The present invention relates to gas compression. More particularly, the invention relates to a method for the compression of gas and most preferably carbon dioxide, for example as part of a carbon capture and storage scheme.

Carbon dioxide Capture and Storage (CCS) has been proposed as an approach for reducing greenhouse gas emissions from the use of fossil fuels.

CCS involves capturing the carbon dioxide produced at sites such as fossil-fuel power generation plants, compressing it into a dense liquid state (typically at a pressure ratio of up to 200) and then transporting it via a pipeline system to storage facilities at which the carbon dioxide is stored in such a way that its release into the atmosphere is prevented. For example, storage facilities may sequestrate the carbon dioxide by injection, in its compressed liquid form, into oil- and gas-field formations, coal seams or saline water-bearing formations. Alternatively, the carbon dioxide may be sequestrated by reaction with magnesium- and calcium- containing minerals to form stable carbonates.

Power generation plants are normally operated so as to produce electrical power at a rate which is directly related to the power-demand at any particular point in time. As the demand for electrical power varies over time, the amount of power produced by the power plant must also vary, and this variation causes a corresponding variation in the amount of waste carbon dioxide which is produced by the plant. This complicates the process of compressing the carbon dioxide for sequestration, as it is difficult to operate compressors effectively in response to a constantly changing supply of carbon dioxide.

Using currently available technology, maintaining the drive power of a CCS compressor to acceptable levels requires the overall compression to be divided into a number of discrete steps, or so-called "sections", with intercooling provided between successive sections.

For a modern coal-fired power station, it has been found that the power required to drive the compressor can typically represent as much as 8% of the total power output of the station, even with a compressor having 8 intercooled steps. This represents a significant parasitic load on the power station.

Previously proposed compressor arrangements of the type mentioned above are typically configured such that each discrete section comprises one or more compressor stages, which are mostly of centrifugal configuration. The overall compressor system is often referred to as a compressor "train". The large target increase in carbon dioxide pressure through the compressor train requires the rear (downstream) sections to be significantly smaller than the front (upstream) sections. The optimum rotational speed of the rear compressor sections is thus higher than the optimum speed of the front sections. This speed differential can be achieved in various ways, the most common of which uses an "integrally geared" layout in which single centrifugal stages are mounted around a gearbox and driven at different, but fixed, speeds to one another. This type of layout requires only a single input drive and so minimizes cost.

The pipeline network downstream of the compressor system operates at a pressure level which is generally independent of the power and carbon dioxide output of an individual power station. Although the pipeline pressure may, in reality, vary slightly over time (for example as storage facilities fill up over time, or as a function of the carbon dioxide output from other power stations on the same network), it can be considered to remain essentially constant relative to the more significantly varying carbon dioxide output of an individual power station, even when the power plant is only operating at very low power.

A single conventional compressor train is typically able to maintain a constant pressure ratio only down to approximately 75% of its maximum flow rate. In contrast, a power station might sometimes be required to operate at power levels as low as 25% of its maximum power. Under such operating conditions, the power station typically produces only 33% of the flow of carbon dioxide which is produced when the power station is operating at full power.

Previously proposed carbon dioxide compressor systems for use in CCS are designed in such a way as to be optimised for full power operating conditions. Accordingly, they are not well suited to operation with reduced flows of carbon dioxide arising from low power running of a power station. In order to address this, it has been proposed to operate such compressor trains under low power conditions by recirculating carbon dioxide from the outlet of the compressor train back to the inlet of the compressor train. However, when operated in accordance with this type of regime, even at 25% of maximum plant power, the power required to drive the compressor train is still typically as high as 80% of its full power level. At such low plant power, the parasitic load of the compressor on the power station becomes very significant indeed, thereby significantly reducing the overall efficiency of the power generation scheme.

The Applicant's co-pending UK Patent Application 0919771 .6, the details of which are hereby incorporated by reference, proposes a compressor train which is divided into three or more discrete portions, wherein each portion is independently driven so as to allow variable relative rotational speeds there-between. The present invention represents the result of further development work into such a concept.

It is an object of the present invention to provide an improved apparatus and method for compressing gas, such as carbon dioxide.

