EP0431287B1 - Process for optimised operation of two or more compressors in parallel or series operation - Google Patents

Description

The invention relates to a method according to the preamble of patent claim 1.

From US-A-3 527 059 a control method for compressors of a cooling device operated in parallel is known. In this method, the drive powers of the compressors are regulated as a function of the respectively recorded values for the individual flows of cooling medium through the compressors in such a way that all flows are or remain the same among one another. Provided that the same compressors are used, a uniform load or performance of the compressors is achieved. The method is not suitable for the joint use of different compressors; Even a uniform load on the compressors does not necessarily result in optimal operation in economic terms.

A similar method is known from FR-A-2 108 039, which is used to control parallel, electromotive driven compressors of a cooling device. The aim of this method is also to equalize the load or performance of the individual compressors. For this purpose, the values for the Current consumption of the drive motors of the compressors is recorded and compared with one another, and control signals are derived therefrom which regulate the motors in such a way that their current consumption is or remains the same among one another. The aforementioned disadvantages occur here in the same way.

A method for operating two compressors in series is known from US-A-4,255,089. This method distributes the load to the two compressors on the basis of fixed, predetermined storage data as a function of a control signal which represents the requirements placed by a downstream system or process. The required storage data is shown here in the form of a series of sequences of linear functions. The disadvantage of this method is that it requires very large data storage capacities and, because of the large number of required memory accesses, it is relatively slow and can therefore hardly be used for more than two compressors. Also, the method cannot react to changes in the compressors after the storage data have been determined, such as that which occurs, for. B. occur as a result of aging or pollution or conversions. To take this into account, a very complex generation and input of new storage data would be required.

In the article "Advances in Instrumentation, Volume 31, No. 1, 1976, pages 585.1 to 585.7, ISA AC; AE Nisenfeld et al .: Control of Parallel Compressors", the problem of parallel operation of compressors is discussed. The possibilities for optimizing operation are given on the one hand by an even load distribution, as previously mentioned, or by maximizing the overall efficiency. To implement the second option, a dynamic simulation of the compressor parallel operation in one Hybrid computers proposed, but without more detailed information or a specific technical teaching are given.

Finally, EP 0 132 487 B1 discloses a method for operating at least two turbo compressors connected in parallel. The essence of this method is to regulate the setting of the compressors among one another by means of load distribution regulators in such a way that the operating points of all compressors are each at the same distance from their blow-off line. In this case, only one of the compressors is controlled by an assigned pressure regulator, while further compressors are tracked via load distribution regulators. A disadvantage of this method is that it can only guarantee optimal operation for compressors that are identical to one another, but not for unequal compressors. In addition, the optimization that can be achieved with this method is only an approximation and not the absolute optimum.

It is therefore the task of specifying a method of the type mentioned at the outset which, with little technical and time expenditure, ensures economically optimized operation of two or more compressors connected in parallel or in series, including unequal ones.

This object is achieved according to the invention by a method of the type mentioned at the outset with the characterizing features of patent claim 1.

With the method according to the invention, the compressors operated in parallel or in series can be moved during operation into an optimum working point combination with regard to the selected input parameter, it being irrelevant whether the same or different compressors can be operated together. Differences can arise from different sizes or types, but also from different operating times or construction tolerances. Since the process only takes a very short time for each of its cycles, the optimization is practically continuous and without delay. The data required for the method, here operating parameters, are usually continuously recorded anyway, so that no additional effort is required. The incremental change in the individual volume flows or individual pressure ratios can also be easily achieved by appropriate adjustment of the manipulated variables, the required amount of adjustment of the manipulated variables being evident from the technical data of the individual compressors. These technical data are known to the operator for each compressor and are usually available in the form of maps. Is carried out by the higher-level control an adjustment of the manipulated variables, z. B. due to changed requirements of the downstream process, then the inventive method immediately leads the compressors back into a new optimal working point combination. Because of the short time of the individual process cycles, systems with more than two compressors can be optimized in their mode of operation with sufficient speed by forming all possible pairs of compressors. The term manipulated variable here means, in particular, a specific speed, guide vane position or throttle position of the compressor, the selection of the specific variable depending on the type of power control of the compressor, which is determined in a technical and constructive manner. The manipulated variable is often understood to be the command from a controller for a subsequent actuator influencing the speed, vane position or throttle position; if no transmission errors occur here, the manipulated variable are in the sense defined above and the positioning command is also identical. If there are transmission errors, they can easily be determined and appropriate corrections can be applied to eliminate their influence.

