DE10330368B4 - Method and arrangement for identifying and / or differentiating substances displayed by sensors in gas mixtures and a corresponding computer program and a corresponding computer-readable storage medium - Google Patents

Abstract

Method for identifying substances in gas mixtures,
wherein parameters of the gas mixture are detected by a semiconductor or metal oxide sensor in the temperature change process, and
wherein at least a portion of the parameters detected by this sensor are evaluated by a computer program installed on a data processing device using reference and / or classification patterns, and substances identified by the sensor are identified, and
whereby an alarm decision is made depending on the results of the evaluation,
characterized in that
In a first step, a prealarm decision is made by means of a comparison of the values of the detected parameters with predetermined tolerance limits, and
- In response to the pre-alarm in a second step, the evaluation is carried out on the data processing device.

Description

The The invention relates to a method and an arrangement for identification and / or differentiation of substances displayed by sensors in gas mixtures as well as a corresponding computer program and a corresponding computer-readable storage medium, which in particular can be used for monitoring industrial plants, in particular the chemical industry, for large-area monitoring of welding control systems, as leak detection system for Refinery site, for the Intrusion protection of containers, as a monitoring system for emissions and immissions or even for the energy management of buildings and Attachments.

The Sensor development is becoming increasingly important in many areas to. This "stupid" sensors of a new generation of intelligent sensors. The development is through state-of-the-art technologies, high integration density and artificial Intelligence shaped.

• – optische Rauchmelder,
• – Ionisationsrauchmelder,
• – Flammenmelder,
• – Wärmemelder (IR),
• – chemische Gassensoren.

In the area of fire monitoring, the following types of sensors have been most frequently used:

• - optical smoke detectors,
• - ionisation smoke detector,
• - flame detector,
• - heat detector (IR),
• - chemical gas sensors.

When Particularly promising were the recently available semiconductor gas sensors considered. They are robust and long-term stable, are characterized by a high sensitivity (detection in the ppm range), readability and low noise. Furthermore They are offered at comparatively low prices. In spite of These advantages have not yet prevailed on the market. The reason for that This is due to the high sensitivity that causes the alarm to be triggered frequently with a high error rate. The problem of this highly sensitive Semiconductor gas sensors lies in the non-discrimination or inadequate Distinctness of the alarms.

Around to solve this problem, was proposed several, for various pollutants or sensitivities designed semiconductor gas sensors or a sensor with different metal oxide layers to provide. If one or more sensors give a signal, it will from the different, respectively appealing sensors on the kind of the triggering Event closed. Analog becomes from the respective appealing Closed metal oxide layers on the triggering gas mixture. A disadvantage of this method is in particular that in the first If more than one sensor must always be used. In the second case it is a disadvantage that the in a sensor, metal oxide layers used the amount of distinguishable Gas mixtures and thus the applications of such a sensor limit.

Semiconductor gas sensors currently z. As used in the automotive industry, the ventilation flaps in mid-range cars to control so as to ensure that no harmful gases in the passenger compartment penetrate. However, to date professional fire monitoring will be Semiconductor gas sensors hardly used, as a false alarm in this case would have very significant consequences.

In the field of the invention, solutions are already known in the prior art. This is how the German patent application describes DE 198 50 564 A1 a method and a device for fire detection, in which open and covert fires are detected by means of electrochemical gas sensors.

In the German Offenlegungsschrift DE 43 21 736 A1 a sensor device for oxidizable gases is described. Alarm states are detected when critical thresholds are exceeded.

The German disclosure documents DE 101 40 134 A1 and DE 37 39 406 A1 Respect each sensor arrangements with multiple sensors whose signals are evaluated for a state detection.

In the utility model DE 299 11 748 U1 a combination of gas sensors with a GSM modem is presented.

The generic patent EP 0 092 068 B1 concerns an alarm system for gases and / or vapors. According to this solution, resistance values of semiconductor gas sensors are compared with predetermined response curves to detect gas components.

The The object of the invention is thus to provide a method and a Arrangement for identifying and / or differentiating by Sensors displayed substances in gas mixtures and a corresponding Computer program and a corresponding computer-readable storage medium to provide which improve the known solutions, and so in particular to create an intelligent system as a total solution for a safe and fast Detection and detection of events, such as fires and gaseous Air changes.

