CN111504366B - Artificial intelligence-based accurate metering method and metering device for fluid conveying system - Google Patents
Artificial intelligence-based accurate metering method and metering device for fluid conveying system Download PDFInfo
- Publication number
- CN111504366B CN111504366B CN202010208203.5A CN202010208203A CN111504366B CN 111504366 B CN111504366 B CN 111504366B CN 202010208203 A CN202010208203 A CN 202010208203A CN 111504366 B CN111504366 B CN 111504366B
- Authority
- CN
- China
- Prior art keywords
- data
- metering
- model
- value
- flow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01F—MEASURING VOLUME, VOLUME FLOW, MASS FLOW OR LIQUID LEVEL; METERING BY VOLUME
- G01F1/00—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow
- G01F1/05—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow by using mechanical effects
- G01F1/34—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow by using mechanical effects by measuring pressure or differential pressure
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01F—MEASURING VOLUME, VOLUME FLOW, MASS FLOW OR LIQUID LEVEL; METERING BY VOLUME
- G01F1/00—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow
- G01F1/56—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow by using electric or magnetic effects
- G01F1/58—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow by using electric or magnetic effects by electromagnetic flowmeters
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01F—MEASURING VOLUME, VOLUME FLOW, MASS FLOW OR LIQUID LEVEL; METERING BY VOLUME
- G01F1/00—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow
- G01F1/66—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow by measuring frequency, phase shift or propagation time of electromagnetic or other waves, e.g. using ultrasonic flowmeters
Abstract
The invention relates to an artificial intelligence-based accurate metering method and a metering device for a fluid conveying system, wherein the metering method comprises the following steps: the method comprises the following steps: installing a metering device capable of measuring flow Q, pressure P, temperature T, vibration V and noise I in a fluid system to obtain data of five basic parameters of flow Q, pressure P, temperature T, vibration V and noise I as metering model data; step two: carrying out graphical processing on the metering model data; step three: setting the value of the measured data of the subsequent measurement as Dt ═ (Qt Pt Tt Vt At), and comparing the value with the metering model to verify so as to determine whether the value result is correct; if the value is correct, directly outputting data; if the value is abnormal, the number of error data in the value needs to be obtained, and the data is output after the parameter of the abnormal value is adjusted. The invention realizes a miniature artificial intelligence, can accurately measure and check the parameters of the fluid system on line, improves the operation safety of the fluid system, perfects the measuring means and reduces the comprehensive cost.
Description
Technical Field
The invention relates to the field of fluid conveying systems, in particular to an artificial intelligence-based accurate metering method and a metering device for a fluid conveying system.
Background
In terms of production and life, pipeline transportation of fluid media is applied, and the fluid media refer to liquid, gas or gas-liquid two-phase media. In practical engineering, the following data are measured for the fluid conveying system: flow (Q), pressure (P), temperature (T), vibration (V), noise (I) five main categories basic parameter to ensure to carry accurate management of medium.
The five kinds of data are measured by five kinds of sensors, which are respectively: flow sensor, pressure sensor, temperