DE102017206299A1 - Method for controlling an operating device of a building and building automation system - Google Patents

Abstract

The invention relates to a method (14) for controlling (S7) at least one operating device (7) of a building (1) and to a corresponding building automation system (2). In the method (14), an optimal solution of a model of a predictive control problem describing at least one operating process of the operating device (7) is calculated by means of a computing device (5) as a function of predetermined parameters (S2). At least a portion of the predetermined parameters and solution data describing the optimal solution are transmitted from the computing device (5) to a local control unit (13) of the building (S3). Furthermore, at least one current measured value of at least one of the parameters is detected (S4) whose value used in calculating (S2) of the optimal solution deviates from the current measured value. Subsequently, an approximate solution of the model-predictive control problem is calculated by means of the control unit (13) on the basis of the transmitted solution data and the current measured value (S6). Finally, the operating device (7) is controlled according to the approximate solution (S7).

Description

The invention relates to a method for controlling at least one operating device of a building and a building automation system.

Buildings, especially large commercial buildings, today often have a building management or building automation system. Such a system may include, for example, an infrastructure underlying an operation of the building, for example for the management, distribution and control of energy. A structure or operation of a building automation system can be described by multiple hierarchical levels. For example, higher-level functions, such as a visualization of the building infrastructure, can be summarized and perceived at a management level. In an automation level arranged hierarchically thereunder, individual field devices, which can be elements of the infrastructure arranged, for example, distributed in the building, can be brought together and, for example, grouped according to their function and / or their position. Also, such device groups can be controlled in or on the automation level. Below this hierarchically, a field level can be provided, on which the individual components or elements of the building infrastructure are connected to the building automation system.

A subsystem of such a building automation system may be a Building Energy Management System (BEMS). The core functions of the energy management system may include, for example, energy efficient control of the connected infrastructure components, protection of the infrastructure from congestion, and provision of comfort or convenience features, such as a particular indoor temperature in individual areas of the building.

In particular against the background of sustainability certifications and a targeted reduction of emissions, the energy management system should in future also be responsible for coordinating the energy consumption of the building or the building infrastructure and the building's own energy generation. This may include, for example, management of renewable energy sources, such as photovoltaic systems, electrical storage, and controllable loads, such as electric vehicle charging stations. For this purpose, it may be provided that the energy management system also performs electrical measurements, infrastructure monitoring, data analysis and forecasting.

In such an environment that may arise in the future, the tasks or requirements for energy management systems can be formulated as new or supplementary as follows. First: Prediction of the maximum possible self-generated energy of the building and self-consumption of non-controllable loads (base load). Second: Optimized energy purchasing through cost-optimal planning of an operation of controllable loads, in particular electrical and thermal, storage, energy production and energy supply from a supply network. Third, online energy management to maintain a nominal load profile, especially despite disruptions or unforeseen events. Fourth: Automated participation in an energy market by marketing available flexibility in energy procurement or in energy production. Fifth: Optimization of purchasing different forms of energy, such as electricity, gas, district heating, district cooling and the like.

In order to be able to fulfill these tasks or functionalities, the energy management system requires extensive data, for example technical specifications of the individual infrastructure elements and weather forecasts, as well as powerful calculation hardware, since modeling and optimization calculations to be solved are very expensive.

The object of the present invention is to enable a robust control of a building infrastructure with reduced hardware complexity.

This object is achieved by the subject matters of the independent claims. Advantageous embodiments and further developments of the invention are specified in the dependent claims as well as in the description and in the drawings.

In the method according to the invention for controlling at least one operating device of a building or a construction, an optimal solution of a model-predictive control problem or optimization problem describing at least one operating process of the operating device is calculated by means of a computing device (MPC, English: "Model Predictive Control"). At least a portion of the predetermined parameters and at least a portion of solution data describing the optimal solution is transmitted by the computing device to a local control unit of the building, such as a local controller of the facility. This local control unit can in particular - for example, due to their design and / or Computing power - unsuitable for calculating the optimal solution. This can be the case, for example, if the local control unit is a simple controller which, for example, has a predetermined range of functions, which is specialized, for example, for controlling the operating device. Therefore, software programs used to calculate the optimal solution can not be executed on the controller or the control unit. A further method step of the method according to the invention consists in detecting at least one current measured value of at least one of the parameters whose value used in calculating the optimum solution deviates from the current measured value. Such a deviation may arise, for example, if respective actual weather conditions deviate from an assumed weather forecast and thus, for example, provides a photovoltaic system serving as the operating device more or less electrical energy than predicted. On the basis of the transmitted solution data and the current measured value, an approximate solution of the model-predictive control problem is then calculated by means of the local control unit. The at least one operating device is then controlled according to the approximate solution. For this purpose, for example by the control unit corresponding control signals or commands can be generated and transmitted to the operating device.

The operating device may be a, in particular controllable, element or component of an infrastructure, in particular an energy infrastructure, of the building. For example, the operating device may be an energetic load, a power generator, an energy store or a switching element or the like. In concrete terms, the operating device can therefore be, for example, an air-conditioning device or installation, a photovoltaic or wind power plant, a cogeneration plant of the building, a battery or a rechargeable battery, a thermal storage or the like. Several or a plurality of operating devices may be provided in or on the building, ie as part of the building infrastructure, which may be directly or indirectly connected to one another. The method according to the invention can then be provided or used for controlling also several or all of the operating devices. In this case, the plurality of operating devices can be controlled by means of a single or multiple control units. If a plurality of control units are provided, they may be assigned to an operating device or a group or type of operating devices for their control, for example. The control unit or the control units are like the operating facilities also part of the building infrastructure and can thus be directly or indirectly connected or networked with each other. Part of this infrastructure or this network or network can also be at least one, preferably a plurality, of measuring devices for detecting the at least one current measured value. By way of example, these measuring devices may be or include corresponding sensors or sensor systems, which may differ in their type or design depending on the measured value or parameters to be determined.