This object is achieved by the features in claims 1, 13, 21 and 22 are solved. Advantageous embodiments of the invention are contained in the subclaims.

It is for this provided that at least one part is responsive to a sensor the substances detected in gas mixtures characterized by this sensor Parameter and / or at least part of the generated by this sensor Data is automatically evaluated by a computer program and the identification and / or differentiation of substances predetermined criteria and / or patterns.

• – in einem ersten Schritt mittels eines Vergleich der Werte der erfassten Parameter mit vorgegebenen Schwellwerten Toleranzgrenzen eine Voralarmentscheidung getroffen wird, und
• – in Abhängigkeit des Voralarms in einem zweiten Schritt die Auswertung auf der Datenverarbeitungseinrichtung erfolgt.

According to the invention, substances are identified in gas mixtures in such a way that parameters of the gas mixture are detected by a semiconductor or metal oxide sensor in the temperature change process, at least a part of the parameters detected by this sensor by a computer program installed on a data processing device using reference and Evaluated and / or classification patterns and identified by the sensor detected displayed substances are identified, and an alarm decision is made depending on the results of the evaluation. An important aspect of the invention is that the process for the identification of substances in gas mixtures proceeds in two steps, such that

• In a first step, by means of a comparison of the values of the detected parameters with predetermined threshold values of tolerance limits, a prealarm decision is made, and
• - In response to the pre-alarm in a second step, the evaluation is carried out on the data processing device.

• – in einem ersten Schritt mittels eines Vergleich der Werte der erfassten Parameter mit vorgegebenen Toleranzgrenzen eine Voralarmentscheidung getroffen wird, und
• – in Abhängigkeit des Voralarms in einem zweiten Schritt die Auswertung auf der Datenverarbeitungseinrichtung erfolgt.

An arrangement for identifying substances indicated by sensors in gas mixtures is inventively arranged such that it comprises at least one semiconductor or metal oxide sensor and at least one data processing device, wherein the arrangement is arranged such that a method for the identification of substances in gas mixtures is feasible Parameters of the gas mixture are detected by the at least one semiconductor or metal oxide sensor in the temperature change method, at least a part of the parameters detected by the at least one sensor evaluated by a computer program installed on a computer program using reference and / or classification patterns and of the at least a sensor indicated substances are identified, and depending on the results of the evaluation, an alarm decision is made, and wherein the method comprises two steps, wherein

• In a first step, a prealarm decision is made by means of a comparison of the values of the detected parameters with predetermined tolerance limits, and
• - In response to the pre-alarm in a second step, the evaluation is carried out on the data processing device.

• – in einem ersten Schritt mittels eines Vergleich der Werte der erfassten Parameter mit vorgegebenen Toleranzgrenzen eine Voralarmentscheidung getroffen wird, und
• – in Abhängigkeit des Voralarms in einem zweiten Schritt die Auswertung auf der Datenverarbeitungseinrichtung erfolgt.

A computer program for identifying substances displayed by sensors in gaseous mixtures allows a computer, after it has been loaded into the memory of the computer, to carry out a method for identifying substances in gas mixtures, wherein parameters of the gas mixture are determined by a semiconductor or metal oxide sensor in the temperature range. Interchangeable method are detected, and wherein at least a part of the parameters detected by this sensor are evaluated by a computer program installed on a computer program using reference and / or classification patterns and identified by the sensor substances identified, and depending on the results of the evaluation an alarm decision is taken, and wherein the method comprises two steps, wherein

• In a first step, a prealarm decision is made by means of a comparison of the values of the detected parameters with predetermined tolerance limits, and
• - In response to the pre-alarm in a second step, the evaluation is carried out on the data processing device.

For example can these computer programs (for a fee or free of charge, free accessible or password protected) be provided downloadable in a data or communication network. The computer programs thus provided can then be processed by a method be made available, in which a computer program according to claim 21 from an electronic data network such as from the Internet on a connected to the data network data processing device is downloaded.

• – in einem ersten Schritt mittels eines Vergleich der Werte der erfassten Parameter mit vorgegebenen Toleranzgrenzen eine Voralarmentscheidung getroffen wird, und
• – in Abhängigkeit des Voralarms in einem zweiten Schritt die Auswertung auf der Datenverarbeitungseinrichtung erfolgt.

