Indirect gas species monitoring using tunable diode lasers
View Patent ↗A method for indirect gas species monitoring based on measurements of selected gas species is disclosed. In situ absorption measurements of combustion species are used for process control and optimization. The gas species accessible by near or mid-IR techniques are limited to species that absorb in this spectral region. The absorption strength is selected to be strong enough for the required sensitivity and is selected to be isolated from neighboring absorption transitions. By coupling the gas measurement with a software sensor gas, species not accessible from the near or mid-IR absorption measurement can be predicted.
1. A method for indirect gas species monitoring for monitoring a gas sample of a combustion process for burning a combustible comprising, in combination:
using an on-line, analytical technique to measure gas species in the sample of a combustion process, the gas species being selected from species that are detectable by the on-line, analytical technique and that correlate to at least one process parameter to be predicted;
developing the software sensor by analyzing process data using manipulated inputs to obtain desired outputs to generate data for software sensor development;
verifying the compatibility of the initial correlation;
applying off-line filters and statistical analyses as necessary to reduce measurement noise, outliers and offsets from the data;
determining a neural-network based software sensor;
determining the network structure;
training the neural network by determining the values of the weights and biases of the network; and
coupling the gas species measurement with the software sensor to predict the process parameter to be predicted.
2. The method described in claim 1 , further comprising adjusting the ratio of O 2 to the combustible to better achieve a stoichiometric balance.
3. The method described in claim 1 , wherein the on-line analytical technique comprises optical absorption techniques.
4. The method described in claim 3 , wherein the optical absorption technique comprises a diode laser.
5. A method for indirect gas species monitoring for monitoring a gas sample of a combustion process for burning a combustible comprising, in combination:
using optical absorption techniques to measure gas species in the sample of a combustion process, the gas species being selected from the group consisting of O 2 , CO, CO 2 and H 2 O;
analyzing process data using manipulated inputs to obtain desired outputs to generate data for software sensor development;
measuring the H 2 level present using a temporary measuring technique;
verifying the compatibility of the initial correlation between the measured gas species data and measured H 2 data;
applying off-line filters and statistical analyses as necessary to reduce measurement noise, outliers and offsets from the data;
determining a neural-network based software sensor;
determining the network structure;
training the neural network by determining the values of the weights and biases of the network;
coupling the gas species measurement with the software sensor to predict the H 2 level present.
6. The method described in claim 5 , further comprising adjusting the ratio of O 2 to the combustible to better achieve a stoichiometric balance.
7. A method for indirect gas species monitoring for monitoring an off-gas of a combustion process for burning a combustible comprising, in combination:
obtaining an off-gas sample of a combustion process to be monitored, the off-gas sample containing at least one compound selected from the group consisting of O 2 , CO, CO 2 and H 2 O;
launching a beam from a launch module housing optic components constructed and arranged to shape the beam to a desired geometry;
shaping the beam to a desired geometry;
transporting the beam through a first shielding pipe and intersecting the beam with the off-gas sample as the beam exits the first shielding pipe;
receiving the beam with a second shielding pipe and transporting it to a detector module housing a detection system;
generating an absorption signal from the module;
directing the absorption signal to an optical unit housing a data acquisition system;
developing a software sensor; and
coupling the gas species measurement with the software sensor to predict the H 2 level present.
8. A method for indirect gas species monitoring for monitoring an off-gas of a combustion process for burning a combustible comprising, in combination:
obtaining an off-gas sample of a combustion process to be monitored, the off-gas sample containing at least one compound selected from the group consisting of O 2 , CO, CO 2 and H 2 O;
launching a beam from a launch module housing optic components constructed and arranged to shape the beam to a desired geometry;
shaping the beam to a desired geometry;
transporting the beam through a first shielding pipe and intersecting the beam with the off-gas sample as the beam exits the first shielding pipe;
receiving the beam with a second shielding pipe and transporting it to a detector module housing a detection system;
generating an absorption signal from the module;
directing the absorption signal to an optical unit housing a data acquisition system;
analyzing process data using manipulated inputs to obtain desired outputs to generate data for software sensor development;
measuring the H 2 level present using a temporary measuring technique;
verifying the compatibility of the initial correlation between the measured gas species data and measured H 2 data;
applying off-line filters and statistical analyses as necessary to reduce measurement noise, outliers and offsets from the data;
determining a neural-network based software sensor;
determining the network structure;
training the neural network by determining the values of the weights and biases of the network; and
coupling the gas species measurement with the software sensor to predict the H 2 level present.
9. The method described in claim 8 , further comprising adjusting the ratio of O 2 to the combustible to better achieve a stoichiometric balance.
10. The method described in claim 5 , wherein the optical absorption technique comprises a diode laser.
11. A method for indirect gas species monitoring for monitoring a gas sample of a combustion process for burning a combustible comprising, in combination:
using optical absorption techniques to measure gas species in the sample of a combustion process, the gas species being selected from the group consisting of O 2 , CO, CO 2 and H 2 O;
analyzing process data using manipulated inputs to obtain desired outputs to generate data for software sensor development;
measuring a level present of a further gas species selected from the group consisting of NO, NOx, and SOx using a temporary measuring technique;
verifying the compatibility of the correlation between the measured gas species data and the level present of the further gas species;
applying off-line filters and statistical analyses as necessary to reduce measurement noise, outliers and offsets from the data;
determining a neural-network based software sensor;
determining the network structure;
training the neural network by determining the values of the weights and biases of the network; and
coupling the gas species measurement with the software sensor to predict the level of the further gas species present.
12. The method described in claim 11 , further comprising adjusting the ratio of O 2 to the combustible to better achieve a stoichiometric balance.