According to the present invention, there is provided a system for compressing a gaseous flow comprising: a compressor train having an upstream rotor portion and a downstream rotor portion arranged in flow series with said upstream portion, each rotor portion comprising one or more compressor stages and said downstream rotor potion being arranged to rotate at a variable relative speed to that of the upstream rotor portion; sensing equipment arranged to determine one or more flow conditions for the compressor train; and, a controller arranged to receive data indicative of said one or more sensed flow conditions, wherein the controller is arranged to set the rotational speed of the upstream rotor portion based upon the pressure or flow rate of gas upstream of said compressor train and to set operational parameters for said downstream portion so as to maintain a desired surge margin for said upstream rotor portion.

In one embodiment, the rotational speed of the upstream and downstream rotor portions is independently controllable.

The sensing equipment may be used to sense a current operational state, condition or signature for the compressor train and/or either portion thereof. The operational parameters for the downstream rotor portion may be determined based upon a comparison between the current operational state and a desired operational surge margin for the upstream and/or downstream rotor portions.

The flow rate of gas directed into the upstream compressor portion may be determined from the pressure of the gas as measured in the region of the inlet of the upstream compressor portion. Conveniently, the speed of the upstream compressor is reduced in response to a reduction in the flow rate of gas directed into it, the upstream compressor thus producing a reduced outlet flow for direction into the downstream compressor portion. The speed at which the downstream compressor portion is driven may also be varied in dependence on the flow rate of compressed gas directed into it.

According to a preferred embodiment, the system further comprises intercooling apparatus arranged to cool the gas flow between one or more stages of the compressor train. An intercooler may be located between each adjacent compressor stage of the upstream and/or downstream rotor portion.

In one embodiment, the compressor train comprises a gas recirculation conduit arranged to selectively bleed gas from the downstream rotor portion. The gas recirculation conduit may provide a selectively operable gas path between the exit and the inlet of the downstream rotor portion. The gas recirculation conduit may comprise a valve arrangement which is selectively operable under the control of the controller.

The gas recirculation conduit may be selectively used in dependence upon, at least in part, the flow rate of gas directed into the upstream compressor portion. The gas recirculation circuit may be used when a currently sensed compressor operation fails to meet a predetermined surge margin threshold for the upstream rotor portion.

The controller may comprise machine readable instructions to process the received sensor data indicative of said one or more sensed flow conditions and to compare said sensed flow conditions against a predetermined schedule of desired flow conditions for said compressor train.

The controller may be arranged to determine whether a predetermined surge flow margin threshold criterion for the upstream and/or downstream rotor portion has been met by one or more sensed flow conditions.

In one embodiment, the controller sets operational parameters comprising any or any combination of rotational speed, intercooler temperature and/or gas recirculation conduit operation so as to maintain a desired surge margin for said compressor train or portion thereof.

The controller may comprise machine readable instructions to apply an iterative control scheme to converge towards determination of desired operational parameters for said compressor train. The controller may iterate steps of receiving data indicative of said one or more sensed flow conditions; comparing said sensed flow conditions against a predetermined schedule of desired flow conditions; and, outputting control signals to said compressor train to achieve said desired surge margin for at least said upstream rotor portion.

According to a second aspect of the present invention, there is provided a gas sequestration process comprising a system according to the first aspect.

According to a third aspect of the invention, there is provided a carbon dioxide capture system according to the first or second aspects.

In such a method, the carbon dioxide which is directed into the compressor train may obtained directly from a carbon dioxide producing site such as a power generation plant configured to operate at variable power. In such an arrangement, it is envisaged that the flow rate of carbon dioxide directed into the upstream compressor portion will vary in dependence on the output power of the power generation plant. Accordingly, the flow rate of carbon dioxide directed into the compressor train will be reduced when the output power of the power plant is reduced. According to a fourth aspect of the invention, there is provided a method of compressing a gas product using a compressor train having discrete upstream and downstream rotor portions arranged successively in flow series, the method comprising: performing a first compression step by driving the upstream compressor portion to compress the gas based upon the pressure or flow rate of gas upstream of said compressor train; and, performing a second compression step by directing the compressed flow from the upstream compressor portion into the downstream compressor portion and controlling the downstream compressor portion to further compress the gas under operational parameters which are determined so as to achieve a desired surge margin for said upstream rotor portion.

The above-mentioned method is most preferably performed on carbon dioxide, and may thus form part of a carbon dioxide sequestration scheme.

Any of the preferable features described above in relation to any one aspect may be applied to any other aspect of the invention wherever it is practicable so to do.

So that the invention may be more readily understood, and so that further features thereof may be appreciated, embodiments of the invention will now be described by way of example with reference to the accompanying drawings in which:

Figure 1 is a schematic diagram showing a carbon dioxide sequestration system;

Figure 2 is a graph of pressure against enthalpy for a C0 ₂ compressor;

Figure 3 is a schematic diagram showing a compressor train for use with the method of the present invention, comprising two discrete compressor portions;

Figure 4 is an exemplary compressor map representing the performance characteristics of the first, upstream, compressor portion of the compressor train shown in figure 3; and,

Figure 5 is a flow diagram of upper level steps performed in a method or system according to one aspect of the present invention.

Turning now to consider the drawings in more detail, figure 1 shows the general arrangement of a carbon dioxide sequestration system. The system is shown to include two separate fossil fuel power plants 1 which produce unwanted carbon dioxide, but which are specifically configured to capture their carbon dioxide. It should be appreciated, however, that the system could include only a single power plant 1 , or indeed three or even more power plants.

The captured carbon dioxide produced by the power plants 1 is directed into a pipeline network 2 which serves to transport the carbon dioxide to a remote storage facility 3. At the storage facility, the carbon dioxide is injected into a formation 4, such as a geological formation, which is an example of a carbon dioxide storage medium. Each power plant has a compressor arrangement 5 which pressurizes the captured carbon dioxide gas before it is directed into the pipeline network. As already indicated, the compressor arrangements 5 are each configured to compress the carbon dioxide by a pressure ratio (PR) of up to 200, and assuming they are fed with carbon dioxide gas arising from their respective power plants 1 at a pressure of 1 bar, are thus effective to increase the pressure of the carbon dioxide to approximately 200 bar, thereby rendering it a highly compressed liquid.

Each compressor arrangement 5 has a control system 6 which is configured to vary the power of the compressor in response to variations in the output power of the respective plant, as measured by a measuring system 7. Each control system 6 thus acts upon a signal received by the respective measuring system 7 so that as the power generated by the respective plant varies, the power of the compressor also varies. Because the carbon dioxide production rate of the plant is related to the power generated, this has the effect of matching the operating point of the compressor arrangement to the respective plant's carbon dioxide output, thereby allowing the compressor arrangement to continue to operate effectively despite variations in the flow of carbon dioxide directed into it from the power plant.

For example, in the case of a 2GW power plant operating at its maximum power output, the compressor arrangement 5 can be set to compress the maximum amount of carbon dioxide produced by the plant. However, when the power plant reduces its capacity down to 100 MW, for example when power demand falls, the compressor power will also decrease. If the power of the compressor arrangement were not capable of being varied when the power plant output changes in this way, the compressor would at times be improperly matched to the carbon dioxide production rate. This could lead to compressor failure or stoppage, and/or accidental release of carbon dioxide.

The detection of carbon dioxide outputted by the plant could be monitored by the power outputted by the power plant. The skilled person will be aware of other measurable parameters in a power plant which might give an indication of the amount of carbon dioxide being outputted from the power plant. The measuring system can include a carbon dioxide detector which measures more directly the amount of carbon dioxide produced by the plant. For example, suitable measuring signals can be obtained from power station boilers, air intake flow sensors, flue gas flow rate sensors etc. Alternatively or additionally the measuring system can determine the amount of carbon dioxide produced by the plant from control signals used to operate the plant.

Measuring the detection of electrical power provided by the power plant can be advantageous as it can be used to predict the demand on the compression system. Further, monitoring electrical power in a grid network could be used to provide an indication of the carbon dioxide required from a sequestration and storage network which receives power from a number of power plants.

Furthermore, the properties of C02 become highly non-linear near the gas/liquid phase boundary at elevated pressures. Density and hence flow velocities and incidence angles, become very difficult to predict in this region. Figure 2 shows a pressure against enthalpy chart for an exemplary 10-stage integrally geared compressor, showing the C02 phase boundary. It is generally desired to provide intercooling between all compressor stages in order to achieve best efficiency. However the difficulties in predicting behaviour in the vicinity of the C02 phase boundary have prevented existing integrally geared compressor manufacturers from intercooling between at least the rearmost stages. This substantially increases the required drive power versus a fully intercooled solution.

