Flame detection method and flame detection system

Abstract

This invention provides a means to realize a multi-wavelength flame sensor with fewer false detections compared to conventional technologies. [Solution] In the flame detection method according to the present invention, first amplitude time-dependent change data, which shows the amplitude of a signal output from a first sensor that responds to the energy amount of the component in the first wavelength band of light, and second amplitude time-dependent change data, which shows the amplitude of a signal output from a second sensor that responds to the energy amount of the component in the second wavelength band of light, are acquired. Subsequently, a linear function is identified by linearly approximating the combination of the amplitude shown by the first amplitude time-dependent change data and the amplitude shown by the second amplitude time-dependent change data at each of a plurality of time points. The slope of the linear function thus identified represents the ratio of the amplitude of the second wavelength band to the amplitude of the first wavelength band (spectral ratio). Subsequently, the presence or absence of a flame in the monitoring area is determined by whether or not the slope of the linear function representing the spectral ratio is within a predetermined range.

Description

Technical Field

[0001] The present invention relates to a technique for detecting a flame.

Background Art

[0002] There is a device (hereinafter referred to as a "multi-wavelength flame sensor") that determines the presence or absence of a flame in a monitoring area based on the strength of signals output from a plurality of light-receiving sensors that respond to light in different wavelength bands.

[0003] As a patent document that discloses a technique related to a multi-wavelength flame sensor, for example, there is Patent Document 1. Patent Document 1 includes two light-receiving units that output light-receiving signals that have observed different wavelength bands, and when the cross-correlation of the light-receiving signals output by these light-receiving units is strong, the presence or absence of a flame is determined based on the ratio of the integral values of these light-receiving signals, and when the cross-correlation is weak, a flame detection device having a configuration in which the presence or absence of a flame is not determined is described. Patent Document 1 describes that a more reliable determination can be made by determining the presence or absence of a combustion flame based on the simultaneity of the temporal changes of the light-receiving signals output from different light-receiving units.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Generally, a multi-wavelength flame sensor determines the presence or absence of a flame based on whether the ratio of the amplitudes of signals (hereinafter referred to as "light-receiving signals") output by each of a plurality of light-receiving sensors falls within a predetermined range.

[0006] The amplitude of the light signal output by each of the multiple light-receiving sensors in a multi-wavelength flame sensor indicates the amount of energy of the light component in the wavelength band corresponding to that light-receiving sensor. Therefore, the series of light signals continuously output by each light-receiving sensor indicates the change in the amount of energy of the light component in the wavelength band corresponding to that light-receiving sensor over time.

[0007] In this application, the light-receiving signal output by the light-receiving sensor may be either an analog signal showing a continuous value in the amplitude axis direction or a digital signal showing a discrete value in the amplitude axis direction. Furthermore, in this application, the series of light-receiving signals output by the light-receiving sensor may be either a continuous signal in the time axis direction or a discontinuous signal in the time axis direction.

[0008] Figure 9 is a schematic graph illustrating an example of a series of received signals output by each of the two light-receiving sensors in a multi-wavelength flame sensor. In the graphs shown in Figure 9, the horizontal axis represents time, and the vertical axis represents amplitude. Graph G1 in Figure 9 shows the change in amplitude of the received signal output by the first light-receiving sensor over time, and graph G2 shows the change in amplitude of the received signal output by the second light-receiving sensor over time.

[0009] In an example of a conventional multi-wavelength flame sensor, the peak point P1 and bottom point B1 of graph G1, and the corresponding peak point P2 and bottom point B2 of graph G2 are identified. The amplitude difference D1 between the peak point P1 and bottom point B1 of graph G1 and the amplitude difference D2 between the peak point P2 and bottom point B2 of graph G2 are calculated. For example, if the ratio of the amplitude difference D2 to the amplitude difference D1 is within a predetermined range, it is determined that there is a flame within the monitoring area; if the ratio is outside the predetermined range, it is determined that there is no flame within the monitoring area.

[0010] In the conventional multi-wavelength flame sensor described above, the peak and bottom points of graphs G1 and G2 need to be correctly identified, but correctly identifying these points is not easy for reasons such as the following.

[0011] (1) The received signal shown in graph G1 and the received signal shown in graph G2 each contain different unknown DC levels (bias values), so it is necessary to identify these DC levels. Methods such as calculating a moving average or low-pass filtering can be used to identify the DC levels, but the DC levels identified will differ depending on the method and parameters used (e.g., the cutoff frequency of the low-pass filter), and it is not always possible to uniquely identify the DC levels. In other words, there is an element of arbitrariness in identifying the DC levels. (2) It is necessary to determine the period of the waves shown in graph G1 and graph G2, but since the waves shown in graph G1 and graph G2 include waves of varying magnitudes, it is not always possible to uniquely determine the period of these waves. For example, one method to determine the period of graph G1 and graph G2 is to identify the zero-crossing points of graph G1 and graph G2 when the DC level is set to zero, and to define the time between adjacent zero-crossing points in the time axis as half a period. However, the period determined will differ depending on whether or not smoothing is performed on graph G1 and graph G2 for noise reduction, and on the method and parameters of the smoothing (e.g., the cutoff frequency of the four-pass filter). Therefore, it is not always possible to uniquely determine the period. In other words, there is an element of arbitrariness in determining the period. (3) Graphs G1 and G2 may show the time-dependent changes in signals in response to light from different light sources. In such cases, graphs G1 and G2 will not be synchronized. However, it is not possible to uniquely determine how similar the waveforms of graphs G1 and G2 must be to determine if they are synchronized, or how different they must be to determine if they are not synchronized. In other words, there is an element of arbitrariness in determining whether or not two waveforms are synchronized. (4) If the amplitude difference between the peak and bottom points of the identified graph G1, or between the peak and bottom points of the identified graph G2, is small, then those peak and bottom points should be removed as noise peaks and bottom points. However, it is not possible to uniquely determine how small the amplitude difference must be for a peak and bottom point to be removed as noise. In other words, there is an element of arbitrariness in determining which peaks and bottom points should be removed as noise.

[0012] Therefore, in the conventional multi-wavelength flame sensor described above, parameters such as thresholds need to be adjusted according to the environment in which the multi-wavelength flame sensor is placed in order to identify the correct peak and bottom points. If this adjustment is not performed properly, false flame detection is likely to occur.

[0013] Furthermore, in the conventional multi-wavelength flame sensors described above, there is a risk that abnormal values ​​caused by noise may be mistaken for peak or bottom points. Therefore, false flame detection is likely to occur due to noise.

[0014] In view of the circumstances described above, the present invention provides a means for achieving flame detection with fewer false detections compared to the case using a conventional multi-wavelength flame sensor. [Means for solving the problem]

[0015] To solve the above problems, the present invention proposes a flame detection method and a flame detection system for executing the flame detection method, comprising the steps of: acquiring first amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the amount of energy of components in a first wavelength band of light during a certain period; acquiring second amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the amount of energy of components in a second wavelength band of light during the same period; identifying a linear function that linearly approximates the relationship between amplitudes in a combination of the amplitude at a given time shown by the first amplitude time-dependent change data and the amplitude at a given time shown by the second amplitude time-dependent change data for each of a plurality of time points within the period; and determining the presence or absence of a flame based on the slope of the linear function.

[0016] Furthermore, the present invention proposes a flame detection method and a flame detection system for executing the flame detection method, comprising the steps of: acquiring first amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the amount of energy of components in a first wavelength band of light during a certain period; acquiring second amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the amount of energy of components in a second wavelength band of light during the same period; identifying a linear function that linearly approximates the relationship between amplitudes in a combination of the amplitude of a frequency represented by the first frequency-dependent change data and the amplitude of a frequency represented by the second frequency-dependent change data, when the frequency-dependent change data of the wave represented by the first amplitude time-dependent change data is defined as the first frequency-dependent change data and the frequency-dependent change data of the wave represented as the second frequency-dependent change data, for each of a plurality of frequencies; and determining the presence or absence of a flame based on the slope of the linear function.

[0017] Furthermore, the present invention proposes a flame detection method and a flame detection system for executing the flame detection method, comprising the steps of: acquiring first amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the amount of energy of components in a first wavelength band of light during a certain period; acquiring second amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the amount of energy of components in a second wavelength band of light during the same period; calculating a correlation coefficient between phases in a combination of the phase of a frequency represented by the first frequency-dependent change data and the phase of a frequency represented by the second frequency-dependent change data, with respect to each of a plurality of frequencies, where the frequency-phase characteristics of the wave represented by the first amplitude time-dependent change data are defined as the first frequency-phase characteristics and the frequency-phase characteristics of the wave represented by the second amplitude time-dependent change data are defined as the second frequency-phase characteristics; and determining the presence or absence of a flame based on the correlation coefficient.

[0018] Further, the present invention provides a flame detection method comprising steps of: obtaining first amplitude change data over time representing a change over time of an amplitude indicated by a series of signals output from a first sensor that responds to an energy amount of a component in a first wavelength band of light during a certain period, and second amplitude change data over time representing a change over time of an amplitude indicated by a series of signals output from a second sensor that responds to an energy amount of a component in a second wavelength band of light during the period; when the frequency-phase characteristics of the wave represented by the first amplitude change data over time are defined as first frequency-phase characteristics and the frequency-phase characteristics of the wave represented by the second amplitude change data over time are defined as second frequency-phase characteristics, specifying a linear function that linearly approximates the relationship between phases in a combination of the phase of the frequency indicated by the first frequency-phase characteristics and the phase of the frequency indicated by the second frequency-phase characteristics for each of a plurality of frequencies; and determining the presence or absence of a flame based on the slope of the linear function. The present invention also proposes a flame detection system that executes the flame detection method.

Advantages of the Invention

[0019] According to the flame detection method or the flame detection system according to the present invention, flame detection with less false detection is realized as compared with the case of using a multi-wavelength flame sensor according to the prior art.

Brief Description of the Drawings

[0020] [Figure 1] A diagram showing the overall configuration of a flame detection system according to an embodiment. [Figure 2] A diagram showing the functional configuration of a data processing unit according to an embodiment. [Figure 3] A graph for explaining the processing of a data processing unit according to an embodiment. [Figure 4] A graph for explaining the processing of a data processing unit according to an embodiment. [Figure 5] A graph for explaining the processing of a data processing unit according to an embodiment. [Figure 6] A graph for explaining the processing of a data processing unit according to a modification. [Figure 7] A graph for explaining the processing of a data processing unit according to a modification. [Figure 8] A graph for explaining the processing of a data processing unit according to a modified example. [Figure 9] A graph for explaining the processing of a multi-wavelength flame sensor according to an example of the prior art.

