DE102021213591A1 - Refrigeration device and method for operating a refrigeration device - Google Patents

Abstract

A refrigeration appliance (1) has a refrigeration circuit with a variable-speed compressor (4), a condenser (7), at least one evaporator (2, 3) and at least one throttle element (5, 6), and also a refrigerated space with at least one temperature zone, the at least one evaporator is assigned in each case, a temperature sensor per temperature zone for sensing actual values (TF1, TF2) of an associated zone temperature, a set of actuators (4-6) that can be variably adjusted via manipulated variables (n, d1, d2), including the compressor , which have a noticeable influence on the operation of the refrigeration circuit, the number of actuators being higher than the number of evaporators, and a control unit (9) which is set up to calculate values of the manipulated variables of the actuators based on deviations between the actual values of the zone temperatures and specified setpoint values (TF1,setpoint, TF2,setpoint) thereof, with the values of the manipulated variables being calculated using a circuit model.

Description

The invention relates to a refrigeration appliance, having a refrigeration circuit with a speed-controllable compressor, a condenser, at least one evaporator and at least one throttle element, a refrigerated space with at least one temperature zone, to which at least one evaporator is assigned, at least one temperature sensor per temperature zone for sensing actual Values of an associated zone temperature, a set of actuators that can be set via actuating variables, including the compressor, which have a noticeable influence on the operation of the refrigeration circuit, the number of actuators being higher than the number of evaporators, and a control unit that is set up to To specify manipulated variables of the actuators based on deviations between the actual values of the zone temperatures and specified target values thereof, the control unit being data-technically coupled to a computer-readable storage medium which includes commands which, when executed by the control unit, cause this to change the values of the manipulated variables to calculate the actuators on the basis of a circuit model, and the circuit model describes a control circuit for setting the actual values of the zone temperatures to respectively predetermined target values. The invention also relates to a corresponding method for operating a refrigeration device. The invention can be applied particularly advantageously to household refrigeration appliances in which the compressor is operated continuously.

The article by Alexsandro S. Silveira, Marcelo D.C. de Oliveira, Alexandre Trofino Neto, Christian J.L. Hermes: "Least power point tracking (LPPT) control for refrigeration systems running with variable-speed compressors", International Journal of Refrigeration, available online January 9, 2021, discloses Least Power Point Tracking (LPPT) control algorithm applied to a refrigeration system. The LPPT is an adaptive version of a dual speed controller, which is an on-off controller with two non-zero speed values: the lower speed value is chosen to minimize power consumption, while the higher speed value is used to avoid strong disturbances to suppress. The LPPT and dual-speed results are compared to other control techniques that can be used in the field of refrigeration systems, namely the on-off (baseline) approach and the proportional-integral approach. Experimental analysis showed that the LPPT control uses approximately 16% less energy than the baseline approach.

In the article by Daniel J. Burns, Christopher R. Laughman and Martin Guay: "Proportional-Integral Extremum Seeking for Vapor Compression Systems", IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY, Vol.28, No.2, March 2020, a power consumption of a Vapor compression system (VCS) through the application of a proportional-integral extremum search controller (PI-ESC) that converges on the same time frame as the process. This extremum search method uses time-varying parameter estimation to determine the local gradient in the map from manipulated inputs to power outputs. In addition, the extremum search method contains terms proportional to the estimated gradient, which requires subsequent modification of an estimation routine to avoid bias. The PI-ESC algorithm is derived and compared to other methods on a benchmark example demonstrating the convergence rate of PI-ESC. PI-ESC is applied to the problem of selecting the compressor discharge temperature setpoint for a VCS so that power consumption is kept to a minimum. A physics-based simulation model of the VCS is used to show that with PI-ESC, convergence to the optimal operating point occurs faster than the bandwidth of typical perturbations - enabling the application of extremum seek control to VCSs in environments under realistic operating conditions. Finally, experiments on a production room air conditioner installed in an adiabatic test facility validate the approach in the presence of significant noise and actuator and sensor quantization.

It is the object of the present invention to at least partially overcome the disadvantages of the prior art and, in particular, to provide a possibility for particularly energy-saving operation of a refrigeration device with several refrigeration compartments, which can have different compartment temperatures, on the basis of a controller design that is comparatively easy to implement.

This object is solved according to the features of the independent claims. Preferred embodiments can be found in particular in the dependent claims.

• - einen Kältekreislauf mit einem drehzahlregelbaren Verdichter, einen Verflüssiger, mindestens einen Verdampfer und mindestens ein Drosselorgan,
• - einen Kühlraum mit mindestens einer Temperaturzone, der jeweils mindestens ein Verdampfer zugeordnet ist,
• - mindestens einen Temperatursensor pro Temperaturzone zum Abfühlen von Ist-Werten einer zugehörigen Zonentemperatur,
• - einen Satz von variabel über Stellgrößen einstellbaren Aktoren, einschließlich des Verdichters, welche einen merklichen Einfluss auf einen Betrieb des Kältekreislauf besitzen, wobei die Zahl der Aktoren um mindestens eins höher ist als die Zahl der Verdampfer,
• - ein Steuergerät, das dazu eingerichtet ist, Stellgrößen der Aktoren beruhend auf Abweichungen zwischen den für jede Temperaturzone gemessenen Ist-Werten der Zonentemperaturen und vorgegebenen Soll-Werten der Zonentemperaturen vorzugeben,

