JP3174188B2 - Gas identification device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a gas identification device,
In particular, the present invention relates to a gas discriminating apparatus including a plurality of gas sensors that output according to the type of gas to be detected and information processing means for pattern-recognizing a combination of output values of each gas sensor.

[0002]

2. Description of the Related Art In recent years, there has been an increasing demand in various fields to detect gases and odors with good selectivity. To meet this demand, various gas sensors having excellent stability and sensitivity have been developed. I have. However, from the viewpoint of selectivity to a specific gas or the like, one having sufficiently satisfactory performance has not yet been developed.

In order to detect a specific gas or the like, a plurality of gas sensors having different characteristics are prepared, and a combination of output signals from these gas sensors is regarded as one output pattern, which is used as a neural network. Attempts to identify gas types by performing pattern recognition with various information processing means are generally made.

That is, for all the gas types to be identified, an output pattern is obtained in advance by combining the output signals from the respective gas sensors using the standard gas, and the output pattern is obtained by an information processing means such as a neural network. Recognize and memorize, compare the actually measured output pattern of the detected gas with the output pattern of the standard gas memorized in the information processing means, and determine whether or not there is a similarity. Is determined.

However, a gas discriminating apparatus comprising a plurality of gas sensors and information processing means such as a neural network or the like usually has only one set of information processing means. When the detected gas contains two or more types of gases and has various mixing ratios, it is difficult to perform accurate qualitative / quantitative discrimination.

However, if an analytical instrument such as a gas chromatography device is used, it is possible to sufficiently cope with the case where two or more kinds of gases are contained in the gas to be detected, but such an analytical instrument is expensive. In addition, it is not practical because of poor portability.

The present invention has been made to solve the above-mentioned problems of the prior art, and has a problem in that the detected gas contains two or more types of gases and has various component ratios. Even in such a case, it is an object of the present invention to provide an inexpensive and simple gas discriminating apparatus capable of accurately performing qualitative / quantitative discrimination of a gas constituting a detected gas.

[0008]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides a plurality of gas sensors having different characteristics and at least two sets of information processing for pattern-recognizing combinations of output values of these gas sensors. Means for identifying a gas type constituting the detected gas by at least one set of information processing means among these information processing means, and determining a mixture ratio of the gas type by another information processing means. Means, and the information processing means is a neural network.

[0009]

According to the present invention, by employing the above-described means, even if the gas to be detected is composed of two or more kinds of gases, the gas to be detected is constituted by at least one set of information processing means. The type of gas to be used is identified, and the remaining information processing means can identify the mixing ratio of each gas.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the gas identification device according to the present invention will be described below. FIG. 1 is a block diagram showing an embodiment of a gas identification device according to the present invention. The gas identification device includes a plurality of gas sensors having different characteristics and a combination of output values from these gas sensors. At least two sets of information processing means for pattern recognition.

At least one set of the information processing means identifies (qualitatively identifies) the type of gas contained in the gas to be detected (hereinafter referred to as "component gas type"), and the remaining set includes
It is configured to identify (quantitatively identify) the mixing ratio of the component gas species. In this case, the information processing means includes:
A pattern recognition method using a neural network or a principal component analysis method of multivariate analysis is preferable, but the present invention is not limited thereto, and any method can be used as long as a similar effect can be obtained.

As the gas sensor, 6 having different characteristics are used.
Various kinds of metal oxide semiconductor type gas sensors (commercially available) were used, and as a gas to be detected, a mixture of three kinds of gases of isobutane, ethanol and methane was used.

Next, the operation of the above-described device will be described. First, using the standard gas of each gas (isobutane, ethanol, methane) to be identified, prepare several kinds of mixed gas with various combinations of component gas types and various mixing ratios. An output pattern based on a combination of output signals from the respective gas sensors is obtained, and this is stored in the information processing means as a reference pattern, and the pattern is recognized.

That is, first, the output patterns are classified using the first set of the plurality of information processing means. Here, the output patterns are classified for each component gas type. It is stored so that it is treated as the same group regardless of the type.

Next, using other remaining information processing means, output patterns are classified into component gas types in the same manner as described above, and these output patterns are stored for each component gas in correspondence with the mixing ratio.

As described above, at least one of the information processing means is used as information processing means for determining the component gas type from the output pattern, and the remaining other information processing means is used for determining the mixing ratio of each component gas type. It is configured as a means.

Then, in order to actually measure the gas species constituting the gas to be detected and the mixing ratio thereof, an output pattern of the gas to be detected is obtained from an output signal from the gas sensor, and then the first set of information processing is performed. The component gas type of the gas to be detected is identified from the output pattern using the means, and then the information processing means corresponding to the identified component gas type is used to determine the mixing ratio from the same output pattern.

As described above, since the qualitative identification and the quantitative identification for the gas to be detected are performed by the separate information processing means, the characteristics of the output pattern corresponding to the gas type and the mixing ratio are classified. This reduces the burden of storing and storing, and enables accurate identification of the component gas species and quantitative measurement of the mixing ratio for the mixed gas.