The computing device for calculating the optimal solution can in principle also be part of the building or part of the infrastructure. Preferably, however, the computing device may be a remote, that is to say separate from the building, computing device, for example an external data center or a cloud server. In this case, the computing device can be connected via a data connection with the building infrastructure, in particular with the control unit at least temporarily, in order to enable the transmission of data. By using such a remote computing device can be minimized in conjunction with the inventive method advantageously a respective hardware, operating and cost of operation of the individual building. Advantageously, the computing device, for example, by the calculation of the optimal solution only temporarily claimed or busy so that it is also available for other tasks or purposes, so for example, each individual optimal solutions for a variety of different buildings can calculate. This advantageously reduces unnecessary redundancy and improves efficiency.

The optimal solution uses as input, that is, as input variables or as input data, all necessary and available information, for example, the technical characteristics of the infrastructure or the operating equipment, a load profile or forecasts of a load profile of the building, to environmental conditions, such as building-specific power generation influencing weather conditions, and a more comprehensive system or network in which the building can be integrated into, by the appropriate control of infrastructure or facilities to be reached or realized objectives and the like. Possible objectives may be, for example, a minimization of emissions generated for the operation of the building, a certain occupancy or utilization of the infrastructure or individual operating facilities, a minimization of a cost to operate the building and / or the like more. Especially in the case of complex buildings or building infrastructures, in particular with multiple control units, it is advantageous if these extensive and complex data to be used for the calculation of the optimal solution are bundled only to the computing device, in particular thus not for each individual control unit, or distributed to different control units have to.

Since for the optimal solution, therefore, this extensive data and with the Recheneirichtung a sufficiently powerful hardware are used, it can be specified by them a best possible control strategy for the operating equipment or the operating equipment under the assumed conditions. In particular, since the predictions used are subject to uncertainty and may change after completion of calculating the optimal solution, which may take a significant amount of time, and in particular deviate from the corresponding prediction, a recalculation or update would be for optimal control the optimal solution necessary. However, a time delay is unavoidable due to the calculation time of the optimal solution. In addition, there can often be the problem that the computing device or the connection to the computing device is not permanently and / or not stably available. For example, availability may be given only once a day for a limited period of time and / or infrequently. A hitherto often used individual, simple control for each operating device can fulfill neither the initially outlined tasks or requirements of the energy management system nor one - even only approximately - ensure optimal realization of the specified objective. For example, in the case of an occurring disturbance, that is to say a deviation between a current measured value-for example a power generation by the photovoltaic system-from the corresponding prognosis, a simple control can completely or completely eliminate this disturbance or deviation via a public power supply network to which the building is connected via a battery, so an energy storage, intercept or compensate the building, as long as appropriate capacity or flexibility is available. However, this is not expedient in terms of the desired functionality of the energy management system, since usually a combination of several compensation mechanisms allows an improved fulfillment of the objective.

By calculating the approximate solution on the basis of the optimal solution can be compared to the simple control always at least equivalent in terms of the objective, but on average significantly better, that is closer to the optimal solution, control of the facility or the building infrastructure. Since the solution data already calculated in advance of the optimal solution are used for the calculation of the approximate solution, advantageously a complete recalculation or solution of the model-predictive control problem for calculating the approximate solution is not necessary. As a result, it is advantageously possible to realize much simpler and less extensive calculations relative to the calculation of the optimal solution, compared to conventional methods, for example more efficient, emission-poorer and / or more cost-effective, control of the operating device or of the building infrastructure. In particular, these simple calculations can be carried out by means of the control unit which can be operated at low cost and operating costs since, for example, only basic standard operations such as simple additions and modifications or at most matrix-vector modifications may be required and no complex modeling or optimization software needs to be executed ,

Both the optimal and the approximate solution can deliver as output, ie as output data or result, in particular one or more setpoint values or states of the operating devices to be reached or set. The optimal solution solution data used to calculate the approximate solution may include, for example, a function or matrix that translates a respective input into that output. The complexity and the associated necessary calculation effort of the optimal solution or the model-predictive control problem can be recognized from the fact that, for example, between 10,000 and 20,000 setpoints and / or states must be calculated for a large commercial building, with the respective operating facilities or their operating processes influencing each other or can be dependent on each other.

A difference between the approximate solution and the optimal solution may depend on a considered time horizon. However, for example, with availability of the computing device and calculation of a new or updated optimal solution once a day, control of the operating devices according to the approximate solution may be closer to the optimal solution than a conventional simple control, so that overall Limitations of the control unit with respect to the available data, the available computing capacity or computing power and the available range of functions - can be achieved by the inventive method, a significant improvement in the control of the operating device or more operating equipment of the building.

In an advantageous embodiment of the present invention, the predetermined parameters include a state of the operating device, a goal to be realized by the optimal solution and a prognosis for a load profile and / or an energy generation of the building. The state may be, for example, a state of charge (SOC) or an actual energy consumption or energy output of the operating device indicated by a corresponding measured value. These parameters can already be used to calculate a significantly improved optimal solution compared to conventional simple rules. The optimality of the solution can be based on the accuracy, timeliness, availability and / or the extent of the data and / or specifications used as input, for example with respect to the target, and / or further restrictions or boundary conditions, for example due to technical circumstances and possibilities. For example, a load capacity of the operating device or the building infrastructure, a switching and / or setting speed of a state or the same may be limited or limited in a predetermined manner, so that thereby possible control processes and thus a range of values of the optimal solution may be limited.

In a further preferred refinement of the present invention, at least one desired value of the operating device used in calculating the optimal solution and at least one nominal value determined during calculation of the optimal solution are transmitted by the computing device to the control unit. This advantageously makes it possible to determine the deviation, that is to say an interference with the data or predictions used for the optimal solution, for example by a simple comparison of the transmitted state with the current measured value. Likewise, it may be advantageous, for example, to compare a desired value determined during the calculation of the approximate solution with the desired value of the optimal solution, since this allows, for example, a simple evaluation of the approximate solution.