wobei

• - das Steuergerät mit einem computerlesbaren Speichermedium datentechnisch gekoppelt ist, das Befehle umfasst, die bei der Ausführung durch das Steuergerät dieses veranlassen, die Werte der Stellgrößen der Aktoren auf Grundlage eines Kreislaufmodells eines Kältesystems bestehend aus dem Kältekreislauf und der mindestens einen Temperaturzone, zu berechnen,
• - das Kreislaufmodell einen vollständigen Regelkreis zum Einstellen der Ist-Werte der Zonentemperaturen auf jeweils vorgegebene Soll-Werte beschreibt, in dem
• - das Kreislaufmodell, insbesondere dessen (modellhafte) Regelstrecke des Kreislaufmodells, ein dynamisches Teilmodell der mindestens einen Temperaturzone umfasst, das aus Werten von der jeweiligen Temperaturzonen zugeführten Kälteleistungen zugehörige Ist-Werte der Fachtemperaturen berechnet,
• - dem dynamischen Teilmodell ein, insbesondere ebenfalls der Regelstrecke zugehöriges, stationäres Teilmodell vorgeschaltet ist, das aus Stellgrößen der Aktoren jeweilige Werte der Kälteleistungen berechnet und an das dynamische Teilmodell übergibt sowie Werte mindestens einer weiteren physikalischen Zustandsgröße der Regelstrecke berechnet,
• - dem stationären Teilmodell ein dazu inverses stationäres Teilmodell vorgeschaltet ist, das aus vorgegebenen Soll-Werten der Kälteleistungen der jeweiligen Temperaturzonen sowie Vorgabewerten der mindestens einen weiteren Zustandsgröße Werte der Stellgrößen der Aktoren berechnet und an das stationäre Teilmodell übergibt,
• - dem inversen stationären Teilmodell pro Temperaturzone ein Reglermodell vorgeschaltet ist, das aus Differenzen zwischen den vorgegebenen Soll-Werten der zugehörigen Zonentemperatur und den gemessenen Ist-Werten der zugehörigen Zonentemperatur eine jeweilige Soll-Kälteleistung berechnet und an das inverse stationäre Teilmodell übergibt, und
• - ein Optimierer dazu vorgesehen ist, aus einem gemessenen Ist-Wert einer weiteren Zustandsgröße als Eingangsgröße einen Vorgabewert einer anderen weiteren Zustandsgröße zu bestimmen bzw. zu berechnen, diesen Vorgabewert an das inverse stationäre Teilmodell zu übergeben und die Vorgabewerte iterativ auf einen vorgegebenen Zielzustand der Eingangsgröße hin einzustellen.

The object is achieved by a refrigeration device having

• - a refrigeration circuit with a variable-speed compressor, a condenser, at least one evaporator and at least one throttle element,
• - a refrigeration room with at least one temperature zone, to which at least one evaporator is assigned,
• - at least one temperature sensor per temperature zone for sensing actual values of an associated zone temperature,
• - a set of actuators that can be variably adjusted via manipulated variables, including the compressor, which have a noticeable influence on the operation of the refrigeration cycle, with the number of actuators being at least one higher than the number of evaporators,
• - a control device that is set up to specify manipulated variables of the actuators based on deviations between the actual values of the zone temperatures measured for each temperature zone and specified target values of the zone temperatures,

whereby

• - the control unit is coupled in terms of data technology to a computer-readable storage medium, which includes commands which, when executed by the control unit, cause the latter to calculate the values of the manipulated variables of the actuators on the basis of a circuit model of a refrigeration system consisting of the refrigeration circuit and the at least one temperature zone,
• - the circuit model describes a complete control circuit for adjusting the actual values of the zone temperatures to respectively specified target values, in which
• - the circuit model, in particular its (model) controlled system of the circuit model, includes a dynamic partial model of the at least one temperature zone, which calculates associated actual values of the compartment temperatures from values of the cooling capacities supplied to the respective temperature zones,
• - the dynamic sub-model is preceded by a stationary sub-model, in particular also associated with the controlled system, which calculates respective values of the cooling capacities from manipulated variables of the actuators and transfers them to the dynamic sub-model and calculates values of at least one other physical state variable of the controlled system,
• - the stationary sub-model is preceded by an inverse stationary sub-model, which calculates values of the manipulated variables of the actuators from specified setpoint values of the cooling capacities of the respective temperature zones and default values of the at least one other state variable and transfers them to the stationary sub-model,
• - the inverse stationary partial model is preceded by a controller model for each temperature zone, which calculates a respective target cooling capacity from the differences between the specified target values of the associated zone temperature and the measured actual values of the associated zone temperature and transfers this to the inverse stationary partial model, and
• - an optimizer is provided to determine or calculate a default value of another additional state variable from a measured actual value of a further state variable as an input variable, to transfer this default value to the inverse stationary partial model and to iteratively adjust the default values to a predetermined target state of the input variable to adjust.

This refrigeration device has the advantage that an almost energy-optimal operation can be achieved, which can be maintained despite component scatter and aging effects because these can be compensated for by the control. It is also advantageous that the control via the circuit model is not only effective for individual stationary operating states, which are relevant for an energy classification, for example, but in principle for every stationary operating state. In particular, this results in the advantage that the further state variable of the stationary sub-model, which is used as the input variable of the optimizer, is understood as a controlled variable that is constantly monitored during operation and, if necessary, constantly readjusted or readjusted. A further advantage is that the controlled system is linearized via the inverse stationary partial model. This allows the use of comparatively simple controller design methods for linear systems for controlling the individual zone temperatures. There is also an advantage in that the method described there can only be used for devices with one compartment and one compressor as an actuator. A simple expandability to devices with multiple compartments and multiple, possibly different, actuators does not exist.

The refrigeration device is in particular a domestic refrigeration device. The refrigeration device can be, for example, a refrigerator, a chest freezer or a combination thereof.

The fact that the actuators can be variably adjusted using manipulated variables includes, in a further development, that the manipulated variables can be adjusted continuously or quasi-continuously (i.e. with small increments). This enables a particularly precise achievement of low power consumption in the refrigeration cycle.

A speed-controllable compressor can be understood to mean a compressor which has a variably adjustable speed and/or a variably adjustable delivery volume.

The throttle element can be a valve, a capillary, a nozzle, etc.

The fact that at least one evaporator is assigned to a temperature zone means that this temperature zone is cooled by means of the at least one evaporator. For this purpose, the temperature zone is thermally coupled to the associated at least one evaporator. For example, the evaporator can be present in the temperature zone, represent a wall of the temperature zone, or be arranged directly behind a wall of the temperature zone.

The refrigeration circuit and the temperature zone(s) together form the refrigeration system.

The refrigerator compartment can have one temperature zone or several temperature zones with individually adjustable zone temperatures.

At least one temperature sensor for sensing an actual zone temperature of this temperature zone is assigned to each temperature zone.

An actuator that has or has a noticeable influence on the operation of the refrigeration circuit can be understood in particular as a component of the refrigeration device whose operation or which during operation significantly influence the behavior of the refrigeration circuit, in particular its operating point. In particular, all desired steady-state operating points with regard to zone temperatures and power consumption can in principle be set together by such actuators. Such actuators can include, for example: the compressor, throttle bodies with adjustable flow, fans acting specifically on the refrigeration circuit, etc. Since the refrigeration circuit has at least one evaporator, the set of actuators comprises at least two actuators, namely in particular the compressor and at least one other actuator.

The control unit is used to control the refrigeration circuit and, if necessary, other functions of the refrigeration device. For this purpose, the control unit can in particular have an electronic circuit and at least one data memory, on which commands and possibly data for the operation of the control unit are stored. For example, the electronic circuit can be a microprocessor, ASIC, FPGA, etc.

The control unit can adjust the manipulated variables of the actuators, for example based on deviations between the actual values measured for each temperature zone and specified target values of the associated zone temperatures, in particular within the framework of a control circuit in which the actual values of the zone temperatures measured by the temperature sensors are fed back as control variables, are compared with the target values serving as reference variables, and their differences are fed to a control device as control deviations, which outputs corresponding control variables to the actuators.

The fact that the control unit is coupled in terms of data technology to a computer-readable storage medium can include that the storage medium is a data memory of the control unit, or can include that the storage medium is a separate unit that can exchange data with the control unit, e.g. one that can be used in the refrigeration device Memory card or even an external instance like cloud storage.

The fact that the storage medium includes instructions which, when executed by the control unit, cause the control unit to calculate the values of the manipulated variables of the actuators on the basis of a circuit model includes, in particular, that the circuit model is stored in the storage medium and the control unit can run the circuit model computationally, receives the values of the manipulated variables of the actuators from this and then sets these values on the actuators.

The modeled controlled system of the circuit model includes, in particular, the inverse stationary sub-model connected in series, behind it the stationary sub-model and behind it the dynamic sub-model. The stationary sub-model and the dynamic sub-model together describe the controlled system of the real refrigeration device using state equations. The inverse stationary part model corresponds to an inverse description of the stationary part model.

By dividing the model of the controlled system of the real refrigeration system into a dynamic part and a stationary part, the dynamics of the refrigeration cycle and the temperature zones take place on different time scales. The time constants of the refrigeration cycle are usually in the range of seconds to a few minutes, whereas the dynamics of the temperature zones typically take place in the range of many minutes to a few hours. This is specifically exploited in the present model-based control draft in order to divide the overall model of the controlled system of the refrigeration system into at least two sub-models, namely a sub-model of the refrigeration circuit (also referred to as "refrigeration circuit model") and a sub-model of the temperature zone(s). Since the time constants of the temperature zone(s) are dominant, only this partial model is dynamically modeled, so that it is referred to below as the "dynamic" partial model. The refrigeration circuit, on the other hand, is mapped as a stationary model and is referred to below as a "stationary" partial model. The dynamic partial model for the temperature zone(s) is represented by linear Equations of state described, ie by a set of linear differential equations with constant coefficients. The stationary partial model for the refrigeration circuit, on the other hand, is described by non-linear state equations, which can, however, be purely algebraic equations. A series connection of a non-linear, stationary sub-model and a dynamic, linear sub-model is known in principle as the Hammerstein system and is therefore not described further.

The dynamic sub-model is designed to use the equations of state stored for this purpose to calculate associated actual values of the zone temperatures from calculated values of cooling capacities Q of the respective temperature zones or cooling capacities Q supplied from the respective temperature zones.

The stationary sub-model is designed to calculate respective values of the cooling power Q from the values of the manipulated variables of the actuators and to transfer them to the dynamic sub-model. Depending on the type of state equations used, which physically describe the refrigeration circuit, at least one other physical state variable of the controlled system is calculated by the stationary sub-model, e.g condenser) and/or refrigerant temperatures at various points along the refrigeration cycle, etc.

The combination of a stationary sub-model and a dynamic sub-model thus again maps the controlled system of the real refrigeration system with the values of the manipulated variables of the actuators as manipulated variables and the actual values of the temperature zones as controlled variables.

Because the stationary part model is preceded by an inverse stationary part model that is as accurate as possible and corresponding thereto, the advantage is achieved that the non-linearities of the controlled system are largely (ideally completely) compensated.

So that the inverse stationary partial model is determined unambiguously, it advantageously has the same number of input and output variables, and should therefore advantageously be neither overdetermined nor underdetermined. The output variables of the inverse stationary sub-model are then identical to the input variables of the refrigeration system and thus here of the stationary sub-model. For the selected example, this means that the inverse stationary partial model can clearly determine the values of all manipulated variables of the actuators from the same number of input variables.

As already indicated above, in addition to the refrigeration capacity Q for the respective temperature zones and the electrical power consumption, other physical state variables can also be read from the stationary partial model or refrigeration cycle model, e.g various points along the refrigeration circuit, etc. In principle, there is freedom in the selection of the remaining input variable(s) for the inverse stationary partial model. If the cooling capacities Q are to be specified for the respective temperature zones in the present case, it is advantageous to set these as input variables for the inverse stationary partial model. At least one input variable from the group of other state variables for the inverse model is defined below. The number of other state variables still to be defined corresponds to the difference between the number of actuators and the number of temperature zones and is at least one.

If the inverse stationary sub-model were perfect, the target values of the refrigerating capacities Q specified for the inverse stationary sub-model would be identical to the model (actual) values of the refrigerating capacities Q transferred from the stationary sub-model to the dynamic sub-model. It is a further development that this assumption is made in the present case.

Since the inverse stationary partial model compensates for the non-linearities of the controlled system, the partial controlled systems of the temperature zones modified in this way are then ideally linear and decoupled. This results in the advantage that the temperature controller(s) connected upstream of the inverse stationary partial model on the input side can be implemented particularly easily. The target values of the respective refrigerating capacities Q can thus be generated separately for each temperature zone using a separate model temperature controller (controller model). It is therefore advantageously possible to use associated temperature controllers for each temperature zone, which calculate a respective target cooling capacity from differences between the specified target values of the associated zone temperature and the actual values of the associated zone temperature fed back by the dynamic sub-model and send it to the inverse stationary sub-model hands over.

A development is that the controller model is a model of a linear controller or a controller for linear systems, e.g. a PI controller, a PID controller, etc. A linear controller is particularly robust and easy to implement.

If the model control circuit above is designed correctly, the actual values of the zone temperatures are adjusted asymptotically to the respective target values. At the same time, the default value or the default values for the remaining at least one other state variable, which has been selected as the input variable(s) of the inverse stationary model, must be set at all times in such a way that the associated cooling capacities Q are stationary at this at least one zone temperature from Refrigeration device can be displayed.

The respective default value for the remaining at least one further state variable is advantageously selected such that one of the further state variable(s) calculated in the stationary partial model is set to a predetermined target state. For this purpose, a corresponding optimizer or optimization algorithm is provided, which at least stationary based on a value of a state variable calculated in the stationary circuit model, calculates the default value of another state variable used as an input variable of the inverse stationary partial model and transfers it to the inverse stationary partial model. The optimizer is also set up to iteratively adjust the default values of the state variable used as the input variable of the inverse stationary partial model to a specified or desired target state of the input variable of the optimizer, e.g. by targeted, incremental change of the default value so that the input variable of the optimizer has a maximum - or minimum value reached. To do this, the optimizer can use a gradient method, for example.

The further state variable used as the input variable of the optimizer is in particular a physically different variable than the further state variable transferred to the inverse stationary partial model.

In principle, the number of evaporators, temperature zones and actuators is not limited as long as the number of evaporators is less than the number of actuators.

In one embodiment, the number of actuators is one more than the number of evaporators. This advantageously enables a particularly simple design of the control model.

In one configuration, the input variable of the optimizer includes or is an electrical power consumption, i.e. the electrical power consumption corresponds to the (at least one) additional state variable, on the basis of which the optimizer determines the default value of another additional state variable, which is transferred to the inverse stationary partial model, and the target state is minimum power consumption. So, in addition to the zone temperatures, the power consumption of the refrigeration circuit is specifically considered as an additional setting or control variable. As a result, the advantage is achieved that the refrigeration device can adjust itself automatically with little implementation effort in such a way that, at least in the stationary state, a low, in particular almost minimal, power consumption is achieved. This energy optimization can not only be achieved with individual excellent steady-state operating states (e.g. those that are relevant for an energy classification), but in principle with every steady-state operating state that is determined by the target zone temperatures.

One configuration is that the refrigeration device has a device for determining an electrical power consumption of the refrigeration circuit, the input variable of the optimizer being a power consumption determined from the operation of the refrigeration device (e.g. a measured or estimated, and not just calculated from the stationary model). This achieves the advantage that the energy-minimizing setting of the electrical power consumption can be maintained despite component scatter and aging effects because these effects are compensated for by the optimizer. The device for determining the electrical power consumption can be, for example, a current measuring device and/or a device for estimating the electrical power consumption, e.g. an electronic inverter. For example, the inverter electronics can also be controlled based on a model and can then output an estimated value for the power consumption of the compressor from its model.

The electrical power consumption of the refrigeration circuit can be understood as the electrical power consumption of the actuators of the refrigeration circuit, i.e. the compressor plus the other actuators considered in the refrigeration circuit model, if these consume power in a stationary manner. In particular, if the other actuators considered in the refrigeration circuit model consume only little power, especially in the stationary state, consume no or practically no power, the electrical power consumption of the refrigeration circuit can be approximated by the electrical power consumption of the compressor.

In one embodiment, the default values output by the optimizer to the inverse stationary model (for use as an input variable there) are values of a final compression temperature of the compressor. The use of the compression end temperature as an input variable of the inverse stationary model has the advantage that the power consumption of the refrigeration tekreiss can be adjusted continuously, since it is itself also continuous and always positive. For this reason, it enables a simpler description or compilation of the equations of state of the stationary sub-model as, for example, physical quantities that can also be used in principle, such as superheating or supercooling temperatures, for which negative, physically meaningless values would have to be specified. In addition, it is easily accessible in terms of measurement technology and therefore facilitates the model comparison with measurement data from a real system.

It is a particularly advantageous development that the input variable of the optimizer is the electrical power consumption and the default variable output by the optimizer to the inverse stationary model is the final compression temperature.

In one configuration, the dynamic partial model, the stationary partial model and the inverse stationary partial model describe an ambient temperature of the refrigeration device as a disturbance variable. As a result, one of the most influential variables for the operation of the refrigeration system, in particular a power consumption, is advantageously taken into account. For example, at least some of the state equations on which the partial models are based can have the ambient temperature as a parameter. For example, the stationary sub-model can use the manipulated variables of the actuators to calculate cooling capacity values that are dependent on the ambient temperature and/or the dynamic sub-model can calculate the actual values of the compartment temperatures as a function of the ambient temperature, etc.

In one configuration, the manipulated variable of the compressor is its speed. For this purpose, the compressor can have, for example, a speed-controllable electric motor (e.g. a BLDC motor) as a drive unit.

In one configuration, the set of actuators includes at least one throttle element in the form of a variably adjustable expansion valve, the manipulated variable of which corresponds to the position of the expansion valve, in particular a flow cross section. Alternatively or additionally, the throttling element can, for example, comprise a capillary, etc., cooled by a fan.

In one configuration, at least one temperature zone corresponds to a refrigerated compartment and the associated zone temperatures correspond to compartment temperatures.

However, the type of temperature zone is not limited in principle. In a further development, the refrigerated space of the refrigeration device can represent a single temperature zone, or the refrigerated space can correspond to a temperature zone, and one or more refrigerated compartments separated from it (e.g. for vegetables, meat, etc.) can correspond to further temperature zones.

In one configuration, the refrigerated space has a plurality of temperature zones, each of which is assigned exactly one evaporator. This simplifies a structural design of the refrigeration device and the model description of the refrigeration system.

In a further development, at least one evaporator is connected in series before or after each throttle element in the flow of the refrigerant of the refrigeration cycle. It is a further development that a number of throttling elements corresponds to a number of evaporators. In a further development, a throttle element and an evaporator are each connected in series without a branch in between.

In one configuration, the refrigeration device is set up to operate the compressor continuously. This advantageously enables the optimizer to work continuously or quasi-continuously and thus to adjust its input variable to the target state. In the case of a speed-controlled compressor, continuous operation of the compressor can be understood as an operating mode in normal use, with compartment temperatures being several °C lower than the ambient temperature, so that heat is constantly being introduced into the refrigeration device, the dissipation of which requires at least the minimum speed of the compressor. At low ambient temperatures, the compressor switches off.

Both for an operating mode with continuous operation of the compressor and for an operating mode with an intermittently working compressor, the compressor switches off when a minimum adjustable total cooling capacity is higher than a stationary required total cooling capacity.

• - Werte der Stellgrößen der Aktoren beruhend auf Abweichungen zwischen den für jede Temperaturzone gemessenen Ist-Werten der Zonentemperaturen und vorgegebenen Soll-Werten der Zonentemperaturen vorgegeben werden,
• - die Werte der Stellgrößen der Aktoren auf Grundlage eines Kreislaufmodells berechnet werden, das einen vollständigen Regelkreis zum Einstellen der Ist-Werte der Zonentemperaturen auf die Soll-Werte der Zonentemperaturen beschreibt, in dem
• - eine Regelstrecke des Regelkreises des Kreislaufmodells ein dynamisches Teilmodell der mindestens einen Temperaturzone umfasst, das aus Werten von Kälteleistungen der jeweiligen Temperaturzonen zugehörige Ist-Werte der Zonentemperaturen berechnet,
• - dem dynamischen Teilmodell ein stationäres Teilmodell der Regelstrecke vorgeschaltet ist, das aus den Werten der Stellgrößen der Aktoren die jeweiligen Werte der Kälteleistungen berechnet und an das dynamische Teilmodell übergibt sowie Ist-Werte mindestens einer weiteren Zustandsgröße der Regelstrecke berechnet,
• - dem stationären Teilmodell ein dazu inverses stationäres Teilmodell vorgeschaltet ist, das aus vorgegebenen Soll-Werten der Kälteleistungen der jeweiligen Temperaturzonen sowie Vorgabewerten der mindestens einen weiteren Zustandsgröße die Werte der Stellgrößen der Aktoren berechnet und an das stationäre Teilmodell übergibt,
• - dem inversen stationären Teilmodell pro Temperaturzone ein Reglermodell vorgeschaltet ist, das aus Differenzen zwischen den vorgegebenen Soll-Werten der zugehörigen Zonentemperatur und den von dem dynamischen Teilmodell rückgeführten Ist-Werten der zugehörigen Zonentemperatur jeweilige Soll-Werte der Kälteleistungen berechnet und an das inverse stationäre Teilmodell übergibt, und
• - ein Optimierer aus einem gemessenen Ist-Wert einer weiteren Zustandsgröße einen Vorgabewert einer anderen weiteren Zustandsgröße bestimmt, diesen Vorgabewert an das inverse stationäre Teilmodell übergibt und die Vorgabewerte iterativ auf einen vorgegebenen Zielzustand der Eingangsgröße hin einstellt.

The object is also achieved by a method for operating a refrigeration device, the refrigeration device having a refrigeration circuit with a variable-speed compressor, a condenser, at least one evaporator and at least one throttle element, a refrigerated space with at least one temperature zone, which is at least one evaporator is assigned, at least one temperature sensor per temperature zone for sensing actual values of an associated zone temperature and a set of actuators that can be adjusted variably via manipulated variables, including the compressor, which have a noticeable influence on the operation of the refrigeration circuit, the number of actuators is higher than the number of evaporators, wherein in the method

• - Values of the manipulated variables of the actuators are specified based on deviations between the actual values of the zone temperatures measured for each temperature zone and the specified target values of the zone temperatures,
• - the values of the manipulated variables of the actuators are calculated on the basis of a circuit model which describes a complete control circuit for adjusting the actual values of the zone temperatures to the target values of the zone temperatures, in which
• - a controlled system of the control circuit of the cycle model includes a dynamic partial model of at least one temperature zone, which calculates associated actual values of the zone temperatures from values of cooling capacities of the respective temperature zones,
• - the dynamic sub-model is preceded by a stationary sub-model of the controlled system, which calculates the respective values of the cooling capacities from the values of the manipulated variables of the actuators and transfers them to the dynamic sub-model and calculates actual values of at least one other state variable of the controlled system,
• - the stationary sub-model is preceded by an inverse stationary sub-model, which calculates the values of the manipulated variables of the actuators from specified target values of the cooling capacities of the respective temperature zones and specified values of the at least one other state variable and transfers them to the stationary sub-model,
• - the inverse stationary sub-model is preceded by a controller model for each temperature zone, which calculates respective setpoint values for the cooling capacities from the differences between the specified setpoint values for the associated zone temperature and the actual values of the associated zone temperature fed back by the dynamic sub-model and sends them to the inverse stationary sub-model passes, and
• - an optimizer determines a default value of another additional state variable from a measured actual value of a further state variable, transfers this default value to the inverse stationary partial model and iteratively adjusts the default values to a predetermined target state of the input variable.

The method can be designed analogously to the refrigeration device and vice versa, and has the same advantages.

In one embodiment, the optimizer works continuously, that is to say (quasi-)continuously sets the default values iteratively to a default target state of the input variable during operation of the refrigeration circuit or the refrigeration device. This can also be described as the optimizer running continuously to set the target state of its input variable. In the stationary state, this target state will typically be reached after a certain time or after a certain number of iterations, in which case the values of the manipulated variables can then remain the same in particular. If the stationary state is canceled by external influences such as a door being opened, a change in the ambient temperature, etc., the optimizer continues to try to set the target state of its input variable, which may result in other manipulated variable values.

• 1 zeigt eine Skizze eines Kältekreislaufs eines Kältegeräts mit zwei Verdampfern und drei Aktoren;
• 2 zeigt ein Blockdiagramm von Eingangs- und Ausgangsgrößen eines Kältesystems des Kältegeräts aus 1;
• 3 zeigt ein Blockdiagramm eines Modells der Regelstrecke des Kältegerätes aus 1 als Hammerstein-System;
• 4 zeigt ein Blockdiagramm des Modells der Regelstrecke aus 3 mit zusätzlich vorgeschaltetem inversen Teilmodell;
• 5 zeigt ein Blockdiagramm des Modells der Regelstrecke aus 4 unter vereinfachenden Annahmen;
• 6 zeigt ein Blockdiagramm eines Modells eines Regelkreises unter Verwendung der Regelstrecke aus 5; und
• 7 zeigt ein Blockdiagramm einer Umsetzung des Modells des Regelkreises aus 7 in dem Kältegerät aus 1.

The characteristics, features and advantages of this invention described above, and the manner in which they are achieved, will become clearer and more clearly understood in connection with the following schematic description of an exemplary embodiment, which will be explained in more detail in connection with the drawings.

• 1 shows a sketch of a refrigeration circuit of a refrigerator with two evaporators and three actuators;
• 2 shows a block diagram of input and output variables of a refrigeration system of the refrigeration device 1 ;
• 3 shows a block diagram of a model of the controlled system of the refrigeration appliance 1 as Hammerstein system;
• 4 shows a block diagram of the model of the plant 3 with an additional upstream inverse partial model;
• 5 shows a block diagram of the model of the plant 4 under simplifying assumptions;
• 6 FIG. 12 shows a block diagram of a model of a control loop using the plant 5 ; and
• 7 Figure 12 shows a block diagram of an implementation of the control loop model 7 in the refrigerator 1 .

1 shows a sketch of a refrigeration circuit of a refrigeration device 1 with two evaporators 2, 3, which apply the respective refrigeration capacities Q̇ ₁ or Q̇ ₂ for cooling respective temperature zones that are not shown, and three actuators, namely a compressor 4 with adjustable speed n and electrical power consumption P _el , an expansion valve 5 with an adjustable flow cross section d ₁ and an expansion valve 6 with an adjustable flow cross-section d ₂ . The expansion valve 5 is in series with the evaporator 2 and the expansion valve 6 is in series with the evaporator 3. The two valve/evaporator pairs 5, 2 and 6, 3 are in turn connected in series with one another. The refrigeration circuit also has a condenser 7 .

The refrigeration cycle is part of a refrigeration system 10, to which each evaporator 2, 3 is also assigned a temperature zone in the form of a refrigerated compartment (not shown). The dynamics of the refrigeration cycle and the cooling compartments take place on different time scales. The time constants of the refrigeration circuit 2 to 7 are usually in the range from seconds to a few minutes, whereas the time constants of the cooling compartments are typically in the range from many minutes to a few hours.

Extensions to refrigeration circuits with more cooling compartments and more actuators can be implemented analogously, as long as the number of actuators is higher than the number of evaporators, in particular by one more.

2 shows a block diagram of the input and output variables of the refrigeration system 10 of the refrigeration appliance 1. Adjusting the compressor speed n and the flow cross-sections _d1 and _d2 results in the actual values T _F1,meas and T F measured by the respective temperature sensors (not shown). _F2,mess of the compartment temperatures assigned to the evaporators 2 or 3 and the electrical power P _el,mess expended for the operation of the refrigeration circuit 2 to 7 . The ambient temperature T _amb represents a significant influencing or disturbance variable.

The fact that the dynamics of the refrigeration circuit and the cooling compartments have different time scales can be exploited in a model-based control design by dividing the controlled system of the refrigeration system into two sub-models, namely a sub-model MK _stat for the refrigeration circuit and a sub-model MT _dyn for the cooling compartments. Since the time constants of the cooling compartments are dominant, only the "dynamic" partial model MT _dyn corresponding to the cooling compartments is dynamically modeled and the refrigeration cycle is mapped as a "stationary" partial model MK _stat , as in 3 shown as a block diagram.

The dynamic partial model MT _dyn has a linear structure, ie it can be described by a set of linear differential equations with constant coefficients. In contrast, the stationary partial model MK _stat is non-linear, but can be described by purely algebraic equations. Such a series connection of a non-linear, stationary model and a dynamic, linear model is usually referred to as a Hammerstein system. The usual approach when controlling a Hammerstein system is to precede the stationary partial model with a corresponding inverse (partial) model MKinv _stat that is as accurate as possible, so that the non-linearities of the controlled system are largely (ideally completely) compensated, as in 4 represented by a block diagram.

In order for the inverse stationary partial model MKinv _stat to be uniquely determined, it must have the same number of input variables and output variables. The output variables of the inverse stationary sub-model MKinv _stat are identical to the input variables of the refrigeration circuit model or the stationary sub-model MK _stat . For the selected exemplary embodiment, this means that the inverse stationary partial model _MKinvstat determines the values of all three manipulated variables n, d ₁ , d ₂ from the same number of input variables. In addition to the cooling capacities Q̇̇ ₁ and Q̇̇ ₂ for the individual cooling compartments, further physical variables can be read out or calculated from the refrigeration circuit model, namely the power consumption P _el,calc already mentioned above and, for example, a compressor end temperature T _di , degree of filling of heat exchangers 2 , 3, 7 and refrigerant temperatures at various points along the refrigeration circuit (not shown), etc. There is therefore fundamental freedom in the selection of the input variables of the inverse stationary partial model MKinv _stat . However, the input variables should be specified in a suitable way by the control concept.

Since the target refrigerating capacities Q - _{1, target} and Q - _{2, target} for the individual cooling compartments are to be specified in the present exemplary embodiment, they are set as input variables of the inverse stationary partial model MKinv _stat . Since the number of actuators 2, 3, 4 is one greater here than the number of evaporators 2, 3 or cooling compartments, an input variable for the inverse stationary partial model MKinv _stat must also be specified. This input variable is advantageously selected in such a way that the power consumption P _e,calcl can be set as continuously as possible through its specification. For this reason, overheating or undercooling temperatures, for example, are less suitable because negative and therefore physically meaningless values would have to be specified for them. On the other hand, a physical variable that meets the above requirement is the compression end temperature T _di , since this is always positive.

If the inverse stationary partial model MKinv _stat were perfect, then the in 5 Structure shown _by means of a block diagram, in which the actual values of the cooling capacities Q̇̇ ₁ and Q̇ ₂ . that of the inverse stationary part model MKinv _stat given values of the target cooling capacity Q̇ _1, set or Q̇ _2,set correspond, i.e. Q̇̇ ₁ = Q̇ _1,set and Q̇̇ ₂ = Q̇ _2,set applies. The following then also applies to the compression end temperature T _di = T _di,ref . This situation is assumed for the rest of the draft regulation.

Since the partial controlled systems modified by the compensation process are now ideally linear and decoupled from the target refrigerating capacities Q̇ _1,set or Q̇ _2,set to the respective associated model-calculated compartment temperatures T _F1,calc or T _F2,calc , proven and easy-to-implement design methods for model temperature controllers (“controller models”) TR ₁ and TR ₂ are used, such as 6 illustrated. Using the temperature controllers TR ₁ and TR ₂ , the target cooling capacities Q̇ _1,set and Q̇ _2,set for each refrigerated compartment separately based on the specified set values T _F1,set and T _F2,set of the calculated compartment temperatures T _F1,calc and T _F2,calc generated. The temperature controllers TR ₁ and TR ₂ can be modeled as PI controllers, for example.

If the control circuit is designed correctly, the two calculated compartment temperatures _TF1,calc and _{TF2,calc are} set asymptotically to the respective setpoint values _TF1,setpoint and _TF2,setpoint . At the same time, the default value T _di,ref for the compression end temperature T _di is selected at all times in such a way that the two refrigeration capacities Q̇̇ ₁ and Q̇̇ ₂ can be displayed by the refrigeration device 1 in a stationary manner at these compartment temperatures T _F1,calc and T _F2,calc .

In the present exemplary embodiment, the default variable T _di,ref is selected in such a way that an almost minimal power consumption P el _{,soll is achieved as the target state for the power consumption P el, calc} , at least in the stationary state. For this purpose, the signal P _el,calc is made available to an optimization algorithm or "optimizer" OPT, which uses it to determine the value for T _di,ref that is suitable for the lowest power consumption P _el, soll as the target state. This can be achieved, for example, by purposefully incrementally changing T _di,ref until a minimum of P _el,calc is found.

With such a design of the optimizer OPT, the final compression temperature T _di is advantageously used as an operand in the optimizer OPT and in the partial models MKinv _stat and MKstat, but does not need to be measured on the real refrigeration device 1 . A dedicated temperature sensor for measuring the compression end temperature T _di is therefore not required.

7 Figure 12 shows a block diagram of an implementation of the control loop model 6 in the refrigerating appliance 1. The manipulated variables n, d ₁ , d ₂ are set by means of a device control 9 which is coupled to a data memory 8 in terms of data technology, eg has a data memory 8 which includes commands which, when executed by the control unit 8, cause the latter to calculate the values of the manipulated variables n, d ₁ , d ₂ on the basis of the cycle model of the refrigeration system 1 described above. In other words, this is in 6 Shown model of the refrigeration system 10 and modified for operation on the real refrigeration device 1 stored in the data memory 8 programmed.

In particular, the model of the refrigeration system 10 executed by the device controller 9 differs from that in 6 shown model in that the input variable supplied to the optimizer OPT is not the electrical power consumption calculated from the stationary partial model MK _stat , but rather an electrical power consumption P _el,mess determined by the refrigeration appliance 1 from operation (e.g. measured or estimated by querying an electronic inverter). . This advantageously ensures that model errors with regard to the description of the calculated power consumption P _el,calc on the real refrigeration device 1 are compensated for.

The compartment temperatures T _F1,calc , T _F2,calc calculated by the stationary partial model MT _dyn are also not fed back and then fed back negatively with the corresponding setpoint values, but rather the compartment temperatures T _F1,meas , T _F2,meas actually measured by the temperature sensors .

The model stored in the data memory 8 therefore only requires the compartment temperatures T _F1,mess , T _F2,mess typically measured today to run on the device control 9 and the electrical power consumption P _el,mess typically already determined today and is therefore particularly easy to implement in terms of design .

Of course, the present invention is not limited to the embodiment shown.

The control concept was explained in more detail above using a refrigeration device with two temperature zones and three actuators (two expansion valves and one compressor) as an example. However, this concept can be transferred analogously to refrigeration devices with a different number of temperature zones and/or different types of actuators.

In general, "a", "an" etc. can be understood as a singular or a plural number, in particular in the sense of "at least one" or "one or more" etc., as long as this is not explicitly excluded, e.g. by the expression "exactly a" etc.

A numerical specification can also include exactly the specified number as well as a usual tolerance range, as long as this is not explicitly excluded.

Reference List

1

refrigeration device

2

Evaporator

3

Evaporator

4

compressor

5

expansion valve

6

expansion valve

7

condenser

8th

data storage

9

device control

10

refrigeration system

d1

Flow cross section of the expansion valve 5

d2

Flow cross section of the expansion valve 6

MTdyn

Dynamic part model

MKinvstat

Inverse stationary part model

MKstat

Stationary part model

n

number of revolutions

opt

optimizer

Pel

Electric power consumption

Q̇̇1

Cooling capacity of the evaporator 2

Q̇̇2

Cooling capacity of the evaporator 3

Q̇1,set

Desired cooling capacity of the evaporator 2

Q̇2,set

Desired cooling capacity of the evaporator 3

Tamb

ambient temperature

Tdi

Actual value of the final evaporation temperature

Tdi, ref

Default value for the final evaporation temperature

TF1

Actual value of the compartment temperature of a first cooling compartment

TF2

Actual value of the compartment temperature of a second cooling compartment

TF1, target

Target value of the compartment temperature of a first cooling compartment

TF2, target

Target value for the compartment temperature of a second cooling compartment

TR1

First temperature controller

TR2

Second temperature controller

Claims (14)