The compressors can be turbo or screw compressors, for example, which are driven by a drive machine, e.g. B. electric motor or turbine. The parameters that are essential in the respective application, eg. B. the power consumption or the operating costs of the compressors can be used. The power consumption or operating costs of both the drive units of the compressors and additional auxiliary units, such as coolant pumps, condensate pumps when driven by turbines, transformers with electric drives, etc., can be taken into account without any problems, since the corresponding technical data are also known to the operator or can be easily determined .

A first development of the method is specified in claim 2, it being essential here that some of the method steps are not carried out on the compressors themselves but in a simulation. It is hereby achieved that the number of adjustment processes on the compressors themselves is significantly reduced and is limited to the adjustments which bring about the desired effect, while unnecessary adjustments, ie those with an undesirable effect, do not reach the compressors. In addition, the process of each individual process cycle is significantly accelerated because the consequences of manipulated variable changes can be determined more quickly in the simulation than on real compressors. The prerequisite for this is that the input parameter map is stored, which is technically and mathematically not a problem. A number of characteristic curves suffice here save constant manipulated variable; Intermediate values between the individual characteristics can then be determined with sufficient accuracy by interpolation.

Claim 3 specifies a specific embodiment of the method for the parallel operation of two compressors in the form of a sequence of individual method steps, from which a preferred method sequence for this application emerges in detail.

Some additional process configurations are provided specifically for parallel operation, which are explained below.

In order to ensure, on the one hand, that the procedure is as fast as possible and, on the other hand, that the optimum of the input parameter sum is determined as precisely as possible, it is provided according to claim 4 that the increments become smaller when the optimum is approached. This results in a faster process with a greater distance from the optimum and a slower but finer process of the compressor operating points with decreasing distance from the optimum.

Since in practice with two or more compressors operated in parallel on the pressure side due to different line lengths and routing until the junction in the downstream process, different pressure losses occur and since these are also flow-dependent, an individual determination of the pressure ratio of each compressor is provided according to claim 5. As a result, line influences can no longer affect the result of the optimization of the operating mode.

When operating compressors, it is still relatively common for the system pressure and thus the pressure ratio to change due to changed process requirements the compressors becomes necessary. The method development set out in claim 6 serves to take account of such changes in the method without any delay. The associated new values for the manipulated variables are then automatically determined within the course of the method and the compressors are moved to the optimum operating points for the new process requirements.

In addition to changes in pressure, there may also be changed process requirements with regard to the volume flow, the consideration of which in the method is specified in claim 7. With this design of the method, the present method can also take over parts of the tasks of conventional control methods or regulations, as a result of which the effort for regulation and optimization of the compressor operation is kept low overall. The conventional control can be both a flow control and a pressure control, the latter generating the additional increments Y1 and Y2 with the same sign by comparing the total pressure setpoint and the current pressure.

Claim 8 specifies a specific embodiment of the method for the series operation of two compressors in the form of a sequence of individual method steps analogous to claim 3. These steps result in a preferred sequence of the method for the application mentioned, with further refinements of the method being specified specifically for series operation in claims 9 and 10.

According to claim 9, comparable to the feature of claim 4, it is provided to reduce the incremental factors when approaching the optimum in order to achieve both a fast and accurate range Achieve process flow.

Analogous to the process design according to claim 7 for parallel operation, for series operation it is specified in claim 10 how changes in requirements resulting from the process can be taken into account in the process with regard to the overall pressure ratio. Here, too, the method can take on subtasks of conventional control. If the flow rate is to be increased for compressors in series operation, the additional flow rate required must first be converted into mass flow if it has not yet been recorded in this unit. This mass flow must then be converted to the respective volume flow taking into account the standard density as well as the current pressure and the current temperature in the inlet of the respective compressor in the series. Using this volume flow, manipulated variables and input parameters can then be determined from the corresponding characteristic maps. Of course, the conventional regulation here can be not only a pressure ratio regulation but also a flow regulation, the latter generating the additional increments Y 'specified in claim 10 by comparing the total flow nominal value and the current flow.

Process developments that can be used for both parallel and series operation are dealt with below.

For technical reasons, the adjustment speed of the manipulated variable is often limited in practice to a speed that can be understood by the compressor or its actuators. It is therefore expedient to limit the speed of the changes in the manipulated variables that occur through the method. This is limited in the process the increments are reached to suitable values, which depend on the desired maximum adjustment speed and the time required for each process cycle.

Since labor costs for energy, e.g. B. the electric current for the drive of the compressors and any associated auxiliary units are not always constant in time, but rather are often subject to daily or seasonal fluctuations, it is provided to take this into account by providing different input parameter maps in the process.

In order to be able to carry out the method as quickly and effectively as possible even when it is used on more than two compressors, it is further provided according to claim 13 that the adjustment of the manipulated variables is limited to those compressor pairs which produce the relatively greatest effect in the desired direction. This suppresses manipulated variable adjustments that contribute only slightly or insignificantly to the optimization, but which took a relatively long time.

When there are several compressors in a system, operating situations often occur in which the requirements specified by the process with regard to pressure ratio and volume flow can be met with different numbers and / or combinations of compressors. Because of the non-linearity of the characteristic curves of the compressors, it is not readily apparent, for example, whether it is operationally more advantageous to operate a smaller number of compressors in the full or overload range or a larger number of compressors in the part-load range. With different compressors in a system, there is also the question of the optimal combination, if not all compressors have to be operated. The method variant serves to solve these questions according to claim 14, by means of which a clear definition of the best possible number and combination of compressors is made possible in order to fulfill the respective process requirements.

Another situation that occurs in the operation of compressors is the blowing off of one or more compressors under the control of the surge limit control, e.g. B. with a sudden reduction in the volume flow removed from the process. With the method development specified in claim 15, effects of a blow-off on the course of the method and the optimization result are prevented by ensuring that only the relevant, ie. H. the volume flow entering the process enters the process.

As far as described so far, the method is based on the prerequisite that only one optimum exists in the working range of the compressors or the compressor combinations. However, it can also happen that several relative optima exist, which can lead to the method determining a relative optimum, which, however, is not the absolute optimum. One way to rule out this undesirable effect is specified in claim 16. In this design of the method, the entire working area is therefore checked to determine whether there are relative optima in order to select the absolute optimum from this, if necessary.

In the vast majority of use cases of the method that occur in practice, one can be of essentially the same. Composition and inlet temperature of the medium to be compressed are assumed. In order to be able to use the method in cases where the composition and inlet temperature and thus the gas data of the medium can fluctuate greatly, it is advisable to use the instead of a map Pressure ratio over the intake volume flow to choose a map, in which the delivery head or enthalpy difference is used instead of the pressure ratio. The pressure ratio is converted into the head or enthalpy difference based on the known physical relationships and conversion formulas. In the course of the process, pressure conditions are then also recorded and incremental factors are varied; however, the conversion described is carried out before the subsequent determination of the manipulated variables.

The new method can be used for the parallel and series operation of any compressor in many technical areas, for. B. in the chemical industry, in particular petrochemicals, for gas transport in pipelines, in the iron and steel industry, in particular for blast furnace operation and in other, especially industrial areas.

Claims (16)

1. Process for the optimised operation of two or more compressors in parallel or series operation for compressing and conveying gaseous or vaporous media, wherein the operating parameters defining the actual operating point, such as the intake volume flow and the pressure ratio, the correcting variable, such as the compressor rotational speed or vane setting or throttle setting, and the input parameters, such as the power requirement or the compressor operating costs, are detected or determined for each of the compressors, and wherein a superordinated control device is provided which regulates the compressors in accordance with the requirements of a downstream process and preferably includes a surge limit control device, characterised:

   - in that, in parallel operation, the operating points of two compressors in each case are shifted cyclically without influencing the actual overall volume flow and the overall pressure ratio by the additive and incremental variation in counter-current of the individual volume flows, in each case by corresponding adjustment of the correcting variable;

   - in that, in series operation, the operating points of two compressors in each case are shifted cyclically without influencing the actual overall throughflow and the overall pressure ratio by the multiplicative and incremental variation in counter-current of the individual pressure ratios, in each case by corresponding adjustment of the correcting variable;

   - in that, in parallel operation, in accordance with the resultant change in direction of the input parameter total, a further incremental variation of the individual volume flows in the previous directions occurs when the input parameter total is varied in the optimisation direction, ie. in the direction of a reduction in the overall power requirement or operating costs, and an incremental variation of the individual throughflows occurs in each case in the opposite direction when the input parameter total is varied in the opposite direction to the optimisation direction, ie. in the direction of an increase in the total power requirement or operating costs, as a result of a further adjustment of the correcting variable;

   - in that, in series operation, in accordance with the resultant change in direction of the input parameter total, when the latter is changed in the optimisation direction, ie. in the direction of a reduction of the overall power requirement or operating costs, a further incremental variation of the individual pressure ratios occurs in the previous directions and, when it is varied in the direction opposite to the optimisation direction, ie. in the direction of an increase in the overall power requirement or operating costs, an incremental variation of the individual pressure ratios in the inverse direction in each case is performed by corresponding renewed adjustment of the correcting variable;

   - and in that, in parallel and/or series operation, when there are more than two compressors in operation, alternating compressor pairs are formed as a result of all the possible combinations being run through in succession.
2. Process according to Claim 1, characterised:

   - in that the operating points of the compressors are firstly shifted by computation;

   - in that the input parameter total is deduced from the stored data of the existing input parameter performance graphs of the individual compressors and the resultant change in direction of the input parameter total is established;

   - in that a real adjustment of the correcting variables of the compressors only occurs if firstly a change of the input parameter total in the optimisation direction, ie. in the direction of a low overall power requirement or operating costs, is established, or only if secondly an optimum of the input parameter total is established;

   - in that, in parallel operation, in the input parameter performance graph the input parameter is in each case plotted over the intake volume flow or, in series operation, over the pressure ratio;

   - and in that characteristic curves are plotted for correcting variables which are constant in each case.
3. Process according to at least one of Claims 1 and 2, characterised in that, in parallel operation, for two compressors operated in parallel in each case

   a) each correcting variable n1 or n2, such as the compressor rotational speed, blade setting or throttle flap setting, belonging to the actual operating point, deduced from its own stored correcting variable performance graph, such as rotational speed performance graph, blade performance graph or throttle flap performance graph, the pressure ratio being plotted in each case over the intake volume flow and characteristic curves being plotted for correcting variables which are constant in each case in the correcting variable performance graph;

   b) the numerical value determined for the intake volume flow V1 is increased by addition by an increment V and the numerical value determined for the other intake volume flow V1 is decreased by subtraction by the same increment V for the shift by computation of the operating point;

   c) each correcting variable n1* or n2*, belonging in each case to the shifted operating point, is deduced from the stored correcting variable performance graphs;

   d) the input parameters N1 and N2 belonging to the actual operating points and the input parameters N1* and N2* belonging to the shifted operating points of both compressors, corresponding for example to the individual power requirements or operating costs, are deduced from the stored input parameter performance graphs, wherein N1 corresponds to the actual input parameter of the first compressor and N2 to the actual input parameter of the second compressor and N1* corresponds to the input parameter, belonging to the shifted operating point, of the first compressor and N2* corresponds to the input parameter, belonging to the shifted operating point, of the second compressor;

   e) the input parameter totals N = N1 + N2 and N* = N1* + N2*, corresponding for example to the overall power requirements or operating costs for the two operating point combinations, are formed and compared, and in that either

   f1) when N* is smaller than N, the intake volume flow V1 of the first compressor is increased really by corresponding adjustment of its correcting variable by an increment X and the intake volume flow V2 of the second compressor is decreased really by corresponding adjustment of its correcting variable by the same increment X, and the process is continued from step a) or, if N is smaller than N*, the intake volume flow V1 is decreased by computation by an increment X and the intake volume flow V2 is increased by computation by the same increment X and the process is continued from step b), or

   f2) if N* is smaller than N, the intake volume flow V1 is increased by computation by the same increment X and the intake volume flow V2 is decreased by computation by the same increment X, and the process is continued from step b), or, if N is smaller than N*, the intake volume flow V1 is decreased by computation by an increment X and the intake volume flow V2 is increased by computation by the same increment X and the process is continued from step b), or, when an optimum for the input parameter total is determined by continuous comparison of the values determined in each case for the input parameter totals, the correcting variables of the compressors are adjusted accordingly and the compressors are run to the operating points corresponding to the optimum of the input parameter total and the process is continued from step a).
4. Process according to Claim 3, characterised in that, in parallel operation, the increments V and X are varied from process cycle to process cycle in dependence on the differences determined between N and N* of the input parameter totals, the increments V and X being decreased as the differences between N and N* become smaller, ie. when they approach the optimum of the input parameter total, and vice versa.
5. Process according to at least one of Claims 2 and 4, characterised in that, in parallel operation, the pressure ratio is determined separately for each compressor in dependence on the respective pipe length between the compressor outlet and a downstream process, on the respective volume flow, and on the respective pressure loss coefficient of the pipeline.
6. Process according to at least one of Claims 3 to 5, characterised in that, in parallel operation, during the course of the process, the actual pressure ratio in each case is determined as a measuring value by at least one pressure sensor or is taken over by an existing compressor regulator unit as a pressure ratio nominal value.
7. Process according to at least one of Claims 3 to 6, characterised in that, in parallel operation, during the course of the process, the actual overall volume flow in each case is input directly as an overall volume flow nominal value or is taken over as an operating parameter nominal value by an upstream compressor regulator unit; and in that, when the overall volume flow is altered as required, ie. when there is a difference between the overall volume flow nominal value and the total of the individual volume flows V1 and V2, in addition to the incremental variation in counter-current of the individual volume flows V1 and V2, the latter are in each case varied by an increment Y1 or Y2 which is identical to the preceding sign, the total of the increments Y1 and Y2 corresponding to the difference between the overall volume flow nominal value and the total of the individual volume flows V1 and V2.
8. Process according to at least one of Claims 1 and 2, characterised in that, in series operation, for two compressors operated in series in each case:

   a) each correcting variable N1 or N2 belonging to the actual operating point, such as the compressor rotational speed, blade setting or throttle flap setting, is deduced from its own stored correcting variable performance graph, such as rotational speed performance graph, blade performance graph or throttle flap performance graph, the compressor pressure or the pressure ratio in each case being plotted over the intake volume flow and characteristic curves being plotted for correcting variables which are constant in each case in the correcting variable performance graph;

   b) for the shifting by computation of the operating point, the numerical value determined for one pressure ratio π1 is increased by multiplication by an incremental factor π, which is greater than 1, and the numerical value determined for the other pressure ratio π2 is decreased by division by the same incremental factor, and the intake volume flow varied as a result of the variation of the pressure ratio in accordance with the resultant density variation is determined for both compressors in each case on the assumption that the throughflow remains the same;

   c) each regulating variable n1* or n2* belonging to the shifted operating point is deduced from the stored correcting variable performance graph of the two compressors;

   d) the input parameters N1 and N2 belonging to the actual operating point and N1* and N2* belonging to the shifted operating point of both compressors, corresponding for example to the individual power requirements or operating costs, are deduced from the stored input parameter performance graphs;

   e) the input parameter totals N = N1 + N2 and N* = N1* + N2*, corresponding for example to the total power requirements or operating costs for the two operating point combinations, are formed and compared; and in that either

   f1) if N* is smaller than N, the pressure ratio π1 of the first compressor is increased really by multiplication, by corresponding adjustment of its correcting variable, by an incremental factor Z, which is greater than 1, and the pressure ratio π2 of the second compressor is decreased really by division, by corresponding adjustment of its regulating variable, by the same incremental factor Z, and the process is continued from step a) or, if N is smaller than N*, the pressure ratio π1 is decreased by computation by division by an incremental factor Z, which is greater than 1, and the pressure ratio π2 is increased by computation by multiplication by the same incremental factor Z, and the process is continued from step b), or