Dabei enthält der Vektor $\overrightarrow{x}$

wenigstens einen beim Berechnen der optimalen Lösung ermittelten Sollwert, also beispielsweise einen idealerweise zu erreichenden oder einzustellenden Zustand, Energieverbrauch oder dergleichen. Der Vektor $\overrightarrow{b}$

enthält die als Input für die optimale Lösung verwendeten beziehungsweise vorgegebenen Parameter beziehungsweise Vorhersagen. $A_B^{-1}$

ist eine Matrix, welche im Folgenden auch als inverse Basismatrix bezeichnet wird und die Lösungsdaten der optimalen Lösung enthält. In der Matrix $A_B^{-1}$

können beispielsweise einzelne Spalten jeweils einer Betriebseinrichtung, also beispielsweise einem Gerät oder Aggregat der Gebäudeinfrastruktur, zugeordnet sein. Beispielsweise kann eine Spalte einem elektrischen Energiespeicher, also einer Batterie, des Gebäudes oder der Gebäudeinfrastruktur zugeordnet sein und dann gegebenenfalls zur Berechnung des Ladungszustandes beziehungsweise der Sollwerte für den Ladungszustand der Batterie beitragen. Durch die Zeilennummern der Matrix $A_B^{-1}$

beispielsweise zeitliche Verläufe der Sollwerte innerhalb jeweils einer Spalte beschreiben werden, sodass also die Einträge jeweils einer Spalte zu einem jeweiligen zeitlichen Verlauf eine entsprechenden Größe korrespondieren können. So kann beispielsweise ein Element einer Spalte mit einer bestimmten ersten Zeilennummer zur Berechnung eines für einen ersten Zeitpunkt vorgesehenen Sollwertes und ein anderes Element derselben Spalte mit einer anderen Zeilennummer zur Berechnung eines für einen zweiten Zeitpunkt vorgesehenen Sollwertes beitragen. Das Produkt der inversen Basismatrix $A_B^{-1}$

mit dem Vektor $\overrightarrow{b}$

kann als Basislösung bezeichnet werden und insgesamt unter anderem technische Randbedingungen der zu steuernden Betriebseinrichtungen umfassen oder berücksichtigen.In a further advantageous embodiment of the present invention, it is provided that the optimal solution can be displayed or is shown as $\overrightarrow{x}=A_B^{-1}\cdot \overrightarrow{b},$

The vector contains $\overrightarrow{x}$

at least one setpoint determined during calculation of the optimal solution, that is to say, for example, a state that can be ideally reached or set, energy consumption or the like. The vector $\overrightarrow{b}$

contains the parameters or predictions used or given as input for the optimal solution. $A_B^{-1}$

is a matrix, which is also referred to below as the inverse base matrix and contains the solution data of the optimal solution. In the matrix $A_B^{-1}$

For example, individual columns can each be assigned to an operating device, that is to say for example a device or aggregate of the building infrastructure. For example, a column can be assigned to an electrical energy store, that is to say a battery, of the building or of the building infrastructure and then optionally contribute to the calculation of the charge state or the setpoint values for the charge state of the battery. By the line numbers of the matrix $A_B^{-1}$

For example, temporal courses of the setpoint values within each one column will be described, so that therefore the entries in each case of a column can correspond to a respective time course of a corresponding size. For example, one element of a column with a particular first row number may contribute to the calculation of a setpoint provided for a first time and another element of the same column may have a different row number for the calculation of a setpoint provided for a second time. The product of the inverse base matrix $A_B^{-1}$

with the vector $\overrightarrow{b}$

may be referred to as a basic solution and include or take into account, inter alia, technical boundary conditions of the operating equipment to be controlled.

und als auch die Einträge der Matrix $A_B^{-1}$

können im einfachsten Fall Zahlen beziehungsweise Werte oder Variablen sein. Die jeweiligen Bedeutungen der einzelnen Einträge, das heißt also beispielsweise an welcher Stelle einer bestimmten Betriebseinrichtung oder einem bestimmten diskreten Zeitpunkt zugeordnete Einträge oder Angaben stehen, kann beispielsweise durch die Matrix $A_B^{-1}$

selbst kodiert sein oder durch entsprechende Metadaten oder in einer separaten Datei angegeben sein. Diese Bedeutungen oder Zuordnungsdaten können ganz oder teilweise zusammen mit den Lösungsdaten und/oder den Parametern von der Recheneinrichtung zu der Steuereinheit übertragen werden. Ebenso kann es vorgesehen sein, dass diese Zuordnungsdaten bereits in einer Speichereinrichtung der Steuereinheit gespeichert beziehungsweise hinterlegt sind.Both the entries of the vectors $\overrightarrow{x}\text{and}\overrightarrow{b}$

and the entries of the matrix $A_B^{-1}$

In the simplest case, they can be numbers or values or variables. The respective meanings of the individual entries, that is to say, for example, at which point of a particular operating device or a specific discrete point of time are associated entries or indications, can be determined, for example, by the matrix $A_B^{-1}$

self-coded or be specified by appropriate metadata or in a separate file. These meanings or assignment data can be transferred in whole or in part together with the solution data and / or the parameters from the computing device to the control unit. Likewise it can be provided that these assignment data are already stored or stored in a memory device of the control unit.

übertragen wird. Die approximative Lösung wird dann berechnet durch eine Matrix-Vektor-Multiplikation der inversen Basismatrix $A_B^{-1}$

mit einem Vektor $\overrightarrow{\tilde{b}},$

welcher gebildet ist aus dem gemäß dem aktuellen Messwert modifizierten Vektor $\overrightarrow{b}.$

Der aktuelle Messwert beschreibt also eine von den für die optimale Lösung angenommenen Bedingungen abweichende Situation. Der Vektor $\overrightarrow{b}$

kann beispielsweise durch Addition der Abweichung zu dem entsprechenden Eintrag des Vektors $\overrightarrow{b}$

an diese neue Situation angepasst, also zu dem Vektor $\overrightarrow{\tilde{b}},$

modifiziert werden. Bei diesem Vorgehen muss vorteilhaft also die inverse Basismatrix $A_B^{-1}$

nicht modifiziert oder neu berechnet werden, sodass das zu berechnende Matrix-Vektor-Produkt $A_B^{-1}\cdot \overrightarrow{\tilde{b}}$

problemlos durch die Steuereinheit berechnet werden kann. Als Ergebnis ergibt sich die approximative und nicht die optimale Lösung, da wechselseitige Einflüsse oder Rückkopplungen, die mit der Abweichung beziehungsweise der Modifikation des Vektors $\overrightarrow{b}$

