MXPA06004679A - Method for the operation of a technical system - Google Patents

Abstract

The invention relates to a method for the operation of a technical system wherein working parameters are detected during a certain period of time. An operational mode and/or functional mode of the technical system are determined by the temporal behaviour of said operational parameters, using artificial intelligence methods.

Description

METHOD FOR THE OPERATION OF A TECHNICAL INSTALLATION FIELD OF THE INVENTION The invention relates to a method for the operation of a technical system, especially a power plant system. BACKGROUND OF THE INVENTION Modern industrial facilities generally show a multiplicity of installation parts, which interact with each other in a complex manner. In order for an installation to operate, the detection of the operating parameters of the important parts of the installation and the conduction to an automation and / or process control system are similarly understood. These operating parameters can be treated, for example, of input parameters, which are set by the user, to operate a part of the installation in a desired manner. For example, in a turbine, the fuel and air supply must be established in a combustion chamber to extract the desired power from a gas turbine. This power is also an operating parameter of the gas turbine, which can be included as an operating parameter. With the gas turbine, a generator as well as multiple auxiliary facilities are joined or linked. Each portion of the installation then shows multiple operating parameters, which are determined by an operator of the installation or which are produced or inferred as output parameters such as the incorporation of the type given below. Without it being evident, only from the registration of the operation parameter can only be developed by deduction the limiting measures and the measures for the operation of the technical installation. This is highly possible in partial regions, for example in the disconnection of a part of the installation, in case of the current value of an operation parameter of a limit value. An important difficulty is that, in the filling of the data, the set of operating parameters must be recognized, so that the operation of the installation can be positively influenced in total. A solution conjecture from the state of the art consists of simulating the technical installation by means of a model, to find from there which modifications of the operation parameters lead to which modifications of other operation parameters, to understand the exchange action between the parts of the installation or also within a part of the installation. This procedure is then very costly and tends to cause errors because the modeling of a complex technical installation is difficult and is only possible with limited accuracy. The invention is then based on the object of providing a method for operating a technical installation, by means of which the way of operation of a technical installation is determined in a simple manner. The objective is solved according to the invention by means of a method for the operation of a technical installation, wherein during an interval of time an important operation parameter is included which can be freely chosen from a portion or part of the installation and through the Temporal procedure of this operation parameter is determined by means of artificial intelligence methods, which comprises at least one group method (neural networks, fuzzy logic, combined neuro-diffuse method, genetic algorithm) of a form of operation and / or a way of operating the technical installation. The operating parameters then also comprise such capabilities, which for example are determined from state monitoring systems, such as for example a vibration analysis such as measured magnitudes or developed magnitudes and are set up for disposition. The invention is then separated from the consideration, that from a temporary procedure of the operating parameters, which are included and stored during a time interval, the conclusions of the current operation form of the technical installation are extracted, without for this a detailed knowledge of the dependencies of the parameters among themselves. Especially no model of the technical installation is presented so that these extractions can be made. The temporal behavior of the operation parameter can then be included, for example, in that for a current time point and for a later (or also historical) time point each one of a number is temporarily included in the operation parameters and they are combined so that each one of them combines or reunites in a moment / trace absorption, which they can compare. With the help of the methods of artificial intelligence it is possible, if - as provided by a method according to the invention - in addition during a period of time observation that they comprise the operation parameter produced and with that their temporary procedures, the consequences of the modifications of a number of operation parameters prove and qualify in the procedure of other operation parameters. When they are modified, for example during the modification of the determined operation parameter of the recording time interval (for example linear) and other operation parameters that also show a modification (for example quadratic) are determined, then this dependency is localized and quantified through the methods of artificial intelligence, without for example a comparison can be determined by a present model. The known methods of artificial intelligence can learn what is the connon between the parameters within a quantity of data of the operation parameters, in which they analyze their procedures or temporary behaviors. The verified or verified connons and their qualification can be improved, the larger the amount of data that is examined in the operating parameters. As long as a connon between the determined operation parameters is identified or quantified, the artificial intelligence methods are also in the position to determine also for such operation parameters and their modifications, for which no image is presented as a data set anymore. included in the operation