Figure 3 shows the configuration of a preferred compressor arrangement 5 in accordance with the present invention in more detail. The compressor arrangement 5 is provided in the form of a so-called compressor train, and comprises at least two discrete compressor portions, namely; an upstream portion 8, and a downstream portion 10. The portions 8 and 10 are arranged successively in flow series. Accordingly, the flow of carbon dioxide gas 1 1 produced by the power plant 1 and which is to be compressed, is directed into the upstream compressor portion 8 within which an initial compression step occurs. The resulting compressed flow 12 then exits the upstream compressor portion 8 and is directed into the downstream compressor portion 10 for further compression, as will be explained in more detail hereinafter.

Each of the discrete compressor portions 8 and 10 may comprise either a single compressor stage or may comprise a plurality of compressor stages connected in series, as depicted in the example of figure 3, in which each compressor portion has four compressor stages. Each compressor portion may have a different number of stages from one or more other compressor portions.

The multi-stage compressor arrangement incorporates an intercooler 13 between each successive pair of stages, as is generally known in the art. According to different embodiments, intercoolers may be located between only some or all stages of each compressor portion and/or between the compressor portions 8, 10 themselves.

Each discrete compressor portion 8, 10 is driven independently, in the sense that each may be driven at a different speed relative to the, or each, other portion. This also permits the operational speed of each compressor portion to be varied independently during operation of the compressor train, as will be explained in more detail hereinafter. In the particular arrangement of figure 3, each compressor portion 8, 10 is shown in connection with a respective electric motor 14, 15, the motors being arranged to drive each compressor at a variable speed. The motors typically draw their electrical power from the power plant 1 .

As will be appreciated by the skilled person, the compressor arrangement 5 comprises sensing equipment by way of a plurality of sensors 9 for measuring operational parameters such as pressure, flow rate and temperature. Such sensors may be located between each, or selected, stages of compressor portions as well as at the inlet and outlet of the overall compressor arrangement 5.

Sensor readings are fed to a controller 19, which comprises on or more processors arranged to receive sensor reading data and to process said received data according to one or more algorithms. The supervisory controller comprises one or more modules of machine readable code/instructions in order to implement a control strategy as will be described in further detail below. The controller controls operation of the electric motors 14, 15 in order to control the rotational speed of the compressor portions 8 and 10. The controller also controls operation of the intercoolers 13 so as to control the degree of cooling undertaken between compressor stages. This may be achieved by controlling delivery temperatures to the intercoolers 13. Typically the controller has independent control of the intercoolers 13 for the upstream 8 and downstream 10 compressor portions and/or individual compressor pairs thereof.

Figure 4 illustrates an exemplary compressor map representative of the performance characteristics of the upstream compressor portion 8. The map comprises plots of pressure ratio (PR) and isentropic efficiency against the mass flow rate through the compressor portion, as is conventional. The particular numerical values appearing on the map are not intended to be restrictive, and are provided simply by way of example. Accordingly the details discussed in relation to figure 4 should be interpreted as being applicable to a wide variety of compressor configurations according to the present invention.

Within the compressor map, the surge line is shown at 16, which may otherwise be referred to as the stall line, and which is indicative of a pressure ratio at a given flow rate, above which the flow becomes unstable. Accordingly a surge margin is defined such that the maximum permitted operating point of the compressor is maintained sufficiently below the surge margin. It will be understood that a greater surge margin generally indicates a safer compressor operation but to the detriment of the potential operational efficiency of the compressor. Reference point 17 shown on the compressor map represents a proposed maximum-flow operating point for the upstream compressor portion 8, and thus represents a suitable operating speed for the compressor portion 8 when the maximum possible flow rate of carbon dioxide 1 1 directed into it. The maximum-flow operating point 17 thus corresponds to the maximum (i.e 100%) power output condition of the power plant 1 . As will be appreciated, the maximum-flow operating point 17 is carefully selected so as to correspond substantially to a peak efficiency point 18, and to provide a suitable surge margin, falling below the surge line 16.