einhergehen - beziehungsweise in der entsprechenden optimalen Lösung für die abweichende Situation einhergehen würden - und bei einer Berechnung einer neuen optimalen Lösung zu Veränderungen auch in anderen Einträgen führen können, unberücksichtigt bleiben. Da jedoch nur der Wert des durch den aktuellen Messwert beschriebenen Parameters modifiziert wird, ist auch die Differenz zwischen der dieserart berechneten approximativen Lösung und der optimalen Lösung beziehungsweise einer unter Berücksichtigung des aktuellen Messwertes neu berechneten optimalen Lösung begrenzt. Es kann also beispielsweise nur ein Wert in dem Vektor $\overrightarrow{b}$

geändert werden, um den modifizierten Vektor $\overrightarrow{\tilde{b}},$

zu erhalten. Basierend hierauf wird dann durch die Multiplikation mit der inversen Basismatrix $A_B^{-1}$

dennoch eine approximative Lösung für das Gesamtsystem, also beispielsweise die gesamte Gebäudeinfrastruktur oder mehrere Betriebseinrichtungen, berechnet beziehungsweise ermittelt.In a further advantageous embodiment of the present invention, it is provided that from the computing device to the control unit as solution data, the inverse base matrix $A_B^{-1}$

is transmitted. The approximate solution is then calculated by a matrix-vector multiplication of the inverse base matrix $A_B^{-1}$

with a vector $\overrightarrow{\tilde{b}}.$

which is formed from the vector modified according to the current measured value $\overrightarrow{b},$

The current measured value thus describes a situation deviating from the conditions assumed for the optimal solution. The vector $\overrightarrow{b}$

For example, by adding the deviation to the corresponding entry of the vector $\overrightarrow{b}$

adapted to this new situation, ie to the vector $\overrightarrow{\tilde{b}}.$

be modified. In this procedure, therefore, the inverse base matrix must be advantageous $A_B^{-1}$

can not be modified or recalculated so that the matrix-vector product to be calculated $A_B^{-1}\cdot \overrightarrow{\tilde{b}}$

can be easily calculated by the control unit. The result is the approximate and not the optimal solution, because of mutual influences or feedbacks associated with the deviation or modification of the vector $\overrightarrow{b}$

go hand in hand - or would be in the appropriate optimal solution for the deviant situation - and can lead to changes in other entries when calculating a new optimal solution. However, since only the value of the parameter described by the current measured value is modified, the difference between the approximate solution calculated this way and the optimal solution or a newly calculated optimal solution taking into account the current measured value is also limited. For example, it can only have one value in the vector $\overrightarrow{b}$

be changed to the modified vector $\overrightarrow{\tilde{b}}.$

to obtain. This is then based on multiplication by the inverse base matrix $A_B^{-1}$

Nevertheless, an approximate solution for the entire system, so for example, the entire building infrastructure or multiple operating equipment, calculated or determined.

das ursprüngliche modellprädiktiven Regelungsproblems (MPC-Problem) unter Verwendung der Recheneinrichtung optimal löst, wird durch die Multiplikation $A_B^{-1}\cdot \overrightarrow{\tilde{b}}$

ein durch die Abweichung des aktuellen Messwerts gegebenes modifiziertes MPC-Problem approximativ gelöst. Somit kann durch eine einfache Matrix-Vektor-Multiplikation eine Approximation einer optimalen Reaktion auf die Abweichung, also die Störung, berechnet und die Betriebseinrichtung beziehungsweise die Gebäudeinfrastruktur entsprechend gesteuert werden.Particularly advantageously, the approximate solution can also be calculated particularly quickly, in particular faster than an updated optimal solution, by means of this procedure, so that it is possible to react to the deviation determined on the basis of the current measured value very quickly by appropriate control measures. This is advantageous in particular regardless of the connection to the computing device or its availability at any time possible. So while the basic solution $A_B^{-1}\cdot \overrightarrow{b}$

optimally solves the original model predictive control problem (MPC problem) using the calculator is multiplied $A_B^{-1}\cdot \overrightarrow{\tilde{b}}$

an approximately solved by the deviation of the current measured value modified MPC problem solved. Thus, by means of a simple matrix-vector multiplication, an approximation of an optimal response to the deviation, that is to say the disturbance, can be calculated and the operating device or the building infrastructure can be correspondingly controlled.

übertragen. Das bedeutet mit anderen Worten, dass nicht die vollständige inverse Basismatrix $A_B^{-1}$

übertragen wird, sondern lediglich eine Teilmenge der Spalten der inversen Basismatrix $A_B^{-1}.$

Die übertragenen Spalten sind dabei diejenigen Spalten der inversen Basismatrix $A_B^{-1},$

welche Einträge enthalten, die bei der Berechnung des Matrix-Vektor-Produkts $A_B^{-1}\cdot \overrightarrow{\tilde{b}}$

mit einem durch den aktuellen Messwert beziehungsweise dessen Abweichung modifizierten Eintrag oder Element des Vektors $\overrightarrow{\tilde{b}}$

multipliziert werden. Hierdurch kann vorteilhaft eine von der Recheneinrichtung zu der Steuereinheit zu übertragende Datenmenge reduziert werden. Dadurch kann vorteilhaft eine Zeitersparnis, eine schnellere Reaktionsfähigkeit der Steuereinheit und eine geringere Belastung einer verwendeten Datenverbindung erreicht werden. Mit diesen übermittelten Spalten und dem aktuellen Messwert beziehungsweise der Abweichung des aktuellen Messwerts von dem entsprechenden für die optimale Lösung angenommenen Wert kann durch die Steuereinheit so beispielsweise durch Addition der um die Abweichung beziehungsweise Störung modifizierten Spalten eine Approximation der optimalen Reaktion auf die Störung berechnet und die Infrastruktur entsprechend gesteuert werden.In a further advantageous embodiment of the present invention are of the Computing device to the control unit as solution data only to the corresponding parameters described by the current parameter columns of the inverse base matrix $A_B^{-1}$

transfer. In other words, that means not the complete inverse base matrix $A_B^{-1}$

but only a subset of the columns of the inverse base matrix $A_B^{-1},$

The transferred columns are those columns of the inverse base matrix $A_B^{-1}.$

which contain entries in the calculation of the matrix vector product $A_B^{-1}\cdot \overrightarrow{\tilde{b}}$

with an entry or element of the vector modified by the current measured value or its deviation $\overrightarrow{\tilde{b}}$

be multiplied. As a result, an amount of data to be transmitted from the computing device to the control unit can advantageously be reduced. As a result, a time saving, a faster response capability of the control unit and a lower load of a data connection used can advantageously be achieved. With these transmitted columns and the current measured value or the deviation of the current measured value from the corresponding value assumed for the optimum solution, the control unit can calculate an approximation of the optimal response to the disturbance and, for example, add the columns modified by the deviation or disturbance Infrastructure are controlled accordingly.