In order to carry out identification of substances displayed by sensors in gas mixtures, it is advantageous to use a computer-readable storage medium on which is stored a program which, after being loaded into the memory of the computer, enables a computer to identify by sensors In the case of gas mixture mixtures, wherein parameters of the gas mixture are detected by a semiconductor or metal oxide sensor in the temperature change method, at least part of the parameters detected by this sensor are evaluated by a computer program installed on a data processing device using reference and / or classification patterns; identified by the sensor, and depending on the Results of the evaluation is made an alarm decision, and wherein the method comprises two steps, wherein

• In a first step, a prealarm decision is made by means of a comparison of the values of the detected parameters with predetermined tolerance limits, and
• - In response to the pre-alarm in a second step, the evaluation is carried out on the data processing device.

A preferred embodiment the method according to the invention provides that in the evaluation by the computer program procedure the artificial one Intelligence be used. In particular, in the evaluation by the computer program neural networks are used.

The brings with it the advantage that thereby the false alarm rate considerably is reduced. supports This is further characterized by being for identification and / or differentiation of substances a library of criteria and / or patterns is used. The inventive method can be optimized by adding the library (s) to the application area the sensors are (are) adapted. In the libraries are then contain only those criteria and / or patterns, the substances, Describe events or states which in the to be monitored Environment can occur real and relevant to the decision.

After this by the method according to the invention a condition has been identified, it turns out in many cases as advantageous that in Dependence on identified substances signaling the detection of these Substances takes place. This procedure will find particular application when using the method according to the invention for fire monitoring and / or for securing closed rooms, in particular containers, against unauthorized opening and / or intrusion.

alternative or parallel to the signaling can also automatically an action initiated or the results obtained by the procedure only saved.

In another preferred embodiment the method according to the invention is provided that the automatic evaluation by the computer program the reaction on earlier Evaluation results taken into account (self-learning system).

One Another advantage of the invention is that parameters are detected and / or transmitted from which the geographical location of sensors can be determined is. This is advantageously a Global Positioning System (GPS) used. This procedure is particularly appropriate when mobile objects, such as containers, equipped with sensors are, as the respective location of the object determined safely can be.

In addition, proves it is advantageous if the sensors and / or one of the sensors comprehensive monitoring device via communication networks be configured remotely.

Especially for large-scale monitoring Areas will be monitoring improved by that Sensors of the monitoring device reticulate in the monitored Area to be installed.

at the surveillance Particularly important areas can be the reliability of the monitoring and Resilience can be improved by using a redundant transmission the parameter and / or the analysis results.

• – mindestens ein Mittel der künstlichen Intelligenz,
• – mindestens eine Referenzbibliothek,
• – mindestens ein Mittel zur Alarm-Ausgabe und/oder
• – mindestens ein Mittel zur visuellen Datenausgabe

umfasst, wobei diese Komponenten der Anordnung durch Mittel zur Datenübertragung miteinander verbunden sind.A preferred embodiment of the arrangement according to the invention provides that the arrangement

• - at least one means of artificial intelligence,
• - at least one reference library,
• - at least one means for alarm output and / or
• - at least one means of visual data output

comprising, wherein these components of the arrangement are interconnected by means for data transmission.

there it proves to be an advantage if the (the) means of artificial Intelligence are designed as a neural network and / or when the Library (s) Classification pattern for changes in air composition includes.

One Another advantage of the invention is that by appropriate. Selection of classification patterns in the libraries for different uses each provided special, optimally adapted sensor systems become. By the method according to the invention are any air changes identifiable; For example, the method can be used for the Intrusion protection of containers, if the libraries patterns for air changes contain an opening of the container.

In another preferred embodiment of the arrangement according to the invention it is provided that the arrangement comprises means with which determine the geographical location of sensors bar is. It is advantageous in particular if the arrangement comprises means with which a Global Positioning System (GPS) can be used.

A Another preferred embodiment of inventive arrangement provides that the Arrangement is remotely configurable.

In spacious, to be monitored Regions is advantageously used an arrangement, wherein the sensors the arrangement installed like a net in the area to be monitored are.

When Another particular advantage of the invention is to be considered, in particular self-learning systems a flexible and fast adaptation to the respective Allow market requirements. It is also advantageous if the Systems are designed to be remotely configurable.