parameters, with which the procedure or behavior of the same is calculated in dependence on other operating parameters. By means of the method according to the invention, it is therefore possible to determine in a simple manner the operation form and / or the operating mode of the technical installation, especially without being able to know a modeling of the technical function of the installation. The determination of the form of operation and / or operation then occurs through the analysis that is described of the procedure of the operation parameter and its dependencies among themselves. The operating parameters included during the time interval can be understood as the moment or stock record or also the characterization of the parts of the installation or installation ("Fingerprint" of the part of the installation or installation). A fingerprint then consists of a classic model, where according to the method according to the invention, the behavior of the operation parameter is concluded in the manner of operation and / or function of the technical installation by means of artificial intelligence methods . For this purpose, for example in a power installation, the traces for a stop and start as well as for normal operation can be incorporated to learn to identify each one of the forms or ways of operation. In a preferred embodiment of the invention, the operation parameters are included for at least two time slots that are temporarily one outside the other, which are then compared to each other as operation parameters that include data sets and by means of Artificial intelligence methods comprise at least one method from the group (neural networks, diffuse logic, neuro-diffuse combined methods, genetic algorithm) a forecast or forecast is found, how the operating parameters are adapted or adjusted to achieve a desired form of operation of the technical installation. In this embodiment, a comparison of at least two fingerprints is provided, where for example they are examined in comparison in the operation parameters that are modified more strongly. This comparison then helps to check which modifications of the operation parameters are necessary, in order to selectively influence other determined operation parameters. A power plant can be found for example throughout the day in a normal operation and suddenly decreases the given production. A comparison of the fingerprints of the technical history. of the installation, shows that it has been modified (for example, they show the operating parameters for an external atmospheric pressure with a significant decrease) and also, as it works contrary to them, so that at least the production is conserved (for example, the operating parameters also show a decrease for the atmospheric combustion pressure). A forecast is then established that by means of the selected operating parameters selectively placed, a desired form of operation of the power plant is determined. The forecast then preferably includes the task of modifying the operating parameters as well as their adjustment values as a data set, so that the desired form of operation is achieved. The comparison can then also include comparison of the fingerprints comparatively, both of the facilities different from each other, as well as the comparison of the footprints of the facilities equally similar to each other. Especially preferred is then further determined the provision of a confidence measure, which represents for this a probability that it leads to an adjustment of corresponding operation parameters of the forecast for the desired form of operation. A confidence measure of, for example, 100% means that this is calculated with greater certainty, which leads to an adjustment of the operating parameters according to the forecast for a desired form of operation of the technical installations. A high confidence measure of this type then originates, when the current desired forms of operation of the technical installation as well as possible limit conditions (for example, environmental factors) have already been made or are presented in the past, and also the adjusted values used then for the operating parameters are known as fingerprints. In this case, it is also possible to proceed from there with great security, that the technical installation is also currently in position, in order to obtain or achieve the desired form of operation. A confidence measure of, for example, 60% can then mean that compared to the current desired form of operation of the technical installation, none of these desired forms of operation corresponds exactly to the historical operation form as a fingerprint. Then there was a similar form of operation so that there is no inferred with great security, that the adjustment values given in the forecast for the operating parameters reach the desired form of operation, but that there is always a good opportunity for that . A confidence measure close to 0% can also specify, for example, that a desired comparable form of operation of the technical installation has not been considered approximate, and then the adjusted values established in the forecast for the operating parameters with greater uncertainty to the achievement of the desired form of operation with errors. Advantageously the operation forms of the technical installation is established by means of a correlation analysis of the operating parameters, where the modification actions of the operation parameters are established, which corresponding input parameters, in the operation parameters, which correspond to the output parameters. In this embodiment, the selective consequences of a modification of the input parameters in the output parameters dependent on them are detected and qualified. The input parameters are then similarly operating parameters, whose values can either be set by a user of the technical installation or are set through the limit conditions for example of the influence of the environment. The output parameters are such operation parameters, which are produced following an adjustment of the input parameters and that are consequently dependent on these; Correlation analyzes then examine the type of connection and quantify these.

