the piping on the intake side and on the discharge side.
The machine was further equipped with a monitoring device as is known from the prior art. This monitoring device comprises temperature sensors on the valve covers of the intake side and the discharge side, which sensors detect the outside temperature of the valve covers. As soon as the measured temperature is above a limit value of 50° C., an alarm is triggered, which indicates defective valves.
FIG. 3 shows a extract from the operating data information system of the compressor in the period from September 2017 to March 2018. The following variables are shown in the diagram, wherein the left-hand scale indicates temperatures in degrees Celsius and the right-hand scale indicates the pressure ratio (p2/p1) as a dimensionless number:


top curve (solid)	end temperature (T2)
second curve from the top (solid)	calculated end temperature (T2b)
third curve from the top (dotted)	pressure ratio (p2/p1)
second curve from the bottom (dashed)	valve cover temperatures 1 and 2
bottom curve (dot-dashed)	intake temperature (T1)

Measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side and end pressure (p2) and end temperature (T2) on the discharge side were continuously acquired and recorded. One measurement point per six hours was used to prepare the diagram. For better clarity, the measurement points have been faded out and joined together by interpolation.
As the target variable which represents a good state of the compressor, a calculated end temperature (T2b) was determined as a function of the measurement data of the end pressure (p2), the intake pressure (p1) and the intake temperature (T1). The calculated end temperature (T2b) was determined according to an isentropic compression model including the isentropic exponent (κ) of the gas to be compressed and a correction factor (η). The isentropic exponent for hydrogen was set in the relevant pressure and temperature range at 1.4.
The calculated end temperature (T2b) was determined in accordance with the equation T2b=T1/η·(p2/p1){circumflex over ( )}(1−1/κ). The correction factor (η) was adjusted on the basis of historical data to the value 0.975.
For the comparison with the target variable, the measured end temperature (T2) was used as the comparison variable.
In the periods from the middle of September to the beginning of October 2017, the end of October 2017 and the end of November 2017 to the end of January 2018, the compressor was not in operation. In October 2017, the calculated end temperature (target variable) and the measured end temperature (comparison variable) were almost identical. This led to the conclusion that the thermodynamic machine elements were completely intact. Following restarting at the end of October 2017, first differences between the measured and the calculated end temperature appeared. On the basis of experience of earlier damage, a difference of about 5° C. with a pressure ratio (p2/p1) of about 1.55 gave reason to expect damage of the working valves with minor spalling at the valve rings. The temperature sensors used for conventional monitoring on the valve covers still showed values at this time which were far below the alarm threshold of 50° C.
Following restarting at the end of January 2018, the difference became greater and in the following two months reached a level of 10° C. From experience, this difference at the low pressure ratio indicates significant valve damage with major spalling at the valve rings. Conventional monitoring did not signal any damage in this period either. The measurement values from the temperature sensors on the valve covers were even still below those in November 2017. In May 2018, the machine was taken out of operation again and the valves were checked. Valve rings with spalling were found on both valve plates.
In this case too, the method in accordance with the invention, based on the apparatus in accordance with the invention, thus detected damage reliably and early during operation of the compressor, whereas conventional monitoring by means of temperature measurement at valve covers gave no indication of possible damage.

Claims

1. A method for identifying damage on a compressor having an intake side and a discharge side, comprising the steps:

(i) acquiring measurement data of the measurement variables intake pressure (p1) and intake temperature (T1) on the intake side and end pressure (p2) and end temperature (T2) on the discharge side;

(ii) determining a calculated end temperature (T2b), a calculated intake temperature (T1b), a calculated end pressure (p2b) or a calculated intake pressure (p1b) as a target variable which represents a good state of the compressor, as a function of the measurement data of a maximum of three of the measurement variables (p1, T1, p2, T2);

(iii) determining a comparison variable from at least one of the measurement variables (p1, T1, p2, T2) not used in step (ii); and

(iv) comparing the comparison variable and the target variable as a measure of damage on the compressor;

characterized in that the target variable determined in step (ii) is determined according to an isentropic compression model including the isentropic exponent (κ) of the gas to be compressed and a correction factor (η), and the correction factor (η) is adjusted on the basis of measurement data.

2. The method as claimed in claim 1, wherein in step (ii) a calculated end temperature (T2b) is determined as the target variable as a function of the measurement data of the end pressure (p2), the intake pressure (p1) and the intake temperature (T1) in accordance with the equation

T2b=T1/η·(p2/p1){circumflex over ( )}(1−1/κ),

and in step (iii) the measured end temperature (T2) is determined as the comparison variable, wherein the correction factor η is calculated in accordance with the equation

η=a·T1+b·p2/p1+c

and the factors a, b and c are determined by regression from measurement data of p2, p1 and T1.

3. The method as claimed in claim 1, wherein in step (ii) a calculated intake temperature (T1b) is determined as the target variable as a function of the measurement data of the intake pressure (p1), the end pressure (p2) and the end temperature (T2) in accordance with the equation

T1b=T2·η·(p1/p2)−(1−1/κ),

and in step (iii) the measured intake temperature (T1) is determined as the comparison variable, wherein the correction factor η is calculated in accordance with the equation

η=a·T2+b·p1/p2+c

and the factors a, b and c are determined by regression from measurement data of end temperature (T2), intake pressure (p1) and end pressure (p2).

4. The method as claimed in claim 1, wherein in step (ii) a calculated end pressure (p2b) is determined as the target variable as a function of the measurement data of the end temperature (T2), the intake pressure (p1) and the intake temperature (T1) in accordance with the equation

p2b=p1·(η·T2/T1)−(κ/(κ−1)),

and in step (iii) the measured end pressure (p2) is determined as the comparison variable, wherein the correction factor η is calculated in accordance with the equation

η=a·p1+b·T2/T1+c

and the factors a, b and c are determined by regression from measurement data of intake temperature (T1), intake pressure (p1) and end temperature (T2).

5. The method as claimed in claim 1, wherein in step (ii) a calculated intake pressure (p1b) is determined as the target variable as a function of the measurement data of the intake temperature (T1), the end pressure (p2) and the end temperature (T2) in accordance with the equation

p1b=p2·(T1/T2/η){circumflex over ( )}(κ/(κ−1)),

and in step (iii) the measured intake pressure (p1) is determined as the comparison variable, wherein the correction factor η is calculated in accordance with the equation

η=a·p2+b·T1/T2+c

and the factors a, b and c are determined by regression from measurement data of intake temperature (T1), end temperature (T2) and end pressure (p2).

6. The method as claimed in claim 1, wherein the compressor has a plurality of compressor stages and method steps (i) to (iv) are carried out for at least two compressor stages.

7. A computer program product with program code which, when the computer program is executed on a suitable computer system, is suitable for carrying out a method as claimed in claim 1.

8. A computer program product having a computer-readable medium and a computer program, stored on the computer-readable medium, with program code means which are suitable, when the computer program is run on a suitable computer system, for carrying out the method as claimed in claim 1.

9. The method as claimed in claim 1, wherein the compressor has a plurality of compressor stages and method steps (i) to (iv) are carried out for all the compressor stages.