BEST MODE FOR CARRYING OUT THE INVENTION

[0021] [Embodiment] The flame detection system 1 according to an embodiment of the present invention will be described below. FIG. 1 is a diagram showing the overall configuration of the flame detection system 1. The flame detection system 1 includes a sensor unit 11 and a data processing unit 12.

[0022] The sensor unit 11 includes a first sensor 111, a second sensor 112, and a communication interface 113.

[0023] The first sensor 111 includes a light receiving element 1111 and a filter 1112. The light receiving element 1111 is an element that outputs an analog signal according to the amount of energy of the received light. The filter 1112 is a band-pass filter that transmits components in a predetermined wavelength band included in the light and does not transmit components in other wavelength bands.

[0024] The second sensor 112 includes a light receiving element 1121 and a filter 1122. The light receiving element 1121 is an element that outputs an analog signal according to the amount of energy of the received light, similar to the light receiving element 1111. The filter 1122 is a band-pass filter that transmits components in a predetermined wavelength band included in the light and does not transmit components in other wavelength bands, similar to the filter 1112. However, the filter 1112 and the filter 1122 have different wavelength bands of the components of the transmitted light. Hereinafter, the wavelength band of the components of the light transmitted by the filter 1112 (an example of the first wavelength band) is referred to as the first wavelength band, and the wavelength band of the components of the light transmitted by the filter 1122 (an example of the second wavelength band) is referred to as the second wavelength band.

[0025] The communication interface 113 converts the analog signal output from the photodetector 1111 into digital data (hereinafter referred to as "first amplitude data") and continuously transmits the first amplitude data to the data processing unit 12 by wire or wireless connection. Hereinafter, the series of first amplitude data continuously transmitted by the sensor unit 11 to the data processing unit 12 will be referred to as the first amplitude time-dependent change data. The first amplitude time-dependent change data represents the time-dependent change in amplitude shown by a series of signals output from the first sensor 111, which responds to the amount of energy of the components in the first wavelength band of light.

[0026] Furthermore, the communication interface 113 converts the analog signal output from the photodetector 1121 into digital data (hereinafter referred to as "second amplitude data") and continuously transmits the second amplitude data to the data processing unit 12 by wire or wireless connection. Hereinafter, the series of second amplitude data continuously transmitted by the sensor unit 11 to the data processing unit 12 will be referred to as second amplitude time-dependent change data. The second amplitude time-dependent change data represents the time-dependent change in amplitude shown by a series of signals output from the second sensor 112, which responds to the amount of energy of the components in the second wavelength band of light.

[0027] The sensor unit 11 is arranged so that the light-receiving elements 1111 and 1121 receive light coming from the monitoring area. The monitoring area is the area where the presence or absence of flames is monitored.

[0028] The data processing unit 12 is a computer. The data processing unit 12 includes a memory 121 for storing data and a processor 122 for performing various data processing according to the program stored in the memory 121. The data processing unit 12 also includes a communication interface 123 for receiving first amplitude data and second amplitude data transmitted from the sensor unit 11, and an input / output interface 124 to which external devices are connected. Notification devices (not shown) such as a display for displaying various information to the user and a speaker for audibly outputting various information to the user are connected to the input / output interface 124.

[0029] Note that the sensor unit 11 and the data processing unit 12 include components such as a battery in addition to the components shown in Figure 1, but Figure 1 only shows the components necessary to explain the features of the flame detection system 1, and other components are omitted.

[0030] Figure 2 is a diagram showing the functional configuration of the data processing unit 12. That is, the processor 122 of the data processing unit 12 performs data processing according to the program of this embodiment, thereby functioning as a device equipped with the components shown in Figure 2. The data processing unit 12 includes, as functional components, an acquisition means 1201, a storage means 1202, a linear function identification means 1203, a correlation coefficient calculation means 1204, and a determination means 1205.

[0031] The acquisition means 1201 acquires the first amplitude time change data and the second amplitude time change data transmitted from the sensor unit 11. The storage means 1202 stores various data. For example, the first amplitude time change data and the second amplitude time change data acquired by the acquisition means 1201 are stored in the storage means 1202.

[0032] The linear function identification means 1203 identifies a linear function that linearly approximates the relationship between amplitudes in combinations of the amplitude at a given time represented by the first amplitude time change data and the amplitude at a given time represented by the second amplitude time change data, for each of several time points within a certain period.

[0033] Figure 3 is a graph showing the time-dependent changes in amplitude represented by the first amplitude time-dependent change data and the second amplitude time-dependent change data over a certain period. Graph E1 shows the time-dependent changes in amplitude represented by the first amplitude time-dependent change data, and graph E2 shows the time-dependent changes in amplitude represented by the second amplitude time-dependent change data.