The Sensors realized as semiconductor gas sensors are robust, long-term stable and own it In addition, a very good price / performance ratio.

neural Networks in conjunction with sophisticated electronics bring one new quality At detection with low false alarm rate, the intelligence it consists of, between natural or artificial harmless changes the air and dangerous To distinguish events. A neural network also detects from the data of widely branched installed sensors online the most important pollutants and fumes or the oxygen content of the air. It can be between absolutely dangerous Fire gases and weather-related changes in air composition be differentiated. With that you can fires or escaping gases are clearly identified geographically safe and fast and the messages are transmitted to the control centers.

• – hohe Flexibilität der Einsatzbereiche,
• – Erkennbarkeit und Klassifikation der Brand-Rauch- und anderer Gase,
• – Brandfrüherkennung. Frühzeitige Erkennung des Brandes bereits im Schwelstadium, wo optische Sensoren noch nicht reagieren,
• – genaue geografische Zuordnung (auch GPS),
• – unabhängig von Wetter und Sichtbedingungen,
• – hohe Zuverlässigkeit mit geringen Fehlalarmmeldungen,
• – gute Auswertbarkeit mittels elektronischer Systeme und
• – vielfältige Einsatzbereiche in Industrie, öffentlichen Einrichtungen und privaten Bereichen.

Briefly summarized, the invention has the following advantages:

• - high flexibility of the application areas,
• - recognizability and classification of fire smoke and other gases,
• - Early fire detection. Early detection of the fire already in the smoldering stage, where optical sensors are not yet reacting,
• - exact geographical allocation (also GPS),
• - regardless of weather and visibility conditions,
• - high reliability with low false alarms,
• - good readability by means of electronic systems and
• - diverse applications in industry, public institutions and private sectors.

The Fields of application of the system are manifold. They range from industry up to environmental monitoring and in the private area.

From In particular, the chemical industry should be mentioned in the industry and laboratories, petrochemicals, large-scale monitoring of welding control systems, Leak detection systems for Refinery site, monitoring of pipelines of different gases and energy controlling system and Energy management of buildings and facilities.

Public Facilities (no smoking zones) can also be monitored are like forest areas (fire and environmental influences) or agglomerations with heavy air pollutant load. Ozone monitoring, pollutant detection in fires (Dioxin), parking garage monitoring (Underground car parks), weather stations are other areas of application.

Furthermore the invention can be used for monitoring be used in or by aircraft. Here especially as Early fire detection system, for monitoring of harnesses, the leakage of fuel or for smoking ban monitoring.

• – der Waldbrandfrüherkennung,
• – der automatisierte Nachkontrolle von abgebrannten Waldstücken und/oder
• – der Erfassung von Schadstoffen.

The invention is the forest monitoring system (WMS)

• - early forest fire detection,
• - Automated follow-up of burnt forest and / or
• - the detection of pollutants.

The Invention is further for an ideal protection of tunnels for fire and dangerous Gases (including smog). This is especially for the transportation (Subway) of importance.

in the Construction issues surveillance of shafts and tubes or follow-up welding is an important application Likewise, the invention can be used for early fire detection in buildings (Libraries, hotels), to recognize and distinguish different Fires (cable fire, Paper fire), for monitoring of electrical distributors and for the early detection of gas line defects. Being special For gas pipes, the early detection of danger and a special monitoring of junctions of pipes such as valves, joints important Aspects are that guaranteed by the invention.

The Invention will be described below with reference to the figures of Drawings on several embodiments explained in more detail. It shows:

1 Flowchart for illustrating the method according to the invention,

2 Illustration of the spatial distribution of the modules of the monitoring system according to the invention and the communication between these modules,

3 An illustration of the modular construction of the monitoring system according to the invention,

4 Representation of an exemplary fire supervision in a subway tunnel,

5 Representation of an exemplary use of the invention for tunnel monitoring.

in the The following is an example of the invention of the special case of fire control explained in more detail become. It is important to take into account that the invention is not limited to fire monitoring, but just as good for monitoring chemical installations or as a leak detection system for refinery premises can.

• – Schadstoffanteil der Luft (Gasanalyse)
• – Temperatur
• – Luftfeuchtigkeit und
• – erforderlichenfalls der geografischer Standort.

