US20130180275A1

US20130180275A1 - Method for Controlling a Heat-Transfer Fluid-Compression Device of a Cryogenic Machine

Info

Publication number: US20130180275A1
Application number: US13/738,528
Authority: US
Inventors: Francois Bonne; Patrick Bonnay
Original assignee: Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Current assignee: Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Priority date: 2012-01-12
Filing date: 2013-01-10
Publication date: 2013-07-18
Also published as: FR2985805A1; EP2615394A1; FR2985805B1

Abstract

Method for controlling a heat-transfer fluid compression device (3) of a cryogenic machine, the compression device comprising a plurality of members including a compressor (4) and a first controlled valve (CV956) mounted in parallel, wherein the method comprises the real-time generation of a control of the compressor and the real-time generation of a control of the first controlled valve.

Description

The present invention relates to a method for controlling or operating a heat-transfer fluid compression device of a cryogenic machine. The invention also relates to a heat-transfer fluid compression device of a cryogenic machine. The invention also relates to a cryogenic machine.
Most of the large cryogenic machines currently designed are not intended to cool strongly variable thermal loads. A minority of them therefore see their applied thermal load vary strongly during operation. Whatever the application (constant thermal load or variable thermal load), the cryogenic machines are dimensioned and designed in the same way, that is to say such that they can supply the maximum load, while always operating at full speed. This appears suitable when the load varies little. However, on the other hand, when the load is variable (for example in an application for cooling superconductive magnets for fusion), this poses three problems:

- an overhead linked to the dimensioning to supply the maximum power permanently,
- an overhead linked to operation, linked to the energy consumption,
- strong agitations of the normally controlled variables, thus affecting the systems that depend thereon.

A cryogenic machine, notably a cryorefrigerator comprises a fluid compression device, also called “hot zone”. The function of this compression device is to generate a flow of heat-transfer fluid, for example gaseous helium, under high pressure, at ambient temperature. This fluid flow is then absorbed and discharged under low pressure by another subassembly, called “cold box” in which the load to be cooled is located.
The compression device is, for example, equipped with various actuators: three valves and a compressor. The hot zone is also equipped with three sensors: the two pressures to be regulated and the outgoing flow rate to the cold box. Note that the flow rate entering into the hot zone is not normally measurable. In the current control strategy, a so-called bypass valve, mounted in parallel with the compressor, is used to redirect to a low-pressure zone the compressed fluid which is not absorbed by the cold box. So-called “input” and “removal” valves serve respectively to add or remove fluid in the circulation circuit of the latter between the cold box and the hot zone. The fluid input is removed from a buffer tank. The removed fluid is stored in the buffer tank. The needs of the applications that use such a cryogenic machine are such that the high and low pressures of the hot zone can follow given trends or, by default, that they can remain constant whatever the state of demand of the cold box.
Such a mode of operation results in a waste of energy. In practice, when the machine is under load, the energy transferred to the fluid in the compressor and not used in the cold box is dissipated at the bypass valve.
The document FR2943768 for example discloses the use of a controller based on linearized equations of a refrigerating system, around an operating point, to control pressures. This controller is not suitable if the method is located at another operating point. The bypass valve is open permanently or quasi-permanently, which results in the abovementioned waste.
The aim of the invention is to provide an operating or control method that makes it possible to remedy the problems described previously and that enhances the methods known from the prior art. In particular, the invention proposes a system that makes it possible to ensure operation of the compression device that meets the demands required by the use of the cryogenic machine, while avoiding wastes of energy. The invention also relates to a compression device operating in accordance with such a method.
According to the invention, the method makes it possible to control a heat-transfer fluid compression device of a cryogenic machine. The compression device comprises a plurality of members including a compressor and a first controlled valve mounted in parallel. The method comprises the real-time generation of a control of the compressor and the real-time generation of a control of the first controlled valve.
The members of the compression device may comprise a second controlled valve and a third controlled valve, the second controlled valve, a heat-transfer fluid tank and the third controlled valve being mounted in series and forming part of a parallel-mounted assembly of the compressor and the method may comprise the real-time generation of a control of the second controlled valve and the real-time generation of a control of the third controlled valve.
The real-time generation of at least one control of a member may comprise the use of the reciprocal steady-state transfer function of the member.
The control of a first member can be modified when a calculation of control of a second member establishes a control value that is impossible to execute.
The real-time generation of at least one control of a member may comprise the use of an estimation of at least one value of a variable of the compression device.
The method may comprise the generation of at least one prepositioning control value.
The method may comprise the use of a linear controller.
The method may comprise the use of a nonlinear observer.
According to the invention, a heat-transfer fluid compression device of a cryogenic machine comprises a plurality of members including a compressor and a first controlled valve mounted in parallel. The method comprises hardware and/or software elements for implementing the method defined previously.
The hardware and/or software elements may comprise an element for generating in real time a control of the compressor and an element for generating in real time a control of the first controlled valve.
The compression device may comprise a second controlled valve and a third controlled valve, the second controlled valve, a heat-transfer fluid tank and the third controlled valve being mounted in series and forming a part of a parallel-mounted assembly of the compressor and the hardware and/or software elements may comprise an element for generating in real time a control of the second controlled valve and an element for generating in real time a control of the third controlled valve.
The hardware and/or software elements may comprise an element for using the reciprocal steady-state transfer function of the member.
The hardware and/or software elements may comprise an element for modifying the control of a first member when a calculation of control of a second member establishes a control value that is impossible to execute by the second member.
The hardware and/or software elements may comprise an observer of at least one value of a variable of the compression device.
The hardware and/or software elements may comprise a controller.
According to the invention, a cryogenic machine comprises a compression device defined previously.

The appended drawing represents, by way of example, two embodiments of a compression device according to the invention.

FIG. 1 is a hydraulic diagram of a first embodiment of a compression device according to the invention.

FIG. 2 is a diagram of the automatic architecture of the embodiment of a compression according to the invention.

FIG. 3 is a diagram illustrating the nonlinearities of a member of a compression device.

FIG. 4 is a time graph of the trends of the value the high fluid pressure in the embodiment of the compression device according to the invention.

FIG. 5 is a time graph of the trends of the value of the low fluid pressure in the embodiment of the compression device according to the invention.

FIG. 6 is a time graph of the trends of the rotation frequency of the compressor of the openings of input and removal valves in the embodiment of the compression device according to the invention.

FIG. 7 is a hydraulic diagram of a second embodiment of a compression device according to the invention.

An embodiment of a cryogenic machine is described below with reference to FIG. 1.
The cryogenic machine 1 mainly comprises a device 3 for compressing a heat-transfer fluid, such as helium, and a cold box 2 comprising an expansion valve and in which is located a thermal load to be cooled.
A function of the compression device is to generate a flow QHP of heat-transfer fluid under high pressure HP. At the outlet of the cold box, a flow QLP of heat-transfer fluid under low pressure LP is collected. The compression device comprises a plurality of members including a compressor 4 and a first controlled valve CV956 hydraulically mounted in parallel. If appropriate, the members of the compression device also comprise a second controlled valve CV952 and a third controlled valve CV953. The second controlled valve, a heat-transfer fluid tank 6 and the third controlled valve are hydraulically mounted in series and form part of a hydraulically parallel-mounted assembly of the compressor.
The hot zone may also be equipped with three sensors: a high pressure HP sensor, a low pressure LP sensor and a cold box incoming flow rate sensor QHP. The flow rate entering into the hot zone cannot be measured. The valve CV956 makes it possible to redirect the fluid from a high pressure zone of the compressor to a low pressure zone upstream of the compressor. The valves CV952 and CV953 are respectively used to add or remove fluid in the circuit in which the latter circulates between the cold box and the hot zone. The added fluid is removed from the tank 6. The removed fluid is stored in the tank 6. The fluid pressure in the buffer tank is therefore between the low fluid pressure prevailing upstream of the compressor and the high fluid pressure prevailing upstream of the compressor.
The compressor comprises a fluid compression element driven by a motor, notably an electric motor, such as an asynchronous motor. The compressor comprises a motor variable speed drive. For example, in the case of an alternating current motor, the latter is preferably powered by the mains via a frequency converter forming part of the variable speed drive. Thus, it is possible to vary the operating speed of the compressor and therefore the fluid compression power. Alternatively, notably in the case of a direct current motor, it is possible to vary the operating speed of the compressor and therefore the fluid compression power by implementing a chopping of the motor power supply signal for example by using controlled switches.
According to another embodiment, it is possible to use the slide valves of the compressor to vary and control the flow rate.
The operation of each element of the compressor, in particular of the motor or of the power supply thereof and of the controlled valve CV956 is driven by a processing logic unit 7. The operation of each of the controlled valves CV952 and CV953 can also be driven by the processing unit.
The compression device comprises hardware and/or software elements making it possible to govern its operation, that is to say implement the control or operating method according to the invention. All or part of these hardware and/or software elements are contained in the processing logic unit 7. The hardware and/or software elements notably comprise an element for generating in real time a control of the compressor 4 and an element for generating in real time a control of the first controlled valve CV956.
The hardware and/or software elements may comprise software elements.
An embodiment of the method for controlling or operating the compression device, and therefore the cryogenic machine comprising such a compression device, is described below.
For the compression device to exhibit satisfactory performance levels, its members have to be monitored or controlled by so-called “advanced” or “model-based” control laws. To satisfy this constraint, a multiple-objective controller or control law is used, said multiple objectives being:

- to ensure maximum performance levels regardless of the state of demand,
- to minimize the energy consumption.

The physical system consisting of the cryogenic machine or of the compression device is of the type that is nonlinear, coupled, saturated by zero value and high value. This constitutes an obstacle to achieving these objectives.
It is possible to partially overcome these obstacles by using, as seen previously, a variable speed drive of the compressor, making it possible to set the adequate fluid flow rate, in real time. The speed of the compressor can therefore be adapted to the load, at each instant. If the compressed fluid flow rate is adequate, then the valve CV956 can be completely closed. This makes it possible to limit the energy consumption of the motor which is, for example, proportional to its speed. If the speed is adapted in real time, less energy is consumed than if the speed were fixed, and, by default, nominal.
Another advantage also appears. It is possible, temporarily and/or within a certain limit, to operate the compressor in overload mode, when using a variable speed drive based on a frequency converter. It is therefore possible to increase the capacity of the machines without having to change the compressor (the cost of which is of the order of 10 times greater than a variable speed drive).
Furthermore, the speed of the compressor has to be adapted at each instant. Since the compressor is not the only member of the compression device, and each speed variation acts both on the high pressure and on the low pressure value, it is necessary to develop a “model-based” multiple-objective control law.
a.- To do this, at least one member is preferably controlled through its reciprocal steady-state transfer function. Thus, the steady-state nonlinearities due to the members can be eliminated.
b.- Also preferably, at least one member has a task allocated to it that another member cannot handle. Thus, the saturations of the members can be eliminated, on the side of the low values of their operating range.
c.- Also preferably, at least one member is pre-positioned in an operating state that is considered to be suitable. Thus, instabilities of the system are avoided during control law transitions.
d.- Also preferably, finally, at least one non-measured variable is estimated in real time. This makes it possible in particular to implement the pre-positioning of at least one member in an operating state that is considered suitable.
To produce the controller, any coherent combination of one, two, three or four of the above proposals a.-, b.-, c.- and d.- can be implemented.
The combination of the four proposals makes it possible to produce a linear controller for the operation or control method. It is possible to use standard generation algorithms (such as Hinf (Hinfinity) and LQG (Linear Quadratic Gaussian)). It is also possible to use simple PI (Proportional Integral), PD (Proportional Derivative) or PID (Proportional Integral Derivative) algorithms. In the case of the compression device, the system is of “non-square” type (non-square system) because the number of variables to be regulated is different from the number of members. Furthermore, the members are not identical and have, for example, control reaction speeds which are different.
The operation of the compression device described above with reference to FIG. 1 is governed by the following equations:
$\dot{L} P = \frac{- Q_{prod} + Q_{952} + Q_{LP}}{K_{LP}}$ $\dot{H} P = \frac{Q_{prod} - Q_{953} - Q_{HP}}{K_{HP}}$
With KLP and KHP varying respectively:
$K_{LP} = \frac{RT}{{MV}_{LP}}$ $K_{HP} = \frac{RT}{{MV}_{HP}}$
with:
VLP the low-pressure volume seen by the device;
VHP the high-pressure volume seen by the device;
R the constant of the ideal gases;
M the molar mass of the heat-transfer fluid; and
T the temperature of the fluid treated,
these quantities being expressed in SI (international system) units.
The set of mathematical laws making it possible to produce the linear controller or the linear control law to drive the system is articulated as represented in FIG. 2. The processing logic unit making it possible to establish the controls for the set 31 of the members mainly comprises an observer 10, an allocation block 20, a pre-positioning block 30, a conversion block 40 and a feedback block 50.
The observer 10 makes it possible to determine estimations of cold box incoming and outgoing flow rates when the latter are not measured. The observer uses the dynamic equations of the system, the controls supplied to the members and measurements notably the HP pressure value downstream of the compressor and the LP pressure value upstream of the compressor.
The non-measured flow rate QLP is determined using the observer or estimator. The equations of the system of FIG. 1 are:
$\dot{L} P = \frac{+ Q_{956} + Q_{952} + Q_{LP} - Q_{nc 1}}{K_{LP}}$ $\dot{H} P = \frac{- Q_{956} + Q_{953} + Q_{HP} + Q_{nc 1}}{K_{HP}}$ $with$ $Q_{prod} = Q_{nc 1} - Q_{956} .$
By linearizing around an operating point, the state model is characterized by the state matrix A:
$A = [\begin{matrix} - \frac{N_{nc 1_{0}} K_{nc 1_{0}}}{K_{LP}} + \frac{Q_{956_{0}}}{K_{LP} . {HP}_{0}} + \\ \frac{N_{nc 1_{0}} K_{nc 1_{0}}}{K_{HP}} - \frac{Q_{956_{0}}}{K_{HP} . {HP}_{0}} \end{matrix}]$
and its input matrix:
$B = [\begin{matrix} \frac{1}{K_{LP}} & \frac{1}{K_{LP}} & 0 & \frac{1}{K_{LP}} & 0 & - \frac{1}{K_{LP}} \\ - \frac{1}{K_{HP}} & 0 & - \frac{1}{K_{HP}} & 0 & - \frac{1}{K_{HP}} & \frac{1}{K_{HP}} \end{matrix}]$
The two states are directly measured, the matrix C is therefore an identity matrix of appropriate size. The direct action matrix is null.
The non-measured flow rates, for example QHP and QLP, are considered to be exogenous (their trend is not governed by a known differential equation) and, in order to be able to estimate them, they will be defined as substantially time-constant states ({dot over (Q)}LP={dot over (Q)}HP=0). In the case where the flow rate is measured, it can be included in the state model to be able to filter it (for example QHP) and it will have to appear in the output matrix C. The matrices of the state mode of the observer are then as follows:
$A_{0} = [\begin{matrix} A & B_{Q} \\ 0_{2.2} & A_{Q} \end{matrix}], B_{0} = [\begin{matrix} B \\ 0_{3.6} \end{matrix}]$ $C_{0} = [\begin{matrix} C & 0_{2.3} \\ 0_{1.4} & 1 \end{matrix}], D = [D 0_{2.3}]$
In the first case where the flow rates are considered to be exogenous, the following applies:
$A_{Q} = 0_{2.2}$ $and$ $B 0 = [\begin{matrix} K_{LP}^{- 1} & 0 \\ 0 & K_{HP}^{- 1} \end{matrix}]$
A return gain is then calculated, for example using the dare(.) or place(.) function of the Matlab® software (after discretization). Once this gain is known, the observer is expressed:
Q₉₅₂=K.CV₉₅₂.P_eq1
Q₉₅₃=K.CV₉₅₃.P_eq2
Q₉₅₆=K.CV₉₅₆.P_eq3
The quantities Peq1, Peq2 and Peq3 being functions of HP, LP and Pcapa.
$[\begin{matrix} \dot{\hat{L}} P \\ \dot{\hat{H}} P \\ {\dot{\hat{Q}}}_{LP} \\ {\dot{\hat{Q}}}_{HP} \\ {\dot{\hat{Q}}}_{956} \end{matrix}] = [\begin{matrix} \frac{+ {\hat{Q}}_{956} + Q_{956} + Q_{952} + {\hat{Q}}_{LP} - Qnc 1}{K_{LP}} \\ \frac{- {\hat{Q}}_{956} - Q_{956} - Q_{953} - {\hat{Q}}_{HP} + Qnc 1}{H_{HP}} \\ 0 \\ 0 \\ 0 \end{matrix}] + L ([\begin{matrix} LP \\ HP \\ Q_{LP} \end{matrix}] - [\begin{matrix} \hat{L} P \\ \hat{H} P \\ {\hat{Q}}_{LP} \end{matrix}])$
with Qnc1=Knc1.Nnc1. {circumflex over (L)}P-Knc10.
A non-linear observer is obtained which is particularly suited to the desired use.
The allocation block 20 makes it possible to produce control compensations between the members. In particular, when the dynamic equations of the system lead to providing a parameter or a variable that does not have physical reality, for example a negative flow rate through a valve, the allocation function compensates by ordering an equivalent configuration on the system by virtue of the other control members. As input, it uses flow rates, notably one or more flow rates that do not have physical reality and supplies, as output, corrective flow rate values.
The possibility of producing such an allocation function is described below. The matrix Bc, control distribution matrix, the control vector u, and the resultant vector on the system v are considered.
$B_{c} = [\begin{matrix} - 1 & + 1 & 0 \\ + 1 & 0 & - 1 \end{matrix}], u = [\begin{matrix} Q_{prod} \\ Q_{952} \\ Q_{953} \end{matrix}], v = [\begin{matrix} Q_{952} - Q_{prod} \\ Q_{prod} - Q_{953} \end{matrix}]$
For example, we can note that, if the Q952 control is negative, it is sufficient to subtract its value from all the elements of the vector. Thus:
$u = [\begin{matrix} Q_{prod} - Q_{952} \\ Q_{952} - Q_{952} \\ Q_{953} - Q_{952} \end{matrix}] = [\begin{matrix} Q_{prod} - Q_{952} \\ 0 \\ Q_{953} - Q_{952} \end{matrix}], v = [\begin{matrix} Q_{952} - Q_{prod} \\ Q_{prod} - Q_{953} \end{matrix}]$
The result on the system is the same, the negative Q952 control has been assigned to other members. Identically, cases where both Q952 and Q953 fluid input and removal flow rates are ordered positive at the same instant, negative at the same instant, or even of opposite signs, can be managed.
This system can be further simplified by reducing the matrix Bc to two columns (the first being a linear combination of the next two) which corresponds to the control of two virtual members which will be transformed into two real members by the allocation function.
The pre-positioning block 30 makes it possible to pre-position at least one member in the state considered to be suitable. This makes it possible to switch from one control law to another without “rushing” the system or making it unstable. Consequently, the pre-positioning block can allow for abrupt variations of the control of at least one member. For input, it uses flow rates, notably one or more estimated flow rates and supplies flow rate values as output.
The regulators usually designed are linear regulators. These linear regulators are designed to be executed about an operating point. The initialization of the output value is therefore a crucial factor, since the system is saturated positively (100% maximum opening of valve, maximum compressor speed selected). Starting up the regulator with null control values would destabilize the loop. To resolve the problem, the system is initialized with the assumed good regulation values. In order to know these values, the steady-state equation of the system is resolved.
{dot over (L)}P=0=Q ₉₅₂ +Q _LP −Q _prod →Q _prod =Q ₉₅₂ +Q _LP
{dot over (H)}P=0=Q _prod +Q ₉₅₃ −Q _HP →Q _prod =Q ₉₅₃ +Q _HP
The solution, for the flow rate to be compressed Qprod, subject to flow rate positivity, is:
if Q _LP >Q _HP →Q _prod =Q _LP
if Q _HP >Q _LP →Q _prod =Q _HP
by positing
ΔQ _LF>0→Q ₉₅₃ =ΔQ _LF , Q ₉₅₂=0
The solution is:
if ΔQ _LF>0→Q ₉₅₃ =ΔQ _LF , Q ₉₅₂=0
if ΔQ _LF<0→Q ₉₅₂ =ΔQ _LF , Q ₉₅₃=0
It is therefore sufficient to order these default flow rates on the machine for the pressures to be stable. The problem is that the QLP flow rate is not measured. To resolve this problem, the QLP flow rate is estimated.
The conversion block 40 makes it possible to translate the fluid flow rates desired in the compression device into commands that can be applied by the members (valve opening, compressor speed). As input, it therefore uses flow rates and supplies controls for all the control members as output.
At the valve level, the form of the non-linearities is known and strictly monotonic. It is as represented in FIG. 3. The relationship linking the opening to its capacity to allow the fluid to pass is expressed as follows:
$C_{V} = \frac{C_{V \max}}{R_{V}} ((e^{\frac{pos}{100} l n (Rv)}) - (1 - \frac{pos}{100}))$
with:
pos: the opening of the valve,
Cv: the capacity to allow the fluid to pass.
It is therefore proposed to control the state of the controlled valves by using their reciprocal steady-state transfer functions.
On the compressor, the speed NnC1 of the compressor NC1 acts in an affine manner on the flow rate Qnc1 according to the following relationship:
Q _nc1=K _nc1 NnC1LP−K _nc10
It is therefore proposed to control the speed NnC1 to its inverse or reciprocal transfer function. If the speed of the compressor is too low (below a critical value at which the compressor can no longer operate), the compressor is made to operate at its limit speed of correct operation and the valve CV956 is opened again so that it allows a flow rate to pass which is equivalent to the excess that the compressor generates. The valve CV956 is controlled as described above.
The feedback block 50 makes it possible to compensate the model errors and the non-modeled disturbances. For input, it uses parameter values measured in the system and supplies flow rate values as output for the allocation block. By virtue of the blocks 10, 20, 30 and 40, the feedback block can be designed using the conventional techniques of pole placement, output return, state return, constrained optimization, etc.
The controller, by virtue of the elements described above, will act linearly on the system. It will therefore be able to be generated using linear processes that are well known, and proven, for example:

- linear quadratic control, or
- Hinfinity control, or
- adaptive Hinfinity control, or
- control by output return (linear, fractional linear).

In the case of a refrigerator of large dimensions, the compression device may comprise at least two compression stages (two input flow rates and one output flow rate), therefore at least two compressors and a plurality of valves. An embodiment of a cryogenic machine 101 is described below with reference to FIG. 5.
The cryogenic machine 101 mainly comprise a device 103 for compressing a heat-transfer fluid, such as helium, and a cold box 102 comprising an expansion valve and containing a thermal load to be cooled.
The function of the compression device is to generate a flow QHP of heat-transfer fluid under high pressure HP. At the output of the cold box, a flow QLP of heat-transfer fluid under low pressure LP and a flow QMP of heat-transfer fluid under medium pressure MP are collected. The compression device comprises a plurality of members including a first compressor 104 a and a second compressor 104 b mounted in series. The compression device also comprises two controlled valves each hydraulically mounted in parallel with one of the compressors. If necessary, the members of the compression device also comprise other controlled valves and a tank 106 of heat-transfer fluid. By using the same methodology, the state model will then be expressed:
$[\begin{matrix} K_{LP} \cdot \dot{LP} \\ K_{MP} \cdot \dot{MP} \\ K_{HP} \cdot \dot{HP} \end{matrix}] = [\begin{matrix} 0 & 1 & 1 \\ 1 & - 1 & 0 \\ - 1 & 0 & - 1 \end{matrix}] [\begin{matrix} Q_{HP \to MP} \\ Q_{MP \to LP} \\ Q_{HP \to LP} \end{matrix}] + [\langle \begin{matrix} - 1 & 0 \\ 1 & - 1 \\ 0 & 1 \end{matrix} \rangle] [\begin{matrix} Qcmp 1 \\ Qcmp 2 \end{matrix}] + [\begin{matrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & - 1 \end{matrix}] [\begin{matrix} Q_{capa \to LP} \\ Q_{MP \to capa} \\ Q_{HP \to capa} \end{matrix}] + [\begin{matrix} Q_{LP} \\ Q_{MP} \\ Q_{HP} \end{matrix}]$
Tests have been carried out on a compression device according to the invention, this device operating according to the method of the invention. This device makes it possible to ensure iso-performance levels regardless of the demand state of the system and to minimize the energy consumption. Algorithms translating the method have been written in the industrial program logic controllers, then executed. A campaign with variable thermal load was conducted and the results in FIGS. 4 to 6 were obtained.
The results obtained show the energy saving objective has been reached since the compression speed was adapted to the load throughout the variation thereof, as shown by the trends of the frequency represented by the top curve of FIG. 6, and the fluid input and removal valves (variables CV952 and CV953, as % opening) are never open at the same time: they open one after the other. It can also be observed that the HP and LP pressures remained constant (or were considered to be constant, in light of the scales) throughout the tests. Furthermore, these devices are stable and the performance levels are enhanced.
If the pressures are constant during the tests and the two fluid input and removal valves do not open at the same time and the valve CV956 is closed, then the speed of the compressor is matched. The energy consumption is therefore minimized.
During these tests, the maximum compressor speed was 53.15 Hz. The average speed was 46.44 Hz. The saving produced during the cycle was 12% on the compressor (the compressor represents ⅔ of the overall system consumption).
These tests were conducted on a cryogenic test station consuming 330 kW at nominal load. For larger requirements, the power levels required to cool the loads can range up to 10 MW (fusion reactor) and even up to 36 MW.
For non-cryogenic cold requirements, the needs vary enormously. They can range from a few hundred watts (household refrigerator), to a few kilowatts (air conditioning for individual homes) and up to a megawatt for collective water or air temperature conditioning installations.
Also, the compression devices make it possible to reduce the deviations of the pressures relative to their set point. In the tests described, the pressures never deviated by more than 50 mbar (for the high pressure HP) and 5 mbar (for the low pressure LP) from their set point. With a conventional regulation system, these deviations would have been doubled.
By virtue of such a compression device, the size of the compressor can be reduced. In practice, the screw compressors intended for the production of industrial cold operate with suction pressures that vary a lot around their average value. Now, the dimensioning of the machines is done on the maximum suction pressure. By regulating the suction pressure to its average value by virtue of a similar algorithm, the compressor can thus be chosen for a nominal suction pressure value, and no longer for a maximum suction pressure value. Its efficiency would also be increased.
Finally, the compression device makes it possible to optimally control its members in steady-state or variable operation mode.
“Real-time generation” means for example a generation within an allowed time, in particular between 10 ms and 1 s.
“Real-time estimation” means for example an estimation within an allowed time, in particular between 10 ms and 1 s.
“Real-time adaptation” means for example an adaptation within an allowed time, in particular between 10 ms and 1 s.
“Transfer function” means a function, in particular a mathematical function, allowing to bind first and second quantities, in particular a position of a valve and a flow. The function is static when it is not dependent on time. A static function is thus not governed by differential equations.
A first function is a reciprocal function of a second function if it makes it possible to determine a first quantity starting from a second quantity, whereas the second function makes it possible to determine the second quantity starting from the first quantity.
“Linear controller” means a controller that can be represented by a linear function.
“Nonlinear observer of a device” means an observer knowing the input controls of the device and receiving output information from the device. The observer determines the state of the device using a model of the device, the model being described by nonlinear functions.

Claims

1. Method for controlling a device, wherein the device is a heat-transfer fluid compression device of a cryogenic machine, wherein the compression device comprises a plurality of members including a compressor and a first controlled valve mounted in parallel and wherein the method comprises the real-time generation of a control of the compressor and the real-time generation of a control of the first controlled valve.

2. Control method according to claim 1, wherein the members of the compression device comprise a second controlled valve and a third controlled valve, the second controlled valve, a heat-transfer fluid tank and the third controlled valve being mounted in series and forming a part of a parallel-mounted assembly of the compressor and wherein the method comprises the real-time generation of a control of the second controlled valve and the real-time generation of a control of the third controlled valve.

3. Control method according to claim 1, wherein the real-time generation of at least one control of a member comprises the use of the reciprocal steady-state transfer function of the member.

4. Control method according to claim 1, wherein the control of a first member is modified when a calculation of control of a second member establishes a control value that is impossible to execute.

5. Control method according to claim 1, wherein the real-time generation of at least one control of a member comprises the use of an estimation of at least one value of a variable of the compression device.

6. Control method according to claim 1, wherein the method comprises the generation of at least one prepositioning control value.

7. Control method according to claim 1, wherein the method comprises the use of a linear controller.

8. Control method according to claim 1, wherein the method comprises the use of a nonlinear observer.

9. Device wherein the device is heat-transfer fluid compression of a cryogenic machine, wherein the compression device comprises a plurality of members including a compressor and a first controlled valve mounted in parallel and wherein the device comprises hardware and/or software elements for implementing the method according to claim 1.

10. Compression device according to claim 9, wherein the hardware and/or software elements comprise an element for generating in real time a control of the compressor and an element for generating in real time a control of the first controlled valve.

11. Compression device according to claim 10, wherein the device comprises a second controlled valve and a third controlled valve, the second controlled valve, a heat-transfer fluid tank and the third controlled valve being mounted in series and forming a part of a parallel-mounted assembly of the compressor and wherein the hardware and/or software elements comprise an element for generating in real time a control of the second controlled valve and an element for generating in real time a control of the third controlled valve.

12. Compression device according to claim 9, wherein the hardware and/or software elements comprise an element for using the reciprocal steady-state transfer function of the member.

13. Compression device according to claim 9, wherein the hardware and/or software elements comprise an element for modifying the control of a first member when a calculation of control of a second member establishes a control value that is impossible to execute by the second member.

14. Compression device according to claim 9, wherein the hardware and/or software elements comprise an observer of at least one value of a variable of the compression device.

15. Compression device according to claim 9, wherein the hardware and/or software elements comprise a controller.

16. A cryogenic machine comprising a compression device according to claim 9.