[0034] The linear function identification means 1203 identifies the amplitudes represented by the first amplitude time-dependent change data at each of the time points t1, t2, ..., tn (where n is any natural number), i.e., amplitudes e1(t1), e1(t2), ..., e1(tn). The linear function identification means 1203 also identifies the amplitudes represented by the second amplitude time-dependent change data at each of the time points t1, t2, ..., tn (where n is any natural number), i.e., amplitudes e2(t1), e2(t2), ..., e2(tn).

[0035] Next, the linear function identification means 1203 combines the amplitude represented by the first amplitude time change data and the amplitude represented by the second amplitude time change data corresponding to the same time point as follows. (e1(t1),e2(t1)) (e1(t2),e2(t2)) ... ... ... (e1(tn),e2(tn))

[0036] Next, the linear function identification means 1203 identifies a linear function that linearly approximates the relationship between the amplitudes in the above combinations. Figure 4 is a scatter plot composed of a collection of plots corresponding to the above combinations, and a graph showing a straight line L1 that approximates those plots. In the graph of Figure 4, the horizontal axis represents the amplitude shown by the first amplitude time change data, that is, the amplitude of the first wavelength band in the time domain, and the vertical axis represents the amplitude shown by the second amplitude time change data, that is, the amplitude of the second wavelength band in the time domain. The linear function identification means 1203 identifies a linear function represented by the straight line L1 (hereinafter referred to as "linear function L1") (an example of the first linear function) by, for example, the least squares method. The linear function L1 thus identified represents the relationship between the amplitude in the time domain of the first wavelength band and the amplitude in the time domain of the second wavelength band.

[0037] The data indicating the linear function L1 identified by the linear function identification means 1203 is stored in the storage means 1202.

[0038] The correlation coefficient calculation means 1204 calculates the correlation coefficient between phases in combinations of the phase of a frequency shown by the frequency-phase characteristics of the wave represented by the first amplitude time-change data (hereinafter referred to as the "first frequency-phase characteristics") and the phase of a frequency shown by the frequency-phase characteristics of the wave represented by the second amplitude time-change data (hereinafter referred to as the "second frequency-phase characteristics") for each of the multiple frequencies.

[0039] Figure 5 is a graph showing the first and second frequency phase characteristics. Graph F1 shows the first frequency phase characteristic, i.e., the frequency phase characteristic of the wave represented by the first amplitude time-dependent data, and graph F2 shows the second frequency phase characteristic, i.e., the frequency phase characteristic of the wave represented by the second amplitude time-dependent data. In graphs F1 and F2, the horizontal axis represents frequency, and the vertical axis represents phase.

[0040] The correlation coefficient calculation means 1204 performs, for example, a Fast Fourier Transform on the first amplitude time-dependent change data for a certain period to identify the first frequency phase characteristic shown in graph F1 of Figure 5. The correlation coefficient calculation means 1204 also performs, for example, a Fast Fourier Transform on the second amplitude time-dependent change data for the same period to identify the second frequency phase characteristic shown in graph F2 of Figure 5.

[0041] The correlation coefficient calculation means 1204 identifies the phases shown by the first frequency phase characteristics at each of the frequencies f1, f2, ..., fm (where m is any natural number), i.e., phases g1(f1), g1(f2), ..., g1(fm). The correlation coefficient calculation means 1204 also identifies the phases shown by the second frequency phase characteristics at each of the frequencies f1, f2, ..., fm (where m is any natural number), i.e., phases g2(f1), g2(f2), ..., g2(fm).

[0042] Next, the correlation coefficient calculation means 1204 combines the phase shown by the first frequency phase characteristic and the phase shown by the second frequency phase characteristic corresponding to the same frequency as follows. (g1(f1),g2(f1)) (g1(f2),g2(f2)) ... ... ... (g1(fm),g2(fm))

[0043] Next, the correlation coefficient calculation means 1204 calculates the correlation coefficient between the phases in the above combination (hereinafter referred to as "correlation coefficient C"). The correlation coefficient C calculated by the correlation coefficient calculation means 1204 is stored in the storage means 1202.

[0044] The determination means 1205 determines the presence or absence of flames in the monitoring area based on the slope of the linear function L1 identified by the linear function identification means 1203 and the correlation coefficient C calculated by the correlation coefficient calculation means 1204.

[0045] The slope of the linear function L1 represents the ratio of the amplitude of the second wavelength band to the amplitude of the first wavelength band (hereinafter referred to as the "spectral ratio").

[0046] Furthermore, the correlation coefficient C indicates the synchronization of the time-dependent changes in amplitude in the first wavelength band and the time-dependent changes in amplitude in the second wavelength band. That is, if the correlation coefficient C is sufficiently close to 1, it can be seen that the time-dependent changes in amplitude in the first wavelength band and the time-dependent changes in amplitude in the second wavelength band are synchronized, and that the first sensor 111 and the second sensor 112 are receiving light from the same light source.

[0047] The determination means 1205 determines, for example, that there is a flame in the monitoring area if the slope of the linear function L1 is within a predetermined range corresponding to the spectral ratio of the flame, and the correlation coefficient C is within a predetermined range with an upper limit of 1 (for example, 0.8 to 1.0, etc.), and determines that there is no flame in the monitoring area otherwise.

[0048] The above is a description of the functional configuration of the data processing unit 12.

[0049] The data processing unit 12 instructs notification devices such as displays and speakers connected to the data processing unit 12 to notify them of the result of determining whether or not there is a flame in the monitoring area, i.e., the result of flame detection. The notification devices, in accordance with the instructions from the data processing unit 12, notify the user of whether or not there is a flame in the monitoring area.

[0050] According to the flame detection system 1 described above, the presence or absence of a flame is determined based on the spectral ratio, which is shown by the slope of a linear function, and the presence or absence of synchronization, which is shown by the correlation coefficient. Therefore, flame detection with high noise immunity is performed regardless of the environment in which the sensor unit 11 is placed. As a result, flame detection with fewer false detections is achieved compared to conventional technology.

[0051] [Differentiation] The embodiments described above are specific examples of the present invention and can be modified in various ways within the scope of the technical idea of ​​the present invention. Examples of such modifications are shown below. Two or more of the following modifications may be combined as appropriate.

[0052] (1) In the above-described embodiment, the spectral ratio, shown by the slope of a linear function L1 that shows the relationship between the amplitude in the time domain of the first wavelength band shown by the first amplitude time change data and the amplitude in the time domain of the second wavelength band shown by the second amplitude time change data, is used to determine whether or not there is a flame. Alternatively, the spectral ratio, shown by the slope of a linear function that shows the relationship between the amplitude in the frequency domain of the first wavelength band shown by the first amplitude time change data and the amplitude in the frequency domain of the second wavelength band shown by the second amplitude time change data, may be used to determine whether or not there is a flame.

[0053] In this modified example, the linear function identification means 1203 identifies a linear function that linearly approximates the relationship between the amplitudes in a combination of the amplitude of a wave represented by the first amplitude time-varying data (hereinafter referred to as the "first frequency amplitude characteristic") and the amplitude of a wave represented by the second amplitude time-varying data (hereinafter referred to as the "second frequency amplitude characteristic") for each of a plurality of frequencies.

[0054] Figure 6 is a graph showing the first and second frequency amplitude characteristics. Graph H1 shows the first frequency amplitude characteristic, i.e., the frequency amplitude characteristic of the wave represented by the first amplitude time-dependent change data, and graph H2 shows the second frequency amplitude characteristic, i.e., the frequency amplitude characteristic of the wave represented by the second amplitude time-dependent change data. In graphs H1 and H2, the horizontal axis represents frequency, and the vertical axis represents amplitude.

[0055] The linear function identification means 1203 performs a Fast Fourier Transform, for example, on the first amplitude time-dependent change data for a certain period to identify the first frequency amplitude characteristic shown in graph H1 of Figure 6. The linear function identification means 1203 also performs a Fast Fourier Transform, for example, on the second amplitude time-dependent change data for the same period to identify the second frequency amplitude characteristic shown in graph H2 of Figure 6.

[0056] The linear function identification means 1203 identifies the amplitudes shown by the first frequency amplitude characteristics at each of the frequencies h1, h2, ..., hp (where p is an arbitrary natural number), i.e., amplitudes j1(h1), j1(h2), ..., j1(hp). The linear function identification means 1203 also identifies the amplitudes shown by the second frequency amplitude characteristics at each of the frequencies h1, h2, ..., hp (where p is an arbitrary natural number), i.e., amplitudes j2(h1), j2(h2), ..., j2(hp).

[0057] Next, the linear function identification means 1203 combines the amplitude shown by the first frequency amplitude characteristic and the amplitude shown by the second frequency amplitude characteristic corresponding to the same frequency as follows. (j1(h1),j2(h1)) (j1(h2),j2(h2)) ... ... ... (j1(hp), j2(hp))

[0058] Next, the linear function identification means 1203 identifies a linear function that linearly approximates the relationship between the amplitudes in the above combinations. Figure 7 is a scatter plot composed of a collection of plots corresponding to the above combinations, and a graph showing the straight line L2 that approximates those plots. In the graph of Figure 7, the horizontal axis represents the amplitude shown by the first frequency amplitude characteristic, that is, the amplitude of the first wavelength band in the frequency domain, and the vertical axis represents the amplitude shown by the second frequency amplitude characteristic, that is, the amplitude of the second wavelength band in the frequency domain. The linear function identification means 1203 identifies a linear function represented by the straight line L2 (hereinafter referred to as "linear function L2") (an example of the first linear function) by, for example, the least squares method.

[0059] The linear function L2, as defined above, represents the relationship between the amplitude in the frequency domain of the first wavelength band and the amplitude in the frequency domain of the second wavelength band. The slope of the linear function L2, similar to the slope of the linear function L1 in the embodiment described above, represents the spectral ratio, i.e., the ratio of the amplitude of the second wavelength band to the amplitude of the first wavelength band.

[0060] The data representing the linear function L2 identified by the linear function identification means 1203 is stored in the storage means 1202.

[0061] In this modified example, the determination means 1205 determines the presence or absence of flames in the monitoring area by using the slope of the linear function L2 identified by the linear function identification means 1203 as described above, instead of the slope of the linear function L1 in the embodiment.

[0062] In other words, the determination means 1205 determines that there is a flame in the monitoring area if the slope of the linear function L2 is within a predetermined range corresponding to the spectral ratio of the flame, and the correlation coefficient C is within a predetermined range with an upper limit of 1 (for example, 0.8 to 1.0, etc.), and determines that there is no flame in the monitoring area otherwise.