Based on 1 and 3 the basic procedure for data acquisition and evaluation as well as the modular structure of the monitoring system according to the invention are clarified:
For gas analysis, semiconductor sensors are preferred 1 used with a long life and very high sensitivity (ppm range). From these semiconductor sensors 1 For example, in a first step, the following physical parameters are recorded during condition monitoring:

• - Pollutant content of the air (gas analysis)
• - temperature
• - Humidity and
• - if necessary, the geographical location.

This can be done, for example, with the thermal cycling process.

In a second step, the acquired data is already in the sensor 1 permanently analyzed. This ensures short reaction times in the data acquisition. The analysis includes, for example, a comparison of the pollutant content determined in the current measurement with the pollutant content of the measurement of the preceding cycle and a "normal value." The values of the current measurement deviate from predetermined tolerance limits-that is, from the two comparison values-by a certain amount , a pre-alarm decision takes place and there is a transfer of the data to a control center / headquarters 2 where, inter alia, a data processing device is installed. The data transmission can be via a bus system 4 or fiber optic networks 5 take place, this type of data transmission will be used in particular when the sensor 1 in a solid structure 6 is installed while installing the sensor 1 in a mobile structure 7 a wireless data transmission via radio 8th or other wireless transmission facilities (eg GPRS) will be used (cf. 2 ).

By a computer program installed on this data processing device, the transmitted data are evaluated. Optionally, the results of the evaluation are provided by means for visual data output 9 visualized. This can be a combined 2D display of spatial sensor data and / or 3D visualization of gas sensors and their evaluation by neural networks 10 respectively.

In the case of a positive pre-alarm decision, then a classification of the data takes place. Advantageously, means of artificial intelligence, such as neural networks, are used 10 used, the z. B. by an extensive reference library 11 get supported. In such a library 11 are sensor data, measuring curves, analysis results u. Ä. For the air changes by a variety of substances, such as gasoline, diesel, oil, paper fire, tire blight, cigarette smoke and the like, stored as a reference or classification pattern.

Depending on the results of the evaluation, an alarm decision is made in the following step and, upon fulfillment of criteria, which generally depend on the monitoring event and can be specified accordingly, via an alarm output means 12 Alarm triggered, signaling the occurrence of an event to be monitored. In a special embodiment of the invention, the alarm simultaneously signals which event is involved. It can be signaled, for example, whether a cable fire, sprayer, a wagon fire, a stuck brake, dioxin, oil fire o. Ä. Was detected.

In every stage of the process the data or analysis results or visualizations also to given Posts transmitted become. Therefor can the internet (e-mail) or SMS are used.

By this approach is done by linking the data with adaptive Networks (neural networks) achieved a high (self-) intelligence of the system. One Neural network recognizes from the data of the sensors online the most important Harmful and flue gases or the oxygen content of the air. It can thereby between absolutely dangerous Fumes and other changes distinguish the air composition.

• – kurze Reaktionszeiten in der Datenerfassung,
• – die Möglichkeit, verschiedene Gase mit einem einzigen Sensorsystem zu erkennen,
• – eine schnelle Sensorregeneration auch bei hohen Fremdgasbelastungen sowie
• – eine geringe Fehlalarmwahrscheinlichkeit durch Erkennung von Klassifikationsmustern.

This procedure ensures in particular

• - short reaction times in the data acquisition,
• The ability to detect different gases with a single sensor system,
• - Fast sensor regeneration even at high external gas loads and
• A low false alarm probability by recognition of classification patterns.

• – Brandschutz, Sprayeralarm durch Lösungsmittel und/oder
• – Emissionserkennung, Immissionsschutz

und den Nutzeffekten außerhalb der Fahrzeuge, wie beispielsweise

• – Brandüberwachung von Tunnelanlagen (vgl. 4)
• – Erkennung von defekten Waggonbremsen und elektrischen

The invention is advantageously used for example for monitoring public transport, as in the application example S- u. Underground becomes clear. It can be between the benefits within the vehicles, such as

• - Fire protection, spray alarm by solvents and / or
• - Emission detection, pollution control

and the benefits outside the vehicles, such as

• - Fire supervision of tunnel installations (cf. 4 )
• - Detection of defective wagon brakes and electrical

cables be distinguished by gas emission (distance about 200 to 500 m).