The compressed flow 12 exiting the upstream compressor portion 8 is fed directly into the compressor portion 10. The downstream compressor portion 10 is thus operated at a speed appropriate to give a suitable surge margin having regard to the surge line for the downstream compressor portion. Accordingly the operational speed of the downstream compressor portion 10 will typically be different to that of the upstream portion 8.

A reduction in the output power of the power plant 1 will have the effect of reducing the flow rate of carbon dioxide gas produced by the plant and thus directed into the compressor train as input flow 1 1. This results in a reduced-flow operating point which corresponds to a reduced power output condition (for example 25% maximum power) of the power plant 1 . As will be appreciated, the reduced-flow operating point is again carefully selected to correspond to a high efficiency level, and to provide a suitable surge margin, falling below the surge line 16.

For the reduced flow condition, this necessitates a reduced operating speed of the upstream compressor portion 8. The upstream compressor portion 8 is thus operated such that its speed is allowed to reduce in response to a reduction of the flow rate of gas 1 1 directed into it (which corresponds to a reduction in output power of the plant 1 ).

The volume flow ratio between the reduced flow condition and the maximum flow condition of the upstream compressor portion 8 can be calculated as follows:

Vol. flow ratio = (M _red /PR _red ) / (M _max / PR _max )

Where: M _red represents the reduced mass flow rate

PRr _ed represents the reduced pressure ratio

M _max represents the maximum mass flow rate corresponding to 100 % plant power, and

Rm _ax represents the maximum pressure ratio

Using the equation above for each of the upstream and downstream portions, it can be demonstrated that responding to varying inlet flow 1 1 conditions, the relative mass and volumetric flow rates form the upstream portion 8 will vary. This means that the output 12 from the upstream compressor portion 8 will be less dense at a reduced flow condition when compared to the maximum flow condition. By operating the upstream compressor portion 8 on a schedule of reducing speed in response to a reduction in plant power (and thus corresponding to a reduction in input flow rate; for example as determined from the pressure of gas measured at the inlet of the upstream compressor portion 8), the variation in volume flow directed into the midstream portion 9 is reduced.

A reduced exit flow 12 form upstream compressor portion 8 arising from the reduced power operation of the upstream compressor portion 8, directed into the downstream compressor portion 10, will result in a reduced inlet flow condition on a compressor map for the mid-stream portion 10. Under such reduced flow conditions, the portion 10 is thus also operated at a reduced speed appropriate to retain a suitable surge margin having regard to its surge line. This results in a reduction in polytropic head across the downstream compressor portion 10.

The downstream compressor portion 10 is thus also operated on a schedule of reducing speed in response to a reduction in plant power (corresponding to a reduction in input flow rate 1 1 directed into the upstream compressor portion 8, and thus giving rise to a reduced flow rate 12).

It is proposed that the upstream and downstream compressor portions 8, 10 will be designed to operate at maximum power conditions such as to increase the pressure of the carbon dioxide gas to a target of approximately 200 bar, without the need for further downstream compression. However, as the power and carbon dioxide output of the plant 1 both reduce such that the upstream and downstream compressor portions 8, 10 are operated at independently reduced speeds as explained above, the overall pressure ratio applied by the two portions in combination will fall.

The inventor has proposed a control scheme for operation of the compressor 5 in order to independently drive only two portions of the compressor train, so as to address variable inlet and discharge conditions in a manner which is power efficient, particularly at maximum flow conditions, whilst retaining desired surge margins. An important factor in the control methodology is to substantially maximize the degree of inter-cooling so as to cool each compressor section or portion to as low a temperature as possible.

The proposed control solution is described in further detail below with reference to figure 5. The following process is described as a proposed hierarchy, such that each stage is dependent on the result of the previous stage. However it is to be noted that the step of setting or determining suitable operating conditions may differ in time from the actual implementation or achievement of such conditions. The stages defined below may or may not accommodate such a time delay between the setting and implementation of desirous operating conditions dependent on the type of control scheme to be employed. Furthermore it will be appreciated that each stage described below is discussed at a high level and

At stage 20, the desired rotational speed of the upstream rotor portion is determined. The suitable rotational speed for the upstream rotor portion 8 is based on the pressure of the inlet flow 1 1 into the compressor arrangement 5. In this embodiment, the upstream rotor speed is also determined in dependence upon the operational level or power output of the power plant. Another conventional indicator of the expected rate of flow or output of C02 from the power plant 1 may be used.