übertragen, die aufgrund ihrer jeweiligen Zeilennummer zu einem durch das Steuern zu erreichenden Sollwert und durch ihre jeweilige Spaltennummer zu dem durch den aktuellen Messwert beschriebenen Parameter, insbesondere also zu der Störung oder Abweichung, korrespondieren. Mit anderen Worten wird also nicht notwendigerweise die gesamte inverse Basismatrix $A_B^{-1}$

übertragen, sondern lediglich ein einzelner Eintrag oder eine Teilmenge von einzelnen Einträgen. Der zu erreichenden Sollwert kann dabei der zuvor genannte durch das Berechnen der optimalen Lösung ermittelte Sollwert sein. Mit den so übertragenen Einträgen der inversen Basismatrix $A_B^{-1}$

und dem aktuellen Messwert beziehungsweise dessen Abweichung gegenüber dem für die optimale Lösung angenommenen entsprechenden Wert kann dann durch die Steuereinheit durch jeweils eine Multiplikation pro entsprechendem Sollwert eine Approximation der optimalen Reaktion auf die Abweichung beziehungsweise Störung berechnet und die Betriebseinrichtung entsprechend, das heißt gemäß der derart ermittelten approximativen Lösung, gesteuert werden.In a further advantageous embodiment of the present invention are from the computing device to the control unit as solution data only such entries of the inverse base matrix $A_B^{-1}$

transmit, due to their respective line number to a setpoint to be achieved by the control and by their respective column number to the parameter described by the current measured value, in particular to the disturbance or deviation corresponding. In other words, it does not necessarily mean the entire inverse base matrix $A_B^{-1}$

but only a single entry or a subset of individual entries. The target value to be achieved can be the aforementioned desired value determined by calculating the optimum solution. With the so transferred entries of the inverse base matrix $A_B^{-1}$

and the current measured value or its deviation from the corresponding value assumed for the optimal solution can then be calculated by the control unit by one multiplication per corresponding setpoint an approximation of the optimal response to the deviation or disturbance and the operating device accordingly, that is according to the thus determined approximate solution, to be controlled.

durch die Störung beeinflusst oder verändert werden beziehungsweise zur Berücksichtigung der Störung relevant sind. Dabei kann beispielsweise ein bestimmter Zeitpunkt oder Zeithorizont vorgegeben sein, für den die approximative Lösung beziehungsweise entsprechende Sollwerte approximativ berechnet werden sollen.By virtue of this procedure, it is advantageously possible to reduce the amount of data to be transmitted from the computing device to the control unit and, if necessary, a calculation effort for calculating the approximate solution. Advantageously so only those entries are transmitted and those individual calculations performed, which in the calculation of the product $A_B^{-1}\cdot \overrightarrow{\tilde{b}}$

are influenced or changed by the disturbance or are relevant for the consideration of the disturbance. In this case, for example, a specific point in time or time horizon can be predetermined, for which the approximate solution or corresponding setpoint values are to be calculated approximately.

zu übertragen, da beispielsweise zum Zeitpunkt der Berechnung der approximativen Lösung bereits in der Vergangenheit zu erreichenden beziehungsweise für einen in der Vergangenheit liegenden Zeitpunkt vorgesehene Sollwerte beziehungsweise die entsprechenden korrespondierenden Matrixeinträge nicht benötigt werden, da sie für die zukünftige Steuerung nicht länger relevant sind. Ebenso kann beispielsweise eine im Rahmen der optimalen Lösung, etwa für Planungszwecke, berechnete Steuerungsstrategie für weiter als der vorgegebene Zeitpunkt oder Zeithorizont in der Zukunft liegende Zeitpunkte oder Zeiträume für die Berechnung der approximativen Lösung ignoriert werden. Wird ein solcher in der Zukunft liegender Zeitpunkt oder Zeitraum erreicht, so kann dann beispielsweise vorteilhaft mit einem dann neu erfassten neuen aktuellen Messwert eine neue approximative Lösung berechnet werden. Ebenso kann es beispielsweise unnötig sein, konstante Einträge beziehungsweise Spalten, welche nicht von dem aktuellen Messwert abhängen oder zu diesem korrespondieren, zum Berechnen der approximativen Lösung zu übertragen. Beispielsweise kann eine Spalte der inversen Basismatrix $A_B^{-1}$

eine maximale Batteriekapazität enthalten, welche - zumindest über den vorgegebenen Zeithorizont - konstant, also unverändert bleibt, da sie beispielsweise nicht von dem aktuellen Ladezustand der Batterie abhängt.In other words, it is not necessary, for example, the entire inverse base matrix $A_B^{-1}$

since, for example, at the time of the calculation of the approximate solution, target values to be reached or intended for a past time point are not required, since they are no longer relevant for future control. Likewise, for example, a control strategy calculated within the framework of the optimal solution, for example for planning purposes, can be ignored for time points or time periods for the calculation of the approximate solution that lie ahead of the predefined time or time horizon in the future. If such a time or period lying in the future is reached, it is then advantageously possible, for example, to calculate a new approximate solution with a newly acquired, new measured value. Likewise, it may be unnecessary, for example, to transmit constant entries or columns, which do not depend on or correspond to the current measured value, for calculating the approximate solution. For example, a column may be the inverse base matrix $A_B^{-1}$

contain a maximum battery capacity, which - at least over the predetermined time horizon - constant, ie unchanged, since it does not depend, for example, on the current state of charge of the battery.