The monitoring of tunnels is in 4 outlined. In the subway tunnel 13 between a first station 14 and a second train station 15 are at intervals, the specific situation in the subway tunnel 13 are adapted, sensors 16 appropriate. These are by data transmission means 17 (Cable or radio) with a data bus 18 which in turn (by data transfer means 17 ) with the central control center 2 connected is. Under certain circumstances, the control center 2 be integrated into already existing (monitoring) systems.

A structural representation for an exemplary tunnel monitoring is in 5 played. Here are the sensors 19 via a fiber-optic BUS system 20 along the tunnel route 21 arranged at a distance of 100 to 500 m. It can prove handy if multiple sensors 19 to a sensor cluster 22 be summarized. The sensors 19 are about these sensor clusters 22 by data connection, e.g. B. via a fiber optic cable 20 which is in the tunnel route 21 with one or more control centers 2 connected. As it turns out, if at each tunnel entrance ever a control center 2 is set up. In order to increase the reliability, can in the control centers 2 redundant server farms (eg Linux cluster). As a redundancy system to further increase the reliability is a radio network 23 used, over which the sensors 19 or sensor cluster 22 with the control center (s) 2 are connected.

• – eine hohe Ausfallsicherheit durch Parallelmeldung über Funk,
• – die Möglichkeit, eine unbegrenzte Anzahl an Sensoren einzusetzen und
• – die Erkennung unterschiedlichster Emissionen, indem eine große Schadstoffbibliothek zugrundegelegt wird.

The advantages that result from this structure are

• - a high reliability by parallel message via radio,
• - the ability to use an unlimited number of sensors and
• - the detection of different emissions by using a large pollutant library.

Both in tunnels and in large-scale forest areas, the sensors can be installed net-like. The sensors constantly analyze the air quality. If a local fire develops, this information is immediately displayed in the control center 2 evaluated. By being able to determine by GPS the exact location of each sensor, it is possible to accurately locate the fire and its size to the meter. For the fire department a time gain. In addition, a follow-up inspection of the affected area can be carried out automatically.

The Sensors detect not only gases, but also humidity and temperatures. For example, forest fire alert levels can be individual and accurately determined and ultimately even objectified accordingly become.

One Another area of application is the monitoring of high quality Technology such. B. the monitoring of IT technology or data cabinet monitoring in data centers. Above all, the early detection play here of critical situations like overheating, smoldering and burn one Role. This will be a timely shutdown of the technique (de-energize) or the deletion (eg with nitrogen) ensured by active control. Furthermore is a surveillance regarding the condition of the batteries, power failure and / or Sabotage possible.

This can be used as remote monitoring respectively; as well the redundant ones mentioned in connection with tunnel monitoring Components (double transmission in parallel over Cable and radio, redundant server farms).

Another field of application of the invention lies in the area of intrusion protection. Since any air changes can be identified by a properly equipped library, the procedure can also be used for the intrusion protection of containers in shipping or air traffic. The libraries used for this purpose must contain patterns for air changes that signal an opening of the container. Connected with Mit This enables a quick and targeted response to unauthorized opening of a container, which determines the geographic location of the containers.

1

Sensor, Semiconductor gas sensor

2

Control center / center

3

module for data processing and visualization

4

bus system

5

fiber-optic network

6

firm structures

7

mobile structures

8th

Wireless

9

medium for visual data output

10

neural network

11

Library, reference library

12

medium for alarm output

13

Subway Tunnel

14

first railway station

15

second railway station

16

sensor

17

Data transmission means

18

bus

19

sensor

20

fiber-optic bus system

21

tunnel route

22

Sensor Cluster

23

radio network

Claims (22)