A predetermined schedule or chart of compressor inlet against the power plant level may be used to indicate a desirous state of operation of the system. Accordingly the rotational speed of the front portion 8 may be set in dependence upon such a schedule. The control of the compressor portion 8 can then be implemented based upon a comparison between the current sensed inlet pressure and the desired inlet pressure for the given power plant conditions. The rotor portion 8 can then be accelerated or decelerated to either decrease or increase compressor inlet pressure respectively as required.

At process 20, the exit pressure from the C02 capture process may be used in addition to, or instead of the power plant operational level. In normal operation, in either embodiment, it is likely to be required to maintain this operating ratio or pressure gradient as substantially constant.

Process stage 21 involves determining the desired rotational speed of the rear compressor portion 10. This is achieved by setting the rear rotor portion 10 to a speed which maintains the desirous surge margin for the front portion 8.

The rotational speed of the rear portion 10 will effect the pressure gradient across the front portion 8 (in particular the pressure at exit flow 12) and thus the acceleration or deceleration of the rear portion 10 can be used to set the pressure ratio for a given flow rate as per the graph of figure 4. The downstream portion 10 rotational speed is typically set so as to maintain the upstream compressor portion 8 at or slightly below the desired operating surge margin. The actual surge margin applied may be specific to the compressor configuration and operating conditions. It may be a constant or variable percentage offset from the surge line 16 and may be determined based (for example) on a desired offset from the onset of surge flow at an optimal operating pressure ratio for the compressor. When applied as a variable percentage offset, the margin is typically set as being proportional to the flow rate. In such an example, the offset would tend to zero for a zero flow rate condition.

For example, with reference to figure 4, a margin of 10% may be applied at the optimal operating pressure ratio for the front portion 8 which may be applied as an offset from the surge line 16 so as to define a control line 21 which runs beneath the surge line.

The setting of the desired downstream rotor portion 10 in this embodiment is based upon data which may be captured in a table (or other matrix) or chart showing the effect of the downstream rotor speed on the upstream portion pressure ratio for given flow conditions. The data may be empirical or else predicted or otherwise prescribed for the specific configuration of compressor 5. Using such information, current pressure measurements from sensors 9 can be compared to the recorded correlations in order to determine a suitable rotational speed for the downstream rotor portion 10. As a result of a comparison with the current rotor speed, the downstream portion 10 speed may either be maintained, increased or decreased as necessary.

At stage 22, the intercooler delivery temperatures are determined. Setting of the intercooler operation is based at least initially on a predetermined schedule. The schedule may be based upon data which may be captured in a table (or other matrix) or chart showing the effect of the intercooler operation on the upstream portion and/or downstream portion pressure ratio for given flow conditions. The data may be empirical or else predicted or otherwise prescribed for the specific configuration of compressor 5. After startup (i.e. during normal operation), current pressure measurements from sensors 9 can be compared to the recorded correlations in order to determine a suitable intercooler delivery temperatures for the upstream and/or downstream portions 8, 10.

The setting of the intercooler temperatures may additionally or alternatively be based upon mixture properties for the gas stream entering the compressor, which may be determined using conventional sensing equipment in the gas stream.

Due to the very high delivery pressures (approximately 50 to 250 bar C02 discharge pressures) and potential consequence for the upstream plant, it is important to maintain acceptable surge margins. The rear portion is potentially most at risk from adverse pressure variations due to downstream pressure changes, and intercooling in the region of C02 phase boundary as shown in figure 2 which causes non-linear properties and unpredictable behaviour.

At stage 23, a check is performed to determine whether the predetermined surge margin conditions have been met by the operation of the compressor 5, including the independently controllable front and rear portions thereof. For this purpose there may be defined at least a lower surge margin threshold to define a minimum margin from the surge line which is acceptable. Preferably there is an upper and lower surge margin threshold. In the even that the pressure ratio for the gas flow rate is found to lie within the acceptable threshold, then further action to protect the surge margins is not required.

The process may then iterate steps 20 to 23 (possibly with a time delay between iterations), in order to maintain a watch over the continuing compressor operation. In such iterations, the steps may allow for recalculation of upstream and/or downstream rotor speeds and/or intercooler temperatures. This may be suitable for dynamic operating conditions (for example, during start-up or during changes to pressure, temperature or flow rate changes at the compressor inlet/outlet). In this regard, the controller may be connected to additional control or sensing equipment for the power plant such that the controller has advance warning of ensuing changes in operating conditions.

Additionally or alternatively, the controller may simply check the current values against the previously determined values if a substantially constant, steady-state mode of operation is required.

In the event that the surge margin threshold(s) are not met at stage 23, then the controller determines action required to control operation of the compressor and thereby restore acceptable surge margins at 24.

In order to achieve the action required at step 24, the compressor arrangement 5 comprises a fast-acting recirculating bleed 25, as shown in figure 3, which provides a gas path from the discharge to the inlet of the rear portion 10. The selective operation of the recirculation gas path 25 is achieved by a valve arrangement 27, which is shown symbolically as a simple valve in figure 3. It will be appreciated that a conventional electromechanically operated valve mechanism will typically be employed. It is desirable to selectively recirculate bleed flow for the rear portion 10 only in order to avoid pressure perturbations of the upstream plant. Also, any steady state, long term bleed would represent a significant power penalty.

However it may also be possible to provide an embodiment in which bleed flow is recirculated for both the front and rear compressor portions either individually or else in combination by providing suitable conduit and valve arrangements under the control of controller 19. Additionally or alternatively a recirculating bleed arrangement may be provided in respect of only one, or else a subset of compressor stages, within the front 8 and/or rear 10 portions. This may provide the possibility to recirculate around any individual stage, combination of stages or section where sensor readings, such as unsteady fluctuations in pressures, indicate a problem.

The operation of the valve to allow recirculation of bleed flow from the outlet to the inlet of the downstream compressor portion 10 would help to restore a desirable pressure balance. Intercooler temperatures in the rear portion would be increased by bypass of either the C02 or the coolant. This would increase the surge margin of the section downstream, at slight expense to that of the section upstream.

In addition to established pressure correlations, lack of surge margin could be detected by a variety of readings, such as unsteady pressure readings, vibration, other correlations based on compressor speed.

Once corrective action has been taken at stage 24 in figure 5, the surge margins are subsequently checked to ensure correct operation. The stages 23 and 24 may thus be repeated as necessary to restore acceptable operation.

An iterative control scheme is employed. Accordingly an initial pass of the process in figure 5, in which initial settings are applied, is followed by subsequent passes in which incremental changes may be made such that the settings converge to an ultimate set of desired operating variables. Thus if the power plant is operating in steady state, so too should the compressor converge to a steady state operation. Modern self tuning control methods such as genetic algorithms could be applied to this end.

The control strategy described above allows a compressor train to absorb minimum power at full load in this demanding duty, without excessive footprint or unit cost. Prior art in C02 compressors are not designed to reduce power to this extent and are typically optimised for other duties.

There are many available variations in compressor types, number of portions / sections, instrumentation readings, etc which could be applied to the above described embodiment. For part of the train could be at fixed speed, or the speeds arranged to vary in a manner that is not fully independent such that one compressor portion can be varied as a function of variation of the other compressor portion.

In another embodiment, the downstream compressor portion may comprise at least one variable-geometry compressor, in which case the adjustment of the geometry of the, or each, said compressor may be applied to affect the operational characteristics of the compressor train.

Whilst the invention has been described above with particular reference to its use in the compression of carbon dioxide as part of a CCS scheme, it is to be noted that the invention could also be used for the compression of other gases, and so is not to be interpreted as restricted to its application to carbon dioxide.

When used in this specification and claims, the terms "comprises" and "comprising" and variations thereof mean that the specified features, steps or integers are included. The terms are not to be interpreted to exclude the presence of other features, steps or integers.

The features disclosed in the foregoing description, or in the following claims, or in the accompanying drawings, expressed in their specific forms or in terms of a means for performing the disclosed function, or a method or process for obtaining the disclosed results, as appropriate, may, separately, or in any combination of such features, be utilised for realising the invention in diverse forms thereof.