Method for converting a measurement of local velocity of a fluid in a channel or a duct into a mean velocity

A device for determining with a processor a mean velocity of a fluid flowing in an open channel or a partially filled duct, said fluid having a free surface of width L1 extending between the walls of the channel or duct, said device comprising:

• an energy source directed at a zone of width L2 covering at least a portion of the surface of the flowing fluid, the energy source generating pulses;
• a sensor for performing velocity measurements based on the pulses being reflected at the surface of the fluid over the zone of width L2 smaller than or equal to the width L1;
• said sensor providing a signal in the form of a spectrum of discrete data expressed in a temporal domain, the signal having a frequency shift proportional to velocity of particles at the surface of the flowing fluid;
• a processor to:
• convert the spectrum of discrete data expressed in the temporal domain into a spectrum of data expressed in a frequency domain via a Fourier transform, the process being configured to fit a Gaussian curve to the spectrum of discrete data expressed in the frequency domain, and to calculate the mean μ and the standard deviation σ of the Gaussian curve;
• the frequency of the mean μ enabling calculating mean surface velocity over the zone of width L2 and the standard deviation σ enabling calculating velocity distribution over the same zone of width L2;
• calculate the ratio L2/L1 between the width L2 of the zone impacted during the measurements and the width L1 of the free surface of the fluid to determine whether one of three cases is met;
• the processor being self-learning by detecting conditions under which the spectrum of discrete data obtained is recordable, and determining the mean velocity of the fluid within the open channel or the partially filled duct free of finite element or finite difference modeling, the mean velocity based on the ratio L2/L1 along with the mean μ and the standard deviation σ of the Gaussian curve, the processor calculating the mean velocity with increased efficiency and precision using fewer calculations from the measurement of the surface velocity using the reduced pulsed zone of width L2;
• wherein, in a first case, L1 is equal to L2, the mean velocity of the fluid within the open channel or partially filled duct is equal to the mean surface velocity of the fluid over the zone of width L2;
• wherein, in a second case, the ratio L2/L1 is greater than or equal to ⅔ and less than 1, the mean velocity of the fluid within the open channel or partially filled duct is equal to the mean surface velocity of the fluid over the zone of width L2 multiplied by a correction factor that depends on the standard deviation σ of the Gaussian curve;
• and
• wherein, in a third case, the ratio L2/L1 is less than ⅔, the mean velocity of the fluid within the open channel or partially filled duct is calculated based on the distribution of velocities measured over the zone of width L2 extrapolated for the missing velocities based on prior recordings of the velocity distribution when said ratio was greater than or equal to ⅔;
• and
• storage for forming a databank from the velocity measurements used to determine the mean velocity of the fluid within the open channel or partially filled duct when the velocity distribution over the width L2 is not representative of the velocity distribution over the width L1;
• the processor's calculation of the mean velocity having increased efficiency and precision through performance of fewer calculations from the measurement of the surface velocity using the reduced pulsed zone of width L2.

A method for determining with a processor free of finite element or finite difference modeling and with reduced calibrations a mean velocity of a fluid flowing in an open channel or a partially filled duct, said fluid having a free surface of width L1 extending between the walls of the channel or duct, said method comprising the following steps:

• performing local velocity measurements at the surface of the fluid over a zone of width L2, by generating a predetermined number of pulses and directing the pulses at a known angle toward the zone of width L2 to reflect the pulses by the surface of the flowing fluid;
• the set of local measurements generating a signal in the form of a spectrum of discrete data expressed in a temporal domain, the signals having a frequency shift proportional to velocity of particles at the surface of the fluid;
• converting the spectrum of discrete data expressed in the temporal domain into a spectrum of data expressed in a frequency domain via a Fourier transform;
• fitting of a Gaussian curve to the spectrum of discrete data expressed in the frequency domain and calculating the mean μ and the standard deviation σ of the Gaussian curve;
• the frequency of the mean μ enabling calculating mean surface velocity over the zone of width L2 and the standard deviation σ enabling calculating velocity distribution over the same zone of width L2;
• calculating a ratio L2/L1 between the width L2 of the zone impacted during the measurements and the width L1 of the free surface of the fluid to determine whether one of three cases is met;
• the processor being self-learning by detecting conditions under which the spectrum of discrete data obtained is recordable, and determining the mean velocity of the fluid within the open channel or the partially filled duct free of finite element or finite difference modeling, the mean velocity based on the ratio L2/L1 along with the mean μ and the standard deviation σ of the Gaussian curve, the processor calculating the mean velocity with increased efficiency and precision using fewer calculations from the measurement of the surface velocity using the reduced pulsed zone of width L2;
• wherein, in a first case, L1 is equal to L2, the mean velocity of the fluid within the open channel or partially filled duct is equal to the mean surface velocity of the fluid over the zone of width L2;
• wherein, in a second case, the ratio L2/L1 is greater than or equal to ⅔ and less than 1, the mean velocity of the fluid within the open channel or partially filled duct is equal to the mean surface velocity of the fluid over the zone of width L2 multiplied by a correction factor that depends on the standard deviation a of the Gaussian curve;
• and
• wherein, in a third case, the ratio L2/L1 is less than ⅔, the mean velocity of the fluid within the open channel or partially filled duct is calculated based on the distribution of velocities measured over the zone of width L2 extrapolated for the missing velocities based on prior recordings of the velocity distribution when said ratio was greater than or equal to ⅔.