Method for identifying substances in gas mixtures, wherein parameters of the gas mixture are detected by a semiconductor or metal oxide sensor in the temperature change method, and wherein at least part of the parameters detected by this sensor are detected by a computer program installed on a data processing device using reference and / or or classification patterns are evaluated and identified by the sensor substances are identified, and wherein an alarm decision is made depending on the results of the evaluation, characterized in that - in a first step by means of a comparison of the values of the detected parameters with predetermined tolerance limits a prealarm decision is made and - in response to the pre-alarm in a second step, the evaluation is carried out on the data processing device. Method according to claim 1, characterized in that that in the evaluation by the computer program method of artificial Intelligence be used. Method according to one of claims 1 or 2, characterized that the automatic evaluation by the computer program the Reaction to earlier ones Evaluation results taken into account (self-learning system). Method according to one of claims 1 to 3, characterized that in the evaluation by the computer program neural networks be used. Method according to one of claims 1 to 4, characterized that for the identification and / or differentiation of substances Library of criteria and / or patterns is used. Method according to claim 5, characterized in that that the library adapted to the application of the sensors becomes. Method according to one of claims 1 to 6, characterized that sensors acquire and / or transmit data that make up the geographical location of these sensors can be determined. Method according to claim 7, characterized in that that for determining the geographical location of sensors Global Positioning System (GPS) is used. Method according to one of claims 1 to 8, characterized that the sensors and / or a monitoring device comprising the sensors via communication networks be configured remotely. Method according to one of claims 1 to 9, characterized that the sensors of the monitoring device reticulate in one to be monitored Area to be installed. Method according to one of claims 1 to 10, characterized that a redundant transmission the parameter and / or the analysis results. Method according to one of claims 1 to 11, characterized that depending on identified substances signaling the detection of these Substances takes place. Arrangement comprising at least one semiconductor or metal oxide sensor ( 1 ) and at least one data processing device, wherein the arrangement is set up such that a method for the identification of substances in gas mixtures can be carried out, wherein parameters of the gas mixture are determined by the at least one semiconductor or metal oxide sensor ( 1 ) in the tem and at least a portion of which of the at least one sensor ( 1 ) are evaluated by a computer program installed on a data processing device using reference and / or classification patterns and by the at least one sensor ( 1 ) and in which an alarm decision is made as a function of the results of the evaluation, and the method comprises two steps, wherein - in a first step, by means of a comparison of the values of the detected parameters with predetermined tolerance limits, a prealarm decision is made, and in response to the pre-alarm in a second step, the evaluation is carried out on the data processing device. Arrangement according to claim 13, characterized in that the arrangement - at least one means of artificial intelligence, - at least one reference library ( 11 ), - at least one alarm output means ( 12 ) and / or - at least one means of visual data output ( 9 ), wherein these components of the arrangement are interconnected by means for data transmission. Arrangement according to claim 14, characterized in that the means (s) of artificial intelligence as neural network ( 10 ) are formed. Arrangement according to one of Claims 13 to 15, characterized that the library (s) classification patterns for changes in air composition includes (include). Arrangement according to one of claims 13 to 16, characterized in that the arrangement comprises means by which the geographical location of sensors ( 1 ) can be determined. Arrangement according to one of claims 13 to 17, characterized that the arrangement comprises means by which a Global Positioning System (GPS) is usable. Arrangement according to one of claims 13 to 18, characterized that the arrangement is remotely configurable. Arrangement according to one of claims 13 to 19, characterized in that the sensors ( 1 ) of the arrangement are net-like installed in an area to be monitored. Computer program that allows a computer to after being loaded into the memory of the computer To carry out a method for the identification of substances in gas mixtures, wherein parameter the gas mixture by a semiconductor or metal oxide in the Temperature change methods are recorded, and at least a part of the parameters detected by this sensor by a a computer program installed under computer Use of reference and / or classification patterns evaluated and substances indicated by the sensor are identified, and in which dependent on the results of the evaluation an alarm decision is made, and the method comprises two steps, wherein - in one first step by comparing the values of the detected Parameters with predetermined tolerance limits a prealarm decision is taken, and - in dependence the pre-alarm in a second step, the evaluation on the data processing device he follows. Computer-readable storage medium on which a program It allows a computer after it has been stored in the memory of the computer Computer has been loaded, a method of identification of To carry out substances in gas mixtures, in which Parameter of the gas mixture by a semiconductor or Metal oxide sensor to be detected in the temperature change process, and wherein at least a part of the detected by this sensor Parameter by an installed on a data processing device Computer program using reference and / or classification patterns evaluated and identified by the sensor substances are identified, and being dependent the results of the evaluation an alarm decision is made, and the method comprises two steps, wherein - in one first step by comparing the values of the detected Parameters with predetermined tolerance limits a prealarm decision is taken, and - in dependence the pre-alarm in a second step, the evaluation on the data processing device he follows.