Patent Application
20080309935
Ling, B., Zeifman, M., Hu, J. (2006), “A Practical and Inexpensive System for Natural Gas Leak Detection,” Proceedings of ISA Technical Conference, Houston, Tex.
O'Brien, J. J., Cao, H. (2002), “Cross sections from Absorption Spectra and Absorption Coefficients for Methane in the 750-940 nm Region Obtained by Intracavity Laser Spectroscopy,” Journal of Quantitative Spectroscopy and Radiative Transfer, 75, pp. 323-350.
Platt, U. (1994) “Differential optical absorption spectroscopy (DOAS),” in Air Monitoring by Spectroscopic Techniques,” M. W. Sigrist, ed., Chemical Analysis Series (Wiley, New York), Vol. 127.
Reichardt, T. A., W. Einfeld, and T. J. Kulp (2002): “Review of Remote Detection for Natural Gas Transmission Pipeline Leaks,” final report on “Evaluation of Active and Passive Gas Imagers for Transmission Pipeline Remote Leak Detection” project report, Sandia National Laboratory.
(1) Field of the Invention
The present invention generally relates to a new method to calculate the spectral absorbance utilizing sunlight without any other illumination sources. This invention more particularly relates to computer and/or electronic and mechanical methods and systems for detecting molecules in air, including methane (CH4), when the wavelength range of interest overlaps with sunlight spectra.
(2) Background Information
In the US natural gas transport system, the underground piping includes approximately 400,000 miles of transmission pipelines and 1.4 million miles of distribution piping, while above-ground piping is located mainly at about 750 gas processing plants and some 3000 compressor stations (Reichardt, Einfeld, and Kulp 2002). These pipelines are often disrupted by leaks. Regulatory pressure is increasing to inspect transmission lines more frequently. Remote gas leak survey is a proactive way to prevent unnecessary loss of human life resulted from natural gas leaks.
Since the existing active detection system is too expensive and passive detection system is less reliable, a spectroscopic-optical-based system has been developed, which utilizes the sunlight to detect the methane within a specific wavelength range such as 850-950 nm (Ling, Zeifman and Hu 2006) or 1300-1700 nm. Being able to detect methane within this range is significant, which makes it possible to dramatically reduce the system cost associated with an expensive sensing instrument, an expensive laser device, and complex and costly system operations.
The leak detection technologies can be roughly divided into in-situ methods and remote sensing methods. The in-situ methods currently being used by gas utility companies include sensor-less methods such as observation of odor, unusual vegetation, or a hissing sound, and sensor-based methods that use either Combustible Gas Indicator (CGI) or Flame Ionization (FI) gas detector. Specifically, the most widely used gas leak survey tool is FI detector that uses a hydrogen fuel to power a small flame in a detector cell. A pump system is used to pass continuous air samples through the detector cell. If the air contains hydrocarbons such as natural gas, they will be burned or ionized in the hydrogen flame. FI detector is highly sensitive when it is close to the sample being tested.
There are two major techniques used for remote gas leak detection: (1) active detection, which requires illuminating the scene with a radiation source such as a laser, and (2) passive detection, which relies on radiative transfer resulted from the temperature difference that usually exists between the background and the target methane (CH4) plume. Active detection removes the constraint of thermal difference, but requires a laser and a scattering surface behind the gas for echo signals. While passive methods allow a long range of detection with a relatively simple thermal imaging device, these methods rely upon a thermal flux between the gas plume and the ground surface below it.
As regulated by individual states in US, gas utility companies must perform gas leak surveys using a gas detection instrument covering various areas including:
Therefore, an efficient and reliable remote natural gas leak detector is desired.
In one aspect, the present invention includes a method to estimate the number of vacuum references required for specific applications such as gas leak detection. In particular, a variable power supply is used to set the photon counts at desired levels. A pipe is used, together with a connected vacuum pump, to generate both air and vacuum environment inside the pipe. With this setting, both pipe references and corresponding vacuum references can be measured.
In another aspect, this invention includes a systematic procedure for collecting references. This procedure is made of six steps including light intensity change, photon counts measurement, vacuuming pipe, and measuring and recording both pipe and vacuum references.