[0063] In this modified example, as described above, the presence or absence of a flame is determined based on the spectral ratio shown by the slope of a linear function L2 that represents the relationship between the amplitude of the first wavelength band and the amplitude of the second wavelength band in the frequency domain. This linear function L2 is identified from the scatter plot exemplified in Figure 7. If the first sensor 111 and the second sensor 112 receive light from the same light source and output the correct signal, the correlation coefficient between the amplitudes of the first wavelength band and the second wavelength band at the same frequency in the frequency domain, as plotted in this scatter plot, should be sufficiently close to 1. Therefore, if the correlation coefficient deviates significantly from 1, it is highly likely that the slope of the linear function L2 does not represent the spectral ratio of the flame, either because the first sensor 111 and the second sensor 112 receive light from different light sources, or because the signal output by the first sensor 111 or the second sensor 112 contains noise.

[0064] Therefore, the correlation coefficient calculation means 1204 may calculate the correlation coefficient between amplitudes in the following combination of amplitudes in the first wavelength band and amplitudes in the second wavelength band in the frequency domain (hereinafter referred to as "correlation coefficient Q"), which the linear function identification means 1203 uses to identify the linear function L2 in the modified example described above, and the determination means 1205 may determine the presence or absence of a flame based on the correlation coefficient Q. (j1(h1),j2(h1)) (j1(h2),j2(h2)) ... ... ... (j1(hp), j2(hp))

[0065] The determination means 1205 determines, for example, that there is a flame in the monitoring area if the slope of the linear function L2 is within a predetermined range corresponding to the spectral ratio of the flame, the correlation coefficient Q is within a predetermined range with an upper limit of 1 (for example, 0.8 to 1.0, etc.), and the correlation coefficient C is within a predetermined range with an upper limit of 1 (for example, 0.8 to 1.0, etc.). Otherwise, it determines that there is no flame in the monitoring area.

[0066] (2) In the above-described embodiment, the presence or absence of a flame is determined based on a correlation coefficient C that shows the correlation between the phase in the frequency domain of the first wavelength band shown by the first amplitude time change data and the phase in the frequency domain of the second wavelength band shown by the second amplitude time change data. Alternatively, or in addition to this, the presence or absence of a flame may be determined based on the slope of a linear function that shows the relationship between the phase in the frequency domain of the first wavelength band shown by the first amplitude time change data and the phase in the frequency domain of the second wavelength band shown by the second amplitude time change data.

[0067] In this modified example, the linear function identification means 1203 uses the following combination of phases shown by the first frequency phase characteristic and the second frequency phase characteristic, which is used by the correlation coefficient calculation means 1204 in the above-described embodiment to calculate the correlation coefficient C. (g1(f1),g2(f1)) (g1(f2),g2(f2)) ... ... ... (g1(fm),g2(fm))

[0068] The linear function identification means 1203 may generate the above combinations, or the linear function identification means 1203 may use the above combinations generated by the correlation coefficient calculation means 1204.

[0069] The linear function identification means 1203 identifies a linear function that linearly approximates the relationship between the phases in the above combination.

[0070] Figure 8 is a scatter plot composed of a collection of plots corresponding to the above combinations, and a graph showing a straight line L3 that approximates these plots. In the graph of Figure 8, the horizontal axis represents the phase shown by the first frequency phase characteristic, that is, the phase of the first wavelength band in the frequency domain, and the vertical axis represents the phase shown by the second frequency phase characteristic, that is, the phase of the second wavelength band in the frequency domain. The linear function identification means 1203 identifies a linear function represented by the straight line L3 (hereinafter referred to as "linear function L3") (an example of a second linear function) for example by the least squares method.

[0071] When the first sensor 111 and the second sensor 112 receive light from the same light source, the slope of the linear function L3 will be a value sufficiently close to 1. Therefore, in this modified example, the determination means 1205 determines, for example, that there is a flame in the monitoring area if the slope of the linear function L1 is within a predetermined range corresponding to the spectral ratio of the flame, and the slope of the linear function L3 is within a predetermined range including 1 (for example, 0.9 to 1.1, etc.), and determines that there is no flame in the monitoring area otherwise.

[0072] (3) In the above-described embodiment, the presence or absence of a flame is determined based on the spectral ratio shown by the slope of a linear function L1 that represents the relationship between the amplitude in the time domain of the first wavelength band and the amplitude in the time domain of the second wavelength band. This linear function L1 is identified from the scatter plot illustrated in Figure 4, and if the first sensor 111 and the second sensor 112 receive light from the same light source and output the correct signal, the correlation coefficient between the amplitudes of the first wavelength band and the second wavelength band at the same time point in the time domain, as plotted in this scatter plot, should be sufficiently close to 1. Therefore, if the correlation coefficient deviates significantly from 1, it is highly likely that the slope of the linear function L1 does not represent the spectral ratio of the flame, either because the first sensor 111 and the second sensor 112 receive light from different light sources, or because the signal output by the first sensor 111 or the second sensor 112 contains noise.