Next, the actual measurement procedure will be described.
First, three kinds of gases, i.e., isobutane, ethanol, and methane, and these are used as component gases, and the mixing ratio is 0 to 100
The output from the gas sensor is measured for various types of mixed gases changed to 10% in increments of 10%, and the neural network learns output patterns obtained from these output values as learning data.

That is, first, component gas type discrimination learning (back propagation learning) is performed using the first set of a plurality of neural networks. At this time, a signal corresponding to only the component gas type is used regardless of the mixing ratio. Is input to the output layer of the neural network (if the component gas types are the same, learning is performed so that the neural network outputs the same determination result regardless of the mixing ratio).

As a result, the number of units in the output layer only needs to be equal to the number of combinations of component gases, and it is not necessary to consider quantitative information on the gas to be detected. The network structure is simple and learning is easy. .

Next, using another remaining set of neural networks, the mixture ratio of the component gas species is learned.

That is, after the output patterns are classified for each combination of the component gas types, the same number of neural networks as the number of the combinations are prepared, and a signal corresponding to the mixture ratio for each combination is used as teacher data. This is given to the output layer of each neural network. As a result, each neural network can output a signal corresponding to the mixing ratio of the corresponding component gas type from the output pattern of the detected gas.

Thus, each neural network only needs to learn the mixing ratio for a specific component gas type, and the learning is facilitated, and the discrimination accuracy for the mixing ratio is improved.

Next, actual measurement results will be described.
First, after measuring the output pattern for the gas to be detected,
Using the first set of neural networks, the component gas type of the gas to be detected is identified from this output pattern, and then the neural network corresponding to the identified component gas type is used. The ratio was determined.

FIG. 2 shows an example of the identification result at this time.
In the figure, 11 sets of upper and lower two sets of graphs show the output state of the neural network when the mixture ratio of isobutane and methane is changed in a range of 0 to 100% in increments of 10% as the gas to be detected. I have.

The upper graph shows the result of discrimination of component gas types by the first set of neural networks, and the vertical axis corresponds to the output layer unit of the neural network indicating the gas type. Further, the lower graph shows a state in which the mixture ratio of the component gas species is determined by the neural network selected from the identification result in the upper column. This vertical axis corresponds to the output layer unit of the neural network indicating the mixing ratio of the component gases. The horizontal axis indicates the output values of these output layer units in both the upper and lower stages.

It can be seen from the upper diagram that the component gas species can be identified for the detected gases having different mixing ratios without being substantially affected by the mixing ratio. Also, as is clear from the lower diagram, the mixing ratio of the component gas species can be read from the output result of the neural network.

It should be noted that the upper and lower tiers 2
By making a comprehensive judgment based on the identification results of a set of neural networks, it is possible to avoid erroneous recognition determinations caused by measurement errors.

For example, in FIG. 2, a false recognition is made that the qualitative identification result suggests the mixing of isobutane with respect to the detected gas of 100% methane. This is presumably because the output patterns of methane 100% and 90% are very similar, and the neural network misjudged. However, looking at the result of the quantitative identification at this time, it is determined that methane is 100%. Therefore, it can be seen from the identification results of these two sets of neural networks that even if isobutane is contained as a component gas, its content is extremely small.

That is, by making the individual neural networks perform qualitative identification and quantitative identification, a relationship is established in which the first neural network is backed up by another neural network, and the identification accuracy is improved. I have.

As described above, in the present invention, a plurality of information processing apparatuses for performing pattern recognition are used, and the roles of the qualitative identification and the quantitative identification of the gas to be detected are shared between them. Compared to the conventional method in which the means performs both qualitative identification and quantitative identification, the configuration of each information processing means can be simplified, and the number of data per set of information processing means can be reduced. Pattern recognition becomes easy. Further, since an effect that a plurality of information processing units have a complementary relationship with each other is obtained, it is possible to significantly improve identification accuracy. Therefore, it is possible to identify the component gas species and quantitatively identify the mixture ratio thereof with high accuracy for various detected gases including the mixed gas.

[0033]

According to the present invention, a plurality of neural networks for performing pattern recognition, that is, a plurality of information processing means, are assigned to perform qualitative identification and quantitative identification of a gas to be detected. Therefore, the structure of each neural network can be simplified as compared with the conventional method in which a single information processing means performs both qualitative identification and quantitative identification. Therefore, there is an excellent effect that learning can be facilitated because a small amount of learning data is required, and the identification accuracy can be greatly improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a gas identification device according to the present invention.

FIG. 2 is an explanatory diagram showing a result measured using the gas identification device shown in FIG.

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int. Cl. ⁷ , DB name) G01N 27/00

Claims (2)

(57) [Claims] 1. A system comprising: a plurality of gas sensors having different characteristics; and at least two sets of information processing means for pattern-recognizing a combination of output values of the gas sensors. A gas discriminating apparatus characterized in that a gas type constituting a detected gas is identified and another information processing means determines a mixing ratio of the gas type. 2. The gas identification device according to claim 1, wherein said information processing means is a neural network.