In yet another aspect, this invention includes method to systematically build two libraries, one for the pipe references, and the other for the vacuum references. References in each library can be retrieved using index.
In still a further aspect, this invention includes a procedure to systematically estimate the vacuum reference associated with one particular solar reference measured. This procedure consists of four steps including solar reference measurement, matching pipe reference in the pipe reference library, retrieving vacuum reference from the vacuum reference library, and finally, use the matched vacuum reference and measured solar reference to estimate the absorbance.
Beer-Lambert's law governs the estimation of spectral absorbance. It specifies the linear relationship between absorbance and concentration of absorbing species, including methane to be detected in the outdoor open air. The general Beer-Lambert's law is usually given as:
Aλ=ελbc (1)
where Aλ is the absorbance, ελ is the absorptivity coefficient at wavelength λ, b is the path length of species (such as methane, CH4), c is the concentration of the same species. It can be observed that absorbance is proportional to both path length and species concentration. One example of theoretical absorbance of methane can be found in (O'Brien 2002).
To obtain the absorbance from a spectrometer, one needs to first take the “dark” scan and “vacuum” scan. Basically, the “dark” scan will measure the light intensity in a totally dark environment while the “vacuum” scan is done in a vacuum environment. Mathematically, the absorbance at pixel n can be calculated using the following formula:
where N is the number of pixels supported by a spectrometer, sn is the real-time light intensity measurement from the spectrometer, vn is the light intensity measurement in the vacuum environment, and dn is the light intensity measured in a totally dark environment.
Therefore, in order to have an accurate absorbance, we must measure the light intensity in both “dark” and “vacuum” environment.
Since dn is close to zero, Eq. (2) can be further simplified as
The value of An is always non-negative since vn≧sn for n=1, 2, . . . , N. Eq. (2) provides a physical basis for the calculation of spectral absorbance. To actually calculate the absorbance An, one must measure the light intensity in a “vacuum” environment to obtain sn, termed as vacuum reference signal.
Since no vacuum reference signal can be obtained in an open-path spectroscopic remote sensing, absorbance An cannot be readily calculated using Eq. (3). Here we disclose a method used to estimate the vacuum reference signal.
Refer to
Refer to
Light measurement device 160 can be any devices capable of measuring the light intensities. For example, a spectrometer can be used. The choice of this device is solely based on applications in interest.
The computation device 180 is used to collect light measurement from the light measurement device 160 through a predefined I/O interface. The absorbance calculation can be done in this device.
As we change the output voltage or current of the variable power supply 122, light intensity from light bulb changes as well, or equivalently, photon counts measured by the light measurement device 160 change. Let Δ be the smallest possible change in photon counts. Δ can be estimated through a series of experiments by varying the power supply output voltage/current and observing the changes in photon counts. Since the light measurement device 160, light bulb 126, and variable power supply 122 operate in certain error tolerances, the photon counts measurement contain noises. The appropriate value of Δ must be larger than the noise band.
Let Smin and Smax be the upper and lower limits of photon counts for specific applications. The number of vacuum references can be calculated using the following equation:
M=round((Smax−Smin)/Δ) (4)
where the function round( ) converts a real number to an integer. For the ith pipe reference, for i=1, 2, . . . , M, the photon counts will be Smin+iΔ.
The procedure of measuring ith vacuum reference for i=1, 2, . . . , M is listed as follows:
Repeat Steps 1-6 to obtain the rest of pipe references and vacuum references. For total M pipe and vacuum references, two libraries will be built: LIBpipe and LIBvacuum. LIBpipe will have M pipe references while LIBvacuum has Mvacuum references. All references in both libraries are arranged in either ascending or descending order. They are also indexed for fast retrieval.
Refer to
Refer to
Refer to Eq. (3). The solar reference, {si, i=1, 2, . . . , N}, is the actual sunlight photon counts measured at difference wavelength. Its corresponding vacuum reference is required to calculate the absorbance.
The procedure of estimating corresponding vacuum reference is listed as follows:
Although this invention has been described according to an exemplary embodiment, it should be understood by those of ordinary skill in the art that modifications may be made without departing from the spirit of the invention. The scope of the invention is not to be considered limited by the description of the invention set forth in the specification, but rather as defined by our claims.