Refrigeration device (1), having - a refrigeration circuit (2-7) with a variable-speed compressor (4), a condenser (7), at least one evaporator (2, 3) and at least one throttle element (5, 6), - a cooling chamber with at least one temperature zone, to which at least one evaporator (2, 3) is assigned, - at least one temperature sensor per temperature zone for sensing actual values (T _F1 , T _F2 ) of an associated zone temperature, - a set of manipulated variables (n, d ₁ , d ₂ ) adjustable actuators (4-6), including the compressor (4), which have a noticeable impact on operation of the refrigeration cycle (2-7), the number of actuators (4-6) being higher than the number of evaporators (2, 3), and - a control unit (9) which is set up to calculate values of the manipulated variables (n, d ₁ , d ₂ ) of the actuators (4-6) based on deviations between the actual values (T _F1 , T _F2 ) of the zone temperatures and specified setpoint values (T _F1,setpoint , T _F2,setpoint ) thereof, wherein - the control unit (9) is data-technically coupled to a computer-readable storage medium (8) which comprises commands, which, when executed by the control unit (9), cause the latter to calculate the values of the manipulated variables (n, d ₁ , d ₂ ) of the actuators (4-6) on the basis of a circuit model, - the circuit model has a complete control circuit for setting the actual - Values (T _F1,calc , T _F2,calc ) of the zone temperatures to predetermined target values (T _F1,target , T _F2,target ) describes, in which - the circuit model is a dynamic partial model (MT _dyn ) of the at least one temperature zone which calculates associated values (T _F1,calc , T _F2,calc ) of the zone temperatures from values (Q̇ ₁ , ,Q̇ ₂ ) of cooling capacities of the respective temperature zones, - the dynamic partial model (MT _dyn ) a stationary partial model (MK _stat ) is connected upstream, which calculates the respective values (Q̇ ₁ , ,Q̇ ₂ ,) of the cooling capacities from the values of the manipulated variables (n, d ₁ , d ₂ ) of the actuators (4-6) and transfers them to the dynamic partial model (MT _dyn ). and values (P _el,calc , T _di ) of at least one other state variable of the circuit model are calculated, - the stationary sub-model (MK _stat ) is preceded by an inverse stationary sub-model (MKinv _stat ) which is based on specified target values (Q̇ _1,setpoint ,Q̇ _2,soll ) of the cooling capacities of the respective temperature zones as well as default values (T _di,ref ) of the at least one other state variable, the values of the manipulated variables (n, d ₁ , d ₂ ) of the actuators (4-6) are calculated and sent to the stationary partial model (MK _stat ) passes, - the inverse stationary partial model (MKinv _stat ) is preceded by a controller model for each temperature zone, which is calculated from the differences between the specified setpoint values (T _F1,setpoint , T _F2,setpoint ) of the associated zone temperature and the measured actual values (T _F1,mess , T _F2,meas ) of the associated zone temperature calculates respective setpoint values (Q̇ _1,setpoint ,Q̇ _2,setpoint ) of the cooling capacities and transfers them to the inverse stationary partial model (MKinv _stat ), and - an optimizer (OPT) is provided for this purpose, to determine a default value (T _di,ref ) of another additional state variable (T _di ) from a measured actual value (P el _, mess ) of a further state variable as an input variable, to transmit this default value (T _di,ref ) to the inverse stationary partial model ( MKinv _stat ) and to set the default values (T _di,ref ) iteratively to a default target state (P _el,soll ) of the input variable (P _el,mess ). Refrigeration device (1) after claim 1 , whereby the number of actuators (4-6) is one higher than the number of evaporators (2, 3). Refrigeration appliance (1) according to one of the preceding claims, wherein the input variable-SSe of the optimizer (OPT) comprises or is an electrical power consumption (P _el,calc , P _el,mess ). Refrigeration device (1) after claim 3 , wherein the refrigeration device (1) has a device for determining the electrical power consumption (P _el,mess ) of the refrigeration circuit (2-7) and the input variable of the optimizer (OPT) is the determined electrical power consumption (P _el,mess ). Refrigeration appliance (1) according to one of the preceding claims, wherein the default values (T _di,ref ) output by the optimizer (OPT) to the inverse stationary model (MKinv _stat ) are values of a compression end temperature (T _di ) of the compressor (4). Refrigeration appliance (1) according to one of the preceding claims, wherein the dynamic partial model (MT _dyn ), the stationary partial model (MK _stat ) and the inverse stationary partial model (MKinv _stat ) describe an ambient temperature ( _Tamb ) of the refrigeration appliance (1) as a disturbance variable. Refrigerating appliance (1) according to one of the preceding claims, wherein the manipulated variable of the compressor (4) is its speed (n). Refrigerating appliance (1) according to one of the preceding claims, wherein the set of actuators (4-6) comprises at least one throttle element in the form of a variably adjustable expansion valve (5, 6), the manipulated variable of which corresponds to a flow cross section (d ₁ , d ₂ ) of the expansion valve ( 5, 6). Refrigeration appliance (1) according to one of the preceding claims, wherein the temperature zones correspond to refrigerated compartments and the zone temperatures correspond to compartment temperatures. Refrigerating appliance (1) according to one of the preceding claims, in which the refrigerated space has a plurality of temperature zones, each of which is assigned precisely one evaporator (2, 3). Refrigerating appliance (1) according to one of the preceding claims, wherein a number of throttle elements (5, 6) corresponds to a number of evaporators (2, 3) and one throttle element (5, 6) and one evaporator (2, 3) each without branching in between are connected in series. Refrigeration device (1) according to one of the preceding claims, wherein the refrigeration device (1) is set up to operate the compressor (4) in continuous operation. Method for operating a refrigeration device (1), wherein the refrigeration device (1) has a refrigeration circuit (2-7) with a variable-speed compressor (4), a condenser (7), at least one evaporator (2, 3) and at least one throttle element (5 , 6), a refrigerator with at least one temperature zone, to which at least one evaporator (2, 3) is assigned, at least one temperature sensor per temperature zone for sensing actual values (T _F1,mess , T _F2,mess ) of an associated zone temperature and a set of variably controlled variables (n, d ₁ , d ₂ ) adjustable actuators (4-6), including the compressor (4), which have a noticeable influence on the operation of the refrigeration cycle (2-7), the number of Actuators (4-6) is higher than the number of evaporators (2, 3), wherein in the method - values of the manipulated variables (n, d ₁ , d ₂ ) of the actuators (4-6) based on deviations between the actual values (T _F1,meas , T _F2,meas ) of the zone temperatures measured for each temperature zone and specified setpoint values (T _F1,set , T _F2,set ) of the zone temperatures are specified, - the values of the manipulated variables (n, i ₁ , d ₂ ) of the actuators (4-6) are calculated on the basis of a circuit model that includes a complete control loop for adjusting the actual values (T _F1,calc , T _F2,calc ) of the zone temperatures to the target values of the zone temperatures ( T _F1,soll , T _F2,soll ) describes, in which - a controlled system of the control loop of the cycle model includes a dynamic partial model (MT _dyn ) of at least one temperature zone, which consists of values (Q̇ ₁ ,Q̇ ₂ ) of cooling capacities of the respective temperature zones Actual values (T _F1,calc , T _F2,calc ) of the zone temperatures are calculated, - the dynamic partial model (MT _dyn ) is preceded by a stationary partial model (MK _stat ) of the controlled system, which consists of the values of the manipulated variables (n, d ₁ , d ₂ ) of the actuators (4-6) calculates the respective values (Q̇ ₁ ,Q̇ ₂ ) of the cooling capacities and transfers them to the dynamic partial model (MT _dyn ) as well as actual values (P _e,calcl , T _di ) of at least one other state variable of the controlled system, - the stationary partial model (MK _stat ) is preceded by an inverse stationary partial model (MKinv _stat ), which is based on specified setpoint values (Q̇ _1,setpoint ,Q̇ _2,setpoint ) of the cooling capacities of the respective temperature zones and default values (T _di,ref ) which calculates the values of the manipulated variables (n, d ₁ , d ₂ ) of the actuators (4-6) at least one further state variable and transfers them to the stationary partial model (MK _stat ), - the inverse stationary partial model (MKinv _stat ) A controller model (TR ₁ , TR ₂ ) is connected upstream for each temperature zone, which is calculated from the differences between the specified setpoint values (T _F1,setpoint , T _F2,setpoint ) of the associated zone temperature and the measured actual values (T _F1,meas , T _F2,meas ) of the associated zone temperature calculates the respective setpoint values (Q̇ _1,setpoint ,Q̇ _2,setpoint ) of the cooling capacities and transfers them to the inverse stationary partial model (MKinv _stat ), and - an optimizer (OPT) from measured actual values ( P _el,mess ) of a further state variable determines a default value (T _di,ref ) of another further state variable (T _di ), transfers this default value (T _di,ref ) to the inverse stationary partial model (MKinv _stat ) and the default values (T _{di ,ref} ) iteratively to a predetermined target state of the input variable (P _el,soll ) out. procedure after Claim 13 , where the optimizer (OPT) is constantly working.

Images