   f2) if N* is smaller than N, the pressure ratio π1 is increased by computation by multiplication by an incremental factor Z, which is greater than 1, and the pressure ratio π2 is decreased by computation by division by the same incremental factor Z, and the process is continued from step b), or, if N is smaller than N*, the pressure ratio π1 is decreased by computation by division by an incremental factor Z, which is greater than 1, and the pressure ratio π2 is increased by computation by multiplication by the same incremental factor Z, and the process is continued from step b), and, when an optimum is determined for the input parameter total by continuous comparison of the values determined in each case for the input parameter totals, the correcting variables of the compressors are adjusted accordingly and the compressors are operated to the operating points corresponding to the optimum of the input parameter total, and the process is continued from step a).
9. Process according to Claim 8, characterised in that, in series operation, the incremental factors π and Z are varied from process cycle to process cycle according to the differences between N and N* determined for the input parameter totals, the incremental factors O and Z being reduced as the differences between N and N* become smaller, ie. as they approach the optimum of the input parameter total, and vice versa.
10. Process according to at least one of Claims 8 and 9, characterised in that, in series operation, during the course of the process, the actual overall pressure ratio in each case is input directly as the overall pressure ratio nominal value or is taken over as an operating parameter nominal value by an upstream compressor regulator unit; and in that, when the overall pressure ratio nominal value is altered as required, ie. when there is a difference between the overall pressure ratio nominal value and the product of the individual pressure ratios π1 and π2, in addition to the counter-current incremental multiplicative variation of the individual pressure ratios π1 and π2, the latter are multiplied in each case by a factor Y' which is identical in terms of amounts and preceding sign and which corresponds to the required factor for increasing the overall pressure ratio.
11. Process according to at least one of Claims 1 to 7 or 8 to 10, characterised in that the increments Y1 and Y2 or Y' are restricted to a maximum value Y1_max and Y2_max or Y'_max corresponding to a required maximum adjusting velocity of the correcting variable.
12. Process according to at least one of Claims 2 to 10, characterised in that a plurality of input parameter performance graphs with different records varying according to the time of day, day of the week and/or season is used for each compressor.
13. Process according to at least one of Claims 2 to 12, characterised in that, if more than two compressors are operated, when all the possible combinations for forming alternating compressor pairs in each case have been run through, only the correcting variable of the compressors of the particular compressor pair which has displayed the greatest variation of the input parameter total in the optimisation direction is adjusted.
14. Process according to at least of one of Claims 2 to 11, characterised in that, when a plurality of compressors is operated under predetermined process requirements with respect to volume flow and pressure ratio for each individual compressor and each possible combination of two or more compressors, the optimum of the input parameter or the input parameter total for the respective requirements is determined; in that the optima of the input parameter total are compared; and in that the compressor or compressor combination in which the absolute optimum of the input parameter or input parameter totals is present remains in operation or is taken and operated to the operating point or operating points.
15. Process according to at least one of Claims 1 to 12, characterised in that, when one or more compressors blows off in each case, the value determined for the intake volume flow is reduced by the partial amount blown off or the volume flow passing to the process is determined directly.
16. Process according to at least one of Claims 1 to 15, characterised:

    - in that, when an optimum for the input parameter total of the compressors has been reached, its associated operating points are retained;

    - in that subsequently the operating points within the performance graph limits are shifted a number of times abruptly, in counter-current, really or by computation without influencing the overall operating parameters, in that the increment for the individual volume flows or the individual pressure ratios is increased a number of times by a factor which is substantially greater than 1;

    - in that, starting from each of these new operating point pairs, an optimum of the input parameter total is again established and each of these newly established optima is compared with the optimum originally established to determine the absolute optimum; and

    - in that subsequently, if necessary, the compressors are operated really by appropriate variation of the individual correcting variables to the operating points which correspond to the absolute optimum of the input parameter total newly determined if necessary.