In a further advantageous embodiment of the present invention, a respective optimal solution for several scenarios, each with different forecasts, for example, for an environment, or weather conditions, for a load profile and / or for an energy production of the operating device and / or the building, first calculated by the computing device , In other words, several optimal solutions are calculated, which are optimal for different possible conditions or situations or developments. This takes into account the inevitable uncertainty of forecasts. From the computing device to the control unit, corresponding parameters and solution data are then transmitted to a plurality of the scenarios. Based on this transmitted data is then determined by the control unit by comparing it with the current measured value of this closest scenario. The approximate solution is then calculated on the basis of this determined closest scenario. In this case, the closest scenario may be that of the several calculated scenarios in which a value used for the parameter described by the current measured value has the smallest difference to the current measured value. If there are several different current measured values, for example different parameters, then, for example, the scenario can be selected which has the overall lowest total deviation or the lowest average deviation over all of these several measured values or parameters. In this case, additional conditions, such as a maximum deviation can be evaluated or taken into account when determining the closest scenario.

oder deren relevante Spalten oder Einträge sowie die jeweiligen ermittelten Sollwerte und Zustände und die oder einige der verwendeten Annahmen beziehungsweise Eingangsgrößen an die Steuereinheit übertragen.The computational effort required to determine the closest scenario is negligible and, for example, by default Nowadays available controllers, which can serve as a control unit in the context of the present invention, manageable, since this example, only comparison operations may be necessary. The multiple calculated scenarios may describe or model several different developments so that the determination or selection minimizes a difference between the approximate solution and the optimal solution. As a result, in particular, without any additional expenditure on hardware in the control unit, the predetermined target to be achieved by controlling the operating device or the building infrastructure can be realized better, ie with a smaller difference to the respective optimum. Depending on the given target, for example, a higher energy efficiency, a lower cost burden, an improved environmental balance or the like can be achieved. Specifically, in this procedure, for example, for each of the scenarios calculated by the computing device, the corresponding base matrices are used $A_B^{-1}$

or the relevant columns or entries as well as the respective determined setpoint values and states and the or some of the assumptions or input variables used are transmitted to the control unit.

In a further advantageous embodiment of the present invention, at least two respectively current measured values, in particular of the same parameter, are detected at a time interval from one another, which respectively lie closest to different scenarios. These different scenarios are then combined by the control unit into a combined scenario and the approximate solution is calculated by the control unit based on the combined scenario. Combining the scenarios here can mean, for example, that the disruption and the associated states, assumed values or forecasts of the scenarios are reconciled with one another. For example, by an averaging or a steady continuation.

For example, by means of the computing device, two different optimal solutions can be calculated, for the optimal solution, an energy production of the photovoltaic system of the building based on a cloud-free sky throughout the day and for the second optimal solution deviating energy production of the photovoltaic system based on the entire Day is used over cloudy sky. Deviating from both scenarios, in reality, for example, the sky may be cloudless in the morning of the corresponding day and cloudy in the afternoon of the corresponding day. A current measured value recorded during the morning would therefore lead to the scenario based on the cloudless sky being determined as the closest scenario. Similarly, a second current reading taken in the afternoon would result in determining the second scenario as closest. Thus, over the entire day, neither of the two calculated scenarios is optimally suited for mapping reality. With the acquisition of the second measured value, it would therefore come to a scenario change. However, since the second scenario does not take into account the increased power generation of the photovoltaic system during the morning due to the sky devoid of the assumption of the second scenario, the control will deviate from an optimal solution based on the actual reality but not previously calculated. In order to obtain an improved approximate solution even in such cases, the two or more different scenarios, which are determined as being closest at different points in time, can be reconciled with one another. For this purpose, the respectively closest scenario can be determined, for example by comparing the disturbance and the associated states, for example an associated state of charge of an energy store of the building infrastructure, and by means of the inverse base matrix of the scenarios being reconciled by the control unit or aligned with one another. For this purpose, already simple mathematical operations can advantageously be sufficient, which in their complexity or their claim do not go beyond matrix-vector multiplications and are accordingly executable by the control unit. In other words, the mechanism described above in connection with other embodiments of the invention can be applied not only to the set values, but also to the states or the input. Thus, in spite of the use of an approximate solution in more different situations, the respectively predetermined target can advantageously be improved, that is to say with less deviation, realized or achieved.

In addition to the method according to the invention, another aspect of the present invention relates to a building automation system. In each case, the building automation system according to the invention comprises at least one controllable operating device, a local control unit and a measuring device. In this case, the control unit is set up to calculate an approximate solution of a model-predictive control problem describing at least one operating process of the operating device. This approximate solution is calculated as a function of provided solution data describing an optimal solution of the model predictive control problem Depending on provided parameters used for the optimal solution and depending on a detected by the measuring device current measurement at least one of these parameters, the value used for the optimal solution deviates from the current measurement. The control unit is additionally configured to control the at least one operating device in accordance with the approximate solution calculated in this way. The solution data and optionally the parameters used can be provided, in particular, by a computing device which may be part of or separate from the building automation system. Likewise, the control unit may alternatively or additionally be arranged for controlling another operating device of the building automation system or of the building.

The embodiments described above and below and the respective advantages of the method according to the invention and of the building automation system according to the invention are each mutually mutually transferable or exchangeable between the method and the building automation system. This applies in particular also to components or devices used or to be used for carrying out the method according to the invention. Thus, the invention also includes developments of all aspects of the invention having features and advantages as described in connection with only one or some of the aspects. For this reason, not all corresponding developments of all aspects of the present invention are explicitly described in all combinations.

Further advantages, features and details of the invention will become apparent from the following description of preferred embodiments of the invention and from the drawings. The features and combinations of features mentioned above in the description as well as the features and feature combinations mentioned below in the description of the figures and / or shown alone in the figures can be used not only in the respectively indicated combination but also in other combinations or in isolation, without the scope of To leave invention.

• 1 eine schematische Darstellung eines Gebäudes mit einem Gebäudeautomationssystem sowie einer separaten Recheneinrichtung; und
• 2 einen schematischen Ablaufplan eines Verfahrens zum Steuern mindestens einer Betriebseinrichtung eines Gebäudes.

Embodiments of the invention are explained below with reference to schematic drawings. Show it:

• 1 a schematic representation of a building with a building automation system and a separate computing device; and
• 2 a schematic flowchart of a method for controlling at least one operating device of a building.

The exemplary embodiments explained below are preferred embodiments of the invention. In the exemplary embodiments, the described components of the embodiments each represent individual features of the invention, which are to be considered independently of one another, which each further develop the invention independently of one another and thus also individually or in a different combination than the one shown as part of the invention. Furthermore, the described embodiments can also be supplemented by further features of the invention already described.

1 shows a schematic representation of a building 1 with a building's own building automation system 2 , The building 1 is via a network connection 3 to an external supply network 4 , for example to a public electricity network. Also shown is one of the building 1 separate server device, which in the present case is a cloud server 5 is. The building 1 or the building automation system 2 and the cloud server 5 are at least temporarily via a wired or wireless data connection 6 connected with each other.

The building automation system 2 in the present case comprises a plurality of operating devices 7 which also forms part of the building 1 are. To these operating facilities 7 include, in particular electrical, consumer 8th , which is a basic load of the building 1 represents a charging station 9 for electric vehicles, a photovoltaic system 10 and a battery 11 , These operating facilities 7 Here are an example of a line system 12 connected and so ultimately connected. The pipe system 12 can serve both for the transmission or transmission of energy as well as for the transmission of data. Such over the pipe system 12 transmitted data may include, for example, control signals, status messages, measured values or the like.

The charging station 9 may exemplify a controllable load, the photovoltaic system 10 can exemplify a building's own power generation and the battery 11 may exemplify a charge and discharge energy storage.

In the present case, each of the operating equipment points 7 a respective control unit 13 This is, for example, a simple controller for controlling a respective functionality or a respective operation or operating process of the respective operating device 7 can act. The control units 13 can do this especially against the cloud server 5 lower Computing capacities and / or capabilities, ie functional scopes have. So can the control units 13 For example, it may be configured to process a specialized controller code and may not be capable of executing, for example, any software programs written in a high-level programming language such as C or C ++. Each of the control units 13 can for controlling the respective operating device 7 be set up, at or in which it is arranged. Likewise, at least one of the control units 13 for controlling several of the operating devices 7 be furnished. Other than shown here, the building automation system 2 So for example only one control unit 13 or fewer control units 13 as operating equipment 7 exhibit.

The building automation system 2 , the operating equipment 7 , the pipe system 12 and other elements, such as measuring devices, not shown here, can together as the infrastructure of the building 1 be designated.

2 shows a schematic flowchart 14 a method for controlling the operating equipment 7 of the building 1 ,

The following will be 1 and 2 explained in more detail together.

In one process step S1 the procedure is started. For example, an installation of the building automation system 2 in the building 1 completed or completed and the building automation system 2 be put into operation. Likewise, for example, a set of parameters can be specified, which in the control of the operating equipment 7 should be taken into account. Such parameters may, for example, technical characteristics of the operating equipment 7 , respective measured values or states, a predicted load curve of the building automation system 2 or the building 1 , a weather forecast and / or the like more. Likewise, an objective can be specified with regard to which the operation of the building automation system 2 by appropriate control of the operating equipment 7 to be optimized.

In order to enable such optimized or optimal operation, therefore, a complex optimization problem, in particular a model-predictive control problem (MPC problem), is to be solved. Since the calculation of an optimal solution to this MPC problem is computation-intensive, the optimal solution becomes in one process step S2 taking into account the specified parameters by means of the cloud server 5 calculated. In the process, or before, by means of the cloud server 5 also calculate one or more predictions to be used in the calculation of the optimal solution. This can be, for example, a historical or previous load profile of the building 1 descriptive data are provided.

As a result of calculating the optimal solution of the MPC problem, setpoints for the individual operating devices result 7 and a solution or base matrix containing solution data describing or characterizing the optimal solution. The basic matrix allows input values, ie an input, to be converted into those for optimum operation of the building automation system 2 provided or to be reached or set target values. Due to their limited functionality and their limited computing capacity, the control units 13 do not solve this MPC problem and therefore do not calculate the optimal solution.

Likewise, by means of the cloud server 5 For example, several optimal solutions for different values of the input variables, for example for different weather developments, are calculated.

The problem is that future forecasts are used for the optimal solution, which are always subject to uncertainty. Another problem may be that the cloud server 5 and / or the data connection 6 are not permanently available for recalculations or updates of the optimal solution or for transferring data.

Nevertheless, a relatively close to the optimum described by the optimum solution operation of the building automation system 2 In this case, the operating equipment should be provided under these conditions 7 be controlled so that the optimal operation is at least approximated.

For this it is envisaged that in a procedural step S3 at least a part of the predetermined parameters and at least a part of the solution data describing the optimal solution or the optimal solutions, that is to say the respective base matrix or a part of the respective base matrix, from the cloud server 5 to the control units 13 be transmitted.

In one process step S4 is by means of a measuring device, not shown here, for example, a sensor, which is also part of the building automation system 2 may be a current reading of one of the parameters detected. For example, as the current measured value, a current power output of the photovoltaic system 10 measured, the current measured value of the assumed for the calculation of the optimal solution value for the power generation of the photovoltaic system 10 may deviate at the time of acquisition of the current measured value. For example, since such measured values can be recorded repeatedly continuously or at predetermined time intervals, the method step can be detected S4 for example, before or after or parallel to the method step S3 be performed.

In the simplest case, the control units could 13 now the operating equipment 7 according to the on the cloud server 5 control calculated optimal solution. However, it is not possible for the control units to respond to disturbances, as in the present example by the deviation between the current measured value and the value assumed in the meantime 13 be reacted. For example, is the measured energy production by the photovoltaic system 10 less than the assumed or predicted value, an optimal response to this disturbance could lie therein, the resulting lack of energy in part due to an increased energy supply from the supply network 4 and partly by a customized operation of the battery 11 compensate. For example, the battery could 11 initially charged or discharged for a certain period of time and recharged during a later period to compensate. In this case, optimal control can be achieved, for example, in the case of dynamic, for example energy prices varying over a course of the day, and if energy cost minimization is specified as the goal, a cost optimal control of the building automation system 2 allow or specify.