[0073] Therefore, the correlation coefficient calculation means 1204 may calculate the correlation coefficient between amplitudes in the combination of amplitudes in the first wavelength band and amplitudes in the second wavelength band in the time domain (hereinafter referred to as "correlation coefficient K"), which the linear function identification means 1203 uses to identify the linear function L1 in the above embodiment, and the determination means 1205 may determine the presence or absence of a flame based on the correlation coefficient K. (e1(t1),e2(t1)) (e1(t2),e2(t2)) ... ... ... (e1(tn),e2(tn))

[0074] The determination means 1205 determines, for example, that there is a flame in the monitoring area if the slope of the linear function L1 is within a predetermined range corresponding to the spectral ratio of the flame, the correlation coefficient K is within a predetermined range with an upper limit of 1 (for example, 0.8 to 1.0, etc.), and the correlation coefficient C is within a predetermined range with an upper limit of 1 (for example, 0.8 to 1.0, etc.). Otherwise, it determines that there is no flame in the monitoring area.

[0075] (4) In the embodiments described above, the flame detection system 1 uses the spectral ratio shown by the slope of a linear function to determine the presence or absence of a flame. Alternatively, the flame detection system 1 may use a spectral ratio used in the prior art instead of the spectral ratio shown by the slope of a linear function to determine the presence or absence of a flame. For example, the flame detection system 1 may calculate the spectral ratio as the ratio of the difference between the peak and bottom values ​​of the amplitude in the time domain of the second wavelength band (see Figure 9) to the difference between the peak and bottom values ​​of the amplitude in the time domain of the first wavelength band, and use this ratio to determine the presence or absence of a flame.

[0076] (5) In the embodiments described above, the flame detection system 1 calculates and uses a correlation coefficient that indicates the synchronization of the time-dependent changes in amplitude of the first wavelength band and the second wavelength band as an index value that indicates the synchronization of the phases of the first wavelength band and the second wavelength band in determining the presence or absence of a flame. The flame detection system 1 may, instead of the correlation coefficient that indicates the synchronization of the time-dependent changes in amplitude of the two wavelength bands, calculate and use an index value that indicates the synchronization of phases used in the prior art (for example, a cross-correlation function between the first amplitude time-dependent change data and the second amplitude time-dependent change data) in determining the presence or absence of a flame.

[0077] Furthermore, the flame detection system 1 may determine the presence or absence of a flame without using phase synchronization, for example, based solely on the spectral ratio.

[0078] (6) In the embodiment described above, the number of light receiving sensors in the flame detection system 1 is assumed to be two, but the flame detection system 1 may be equipped with three or more light receiving sensors and perform flame detection based on three or more signals output by those light receiving sensors.

[0079] (7) In the above-described embodiment, the number of sensor units 11 connected to the data processing unit 12 is assumed to be one, but multiple sensor units 11 may be connected to the data processing unit 12.

[0080] (8) In the embodiments described above, the flame detection system 1 is configured with a sensor unit 11 and a data processing unit 12, but the configuration of the flame detection system 1 is not limited to this. For example, the flame detection system 1 may be configured with a single device that has the functions of a sensor unit 11 and a data processing unit 12.

[0081] (9) In the embodiments or modifications described above, a Fast Fourier Transform is performed to identify the frequency amplitude characteristics or frequency phase characteristics. However, the transformation method used to identify the frequency domain data from the time domain data is not limited to the Fast Fourier Transform. For example, the Discrete Fourier Transform or the like may be used.

[0082] (10) The present invention provides a flame detection system exemplified by the flame detection system 1 described above, a flame detection method exemplified by the processing performed by the data processing unit 12 described above, and a program for causing a computer to execute the processing exemplified by the processing performed by the data processing unit 12 described above. [Explanation of Symbols]

[0083] 1...Flame detection system, 11...Sensor unit, 12...Data processing unit, 111...First sensor, 112...Second sensor, 113...Communication interface, 121...Memory, 122...Processor, 123...Communication interface, 124...Input / output interface, 1111...Light receiving element, 1112...Filter, 1121...Light receiving element, 1122...Filter, 1201...Acquisition means, 1202...Storage means, 1203...Linear function identification means, 1204...Correlation coefficient calculation means, 1205...Determination means.

Claims

1. The steps include acquiring first amplitude time-dependent data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the energy amount of components in a first wavelength band of light during a certain period, and second amplitude time-dependent data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the energy amount of components in a second wavelength band of light during the same period. When the frequency amplitude characteristics of the wave represented by the first amplitude time-dependent change data are defined as the first frequency amplitude characteristics, and the frequency amplitude characteristics of the wave represented by the second amplitude time-dependent change data are defined as the second frequency amplitude characteristics, the step of identifying a linear function that linearly approximates the relationship between the amplitudes in combinations of the amplitude of that frequency shown by the first frequency amplitude characteristics and the amplitude of that frequency shown by the second frequency amplitude characteristics for each of a plurality of frequencies, The steps include determining the presence or absence of a flame based on the slope of the aforementioned linear function, and A flame detection method equipped with [specific features / features].

2. The step of calculating the correlation coefficient between the amplitudes in the above combination is included. In the step of determining the presence or absence of a flame, the presence or absence of a flame is determined based on the correlation coefficient in addition to the slope of the linear function. The flame detection method according to claim 1.