Ideally, all the operating parameters of all the essential parts of the installation are included in a technical installation, so that by means of a method according to the invention, the operation form of the technical installation can be determined and regulated. total in a simple way; The methods according to the invention can then form a regulation system, by means of which one or multiple parts of the installation as well as the total technical installation are regulated by means of a closed control circuit. In the methods according to the invention, an image of the data bank of the operating parameters is generated. This image allows to divert the operation of the technical installation in connection between the operation parameters and the. form of operation of the technical installation, the own knowledge with the data including the form of operation desired, selected and compared of the technical installation to be controlled. Preferably, multiple fingerprints are compared with each other to identify which knowledge of one form of operation can be transmitted to another form of operation. The corresponding results and forecasts can be stored easily as data sets and, if necessary, can be called each time.

BRIEF DESCRIPTION OF THE FIGURE The following is an embodiment of the invention. Figure 1 shows a processing system for carrying out the method according to the invention.

DETAILED DESCRIPTION OF THE INVENTION Figure 1 illustrates a processing system 1, comprising a processing unit 10 for carrying out the method according to the invention. The operating parameters of a technical installation are carried to the processing unit 10, which include the input parameters 5 as well as the output parameters 20. The time marker 25 serves for the choice of a time interval that is of interest, during which the operation parameters 5 must be included. The temporal procedure of the operation parameters 5, during the time interval is examined by means of a neuronal network 30 and / or a neuro-diffuse function unit 35 and / or one or several genetic algorithms and from there it detects and qualifies a connection between at least a part of the operation parameter 15 and at least a part of the output parameter 20. The recognition of this connection finally allows the preparation of a data set 50, which includes adjustment values for at least a part of the operation parameter 5, to achieve a desired form of operation of a part of the installation of a technical installation . This data set 50 represents a forecast, adjusted or determined as determined parameters, to realize the desired form of operation of the technical installation. In addition, a confidence measure, which represents a guarantee for the same, is issued through the processing unit 10, that an adjustment of the parameters is conducted according to the data of the data set 50 for the desired form of operation. Within the processing unit 10, a correlation analysis between the operation parameter 5 and the output parameter 20 takes place, so that the operation and operation form of the technical installation is possible according to the recognition of the procedure or temporary behavior of the input parameter 15 as well as the output parameter 20 which is in connection therewith and the data set 20 can be provided for the desired form of operation of the technical installation, for which no parameter would be included in the past. of operation with the corresponding input parameters 15 and the output parameters 20. The processing unit 10 is susceptible to interpolation.

Claims (4)

NOVELTY OF THE INVENTION
1. Having described the present is considered as a novelty, and therefore, the content of the following is claimed as property:
2. CLAIMS 1. Method for the operation of a technical installation, characterized in that during an interval of time an important operation parameter is included, which can be freely chosen, from at least a part of the installation and, based on the temporal behavior of this operation parameter, by means of artificial intelligence methods that include at least one method of the group (neural networks, fuzzy logic, combined neuronal / fuzzy logic method, genetic algorithm) a form of operation and / or a way of functioning is determined of the technical installation. Method according to claim 1, characterized in that the operating parameters are included during at least two temporary time intervals that are separated from each other, each time the operating parameters included or included as sets of operations are compared with each other. data, and by means of artificial intelligence methods that comprise at least one method of the groups (neural networks, fuzzy logic, combined neuronal / fuzzy logic, genetic algorithm) a forecast is established, of how they are established or adjusted at least a part of the operating parameters to achieve a desired form of operation of the technical installation.
3. 3. Method according to claim 2, characterized in that, in addition to the forecast, a confidence measure is determined, which represents a guarantee for the same, that an adjustment of the corresponding operation parameter leads to the forecast for the desired form of operation . Method according to one of claims 1 to 3, characterized in that the operation form of the technical installation is determined by means of a correlation analysis of the operation parameters, where the consequences and modifications of the parameters are determined of operation, which correspond to the input parameters, of the operating parameters, which correspond to the output parameters.