In one process step S5 If several different optimal solutions have been calculated for different situations or scenarios, this is done by comparing the current measured value with the transmitted data through at least one of the control units 13 Determines which of the calculated scenarios, that is, which of the calculated optimal solutions, comes closest to the actual situation described by the current measurement. This scenario is then selected for further use.

In one process step S6 is then through at least one of the control units 13 calculated on the basis of the transmitted solution data and the current measured value an approximate solution of the MPC problem. For this purpose, for example, the base matrix of the selected optimal solution can be applied to input data modified according to the current measured value, so that the resulting approximate solution or a set of corresponding approximate nominal values is significantly simpler than matrix-vector multiplication compared to the complete optimal solution of the MPC problem can be calculated.

In one process step S7 then some or all of the operating equipment 7 controlled according to the approximate solution. In this case, the optimum solution and the assumptions or input data used for this purpose are assumed and a deviation or disturbance from these assumptions is taken into account by a corresponding modification. Thus, in particular with slight deviations between the current measured value and the value assumed instead in the calculation of the optimal solution, an approximately optimal operation of the building automation system can be achieved 2 surrender or discontinue. In particular, this allows significantly better results in the sense of the respectively specified objective compared to a conventional simple control.

In a conventional simple control, for example, no interdependencies or influences between the individual operating equipment 7 and may also not take into account external factors, such as dynamic energy prices. Instead, in the conventional often used simple regulations, for example, each of the control units 13 the operating equipment assigned to it 7 independently, for example by means of a PI controller, according to a single criterion. In contrast, the approximate solution used herein continues to make the complex dependencies within the building automation system 2 as well as the external factors such as the dynamic energy prices and one for the entire building 1 or the entire building automation system 2 valid objective - at least approximately - taken into account.

At least some of the method steps described can be repeated in a loop or continuously or regularly, which here by a method step S8 is indicated. For example, the process step S2 be done once a day. In particular, the process steps S4 to S7 However, for example, they can also be repeated several times or multiple times during the day in order, for example, to respond to changes in one or more measured values or deviations from the values assumed for the optimum solution in a particularly timely manner on the basis of current measured values.

The method described may, in particular after at least one-time transmission from the cloud server 5 to at least one of the control units 13 takes place, for example, in the case of an abort or interruption of the communication or data connection 6 the building automation system 2 or the building infrastructure with building's own funds further in terms of the specified objective, for example, approximately cost optimal, controlled or operated. Thus, therefore, an improved robustness of the building automation system 2 achieved or a more robust operation of the building automation system 2 or the building 1 allows. For example, a very cost-effective energy management solution can be realized, which only requires a simple controller and an occasional connection to a cloud computing service.

Claims (10)

Method (14) for controlling at least one operating device (7) of a building (1) with the method steps Calculating (S2) an optimal solution of a model-predictive control problem describing at least one operating process of the operating device (7) as a function of predetermined parameters by means of a computing device (5), Transferring (S3) at least part of the predetermined parameters and at least part of solution data describing the optimal solution from the computing device (5) to a local control unit (13) of the building (1), Detecting at least one of the current measured values of at least one of the parameters whose value used in calculating (S2) of the optimal solution deviates from the current measured value, Calculating (S6) an approximate solution of the model-predictive control problem by means of the control unit (13) on the basis of the transmitted solution data and the current measured value, and - Controlling (S7) of the operating device (7) according to the approximate solution. Method (14) according to Claim 1 , characterized in that the predetermined parameters comprise a state of the operating device (7), a target to be realized by the optimal solution and a prognosis for a load profile and / or an energy generation of the building (1). Method (14) according to one of the preceding claims, characterized in that from the computing device (5) to the control unit (13) at least one state of the operating device (7) used in the calculation (S2) of the optimal solution and at least one corresponding thereto Calculate (S2) of the optimum solution determined setpoint. Method (14) according to one of the preceding claims, characterized in that the optimal solution can be represented as $$\overrightarrow{x}=A_B^{-1}\cdot \overrightarrow{b}$$

being the vector $\overrightarrow{x}$

contains at least one setpoint determined during the calculation (S2) of the optimal solution, the vector $\overrightarrow{b}$

contains the given parameters and $A_B^{-1}$

is a matrix called inverse base matrix containing the solution data. Method (14) according to Claim 4 , characterized in that - from the computing device (5) to the control unit (13) as solution data, the inverse base matrix $A_B^{-1}$

and the approximate solution is calculated by matrix-vector multiplication of the inverse base matrix $A_B^{-1}$

with a vector $\overrightarrow{\tilde{b}}.$

which is formed from the vector modified according to the current measured value $\overrightarrow{b},$

Method (14) according to Claim 4 , characterized in that from the computing device (5) to the (13) control unit as solution data only columns of the inverse base matrix corresponding to the parameter described by the current measured value $A_B^{-1}$

be transmitted. Method (14) according to Claim 4 , characterized in that from the computing device (5) to the control unit (13) as solution data only those entries of the inverse base matrix $A_B^{-1}$

are transmitted, which correspond due to their line number to a setpoint to be achieved by the control and by their column number to the parameter described by the current measured value. Method (13) according to one of the preceding claims, characterized in that - by the computing device (5) a respective optimal solution for several scenarios, each with different forecasts, in particular for a load profile and / or for power generation of the operating device and / or the building (1), - from the computing device (5) to the control unit (13) to a plurality of the scenarios corresponding parameters and solution data are transmitted, and - determined by the control unit (13) by comparison with the current measured value, the closest this scenario and the approximate solution is calculated on the basis of this scenario. Method (14) according to Claim 8 , characterized in that - at least two respectively current measured values are detected at a time interval from each other, each of which is closest to different scenarios, - these scenarios are combined by the control unit (13) into a combined scenario, and - the approximate solution by the Control unit (13) is calculated based on the combined scenario. Building automation system (2) each having at least one controllable operating device (7), a local control unit (13) and a measuring device, wherein the control unit (13) is set up for this purpose to calculate an approximate solution of a model-predictive control problem describing at least one operating process of the operating device as a function of provided solution data describing an optimal solution of the model-predictive control problem, provided parameters used for the optimal solution and a current measured value of one of the parameters acquired by the measuring device, whose value used for the optimal solution deviates from the current measured value, and - To control the at least one operating device (7) according to the approximate solution (S7).

Images