3. When the frequency-phase characteristics of the wave represented by the first amplitude-time change data are defined as the first frequency-phase characteristics, and the frequency-phase characteristics of the wave represented by the second amplitude-time change data are defined as the second frequency-phase characteristics, the method includes a step of calculating the correlation coefficient between the phases in the combination of the phase of the frequency represented by the first frequency-phase characteristics and the phase of the frequency represented by the second frequency-phase characteristics for each of a plurality of frequencies. In the step of determining the presence or absence of a flame, the presence or absence of a flame is determined based on the correlation coefficient in addition to the slope of the linear function. The flame detection method according to claim 1.

4. When the aforementioned linear function is the first linear function, When the frequency-phase characteristics of the wave represented by the first amplitude-time change data are defined as the first frequency-phase characteristics, and the frequency-phase characteristics of the wave represented by the second amplitude-time change data are defined as the second frequency-phase characteristics, the method includes the step of identifying a linear function as the second linear function which is a linear approximation of the relationship between the phases in a combination of the phase of the frequency represented by the first frequency-phase characteristics and the phase of the frequency represented by the second frequency-phase characteristics for each of a plurality of frequencies. In the step of determining the presence or absence of a flame, the presence or absence of a flame is determined based on the slope of the second linear function in addition to the slope of the first linear function. The flame detection method according to claim 1.

5. The steps include acquiring first amplitude time-dependent data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the energy amount of components in a first wavelength band of light during a certain period, and second amplitude time-dependent data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the energy amount of components in a second wavelength band of light during the same period. When the frequency-phase characteristics of the wave represented by the first amplitude-time change data are defined as the first frequency-phase characteristics, and the frequency-phase characteristics of the wave represented by the second amplitude-time change data are defined as the second frequency-phase characteristics, the step of calculating the correlation coefficient between the phases in the combination of the phase of the frequency represented by the first frequency-phase characteristics and the phase of the frequency represented by the second frequency-phase characteristics for each of the multiple frequencies, A step of determining the presence or absence of flames based on the correlation coefficient. A flame detection method equipped with [specific features / features].

6. The steps include acquiring first amplitude time-dependent data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the energy amount of components in a first wavelength band of light during a certain period, and second amplitude time-dependent data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the energy amount of components in a second wavelength band of light during the same period. When the frequency-phase characteristics of the wave represented by the first amplitude-time change data are defined as the first frequency-phase characteristics, and the frequency-phase characteristics of the wave represented by the second amplitude-time change data are defined as the second frequency-phase characteristics, the step of identifying a linear function that linearly approximates the relationship between the phases in the combination of the phase of that frequency shown by the first frequency-phase characteristics and the phase of that frequency shown by the second frequency-phase characteristics for each of a plurality of frequencies, The steps include determining the presence or absence of a flame based on the slope of the aforementioned linear function, and A flame detection method equipped with [specific features / features].

7. Acquisition means for acquiring first amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the energy amount of components in a first wavelength band of light during a certain period, and second amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the energy amount of components in a second wavelength band of light during the same period. When the frequency amplitude characteristics of the wave represented by the first amplitude time change data are defined as the first frequency amplitude characteristics, and the frequency amplitude characteristics of the wave represented by the second amplitude time change data are defined as the second frequency amplitude characteristics, a linear function identification means identifies a linear function that linearly approximates the relationship between the amplitudes in a combination of the amplitude of the frequency represented by the first frequency amplitude characteristics and the amplitude of the frequency represented by the second frequency amplitude characteristics for each of a plurality of frequencies, A determination means for determining the presence or absence of a flame based on the slope of the aforementioned linear function, A flame detection system equipped with [the necessary components].

8. Acquisition means for acquiring first amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the energy amount of components in a first wavelength band of light during a certain period, and second amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the energy amount of components in a second wavelength band of light during the same period. When the frequency-phase characteristics of the wave represented by the first amplitude-time change data are defined as the first frequency-phase characteristics, and the frequency-phase characteristics of the wave represented by the second amplitude-time change data are defined as the second frequency-phase characteristics, a correlation coefficient calculation means calculates the correlation coefficient between phases in a combination of the phase of the frequency represented by the first frequency-phase characteristics and the phase of the frequency represented by the second frequency-phase characteristics for each of a plurality of frequencies. A determination means for determining the presence or absence of a flame based on the correlation coefficient. A flame detection system equipped with [the necessary components].

9. Acquisition means for acquiring first amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a first sensor that responds to the energy amount of components in a first wavelength band of light during a certain period, and second amplitude time-dependent change data representing the time-dependent change in amplitude of a series of signals output from a second sensor that responds to the energy amount of components in a second wavelength band of light during the same period. When the frequency-phase characteristics of the wave represented by the first amplitude-time change data are defined as the first frequency-phase characteristics, and the frequency-phase characteristics of the wave represented by the second amplitude-time change data are defined as the second frequency-phase characteristics, a linear function identification means identifies a linear function that linearly approximates the relationship between the phases in a combination of the phase of the frequency represented by the first frequency-phase characteristics and the phase of the frequency represented by the second frequency-phase characteristics for each of a plurality of frequencies, A determination means for determining the presence or absence of a flame based on the slope of the aforementioned linear function, A flame detection system equipped with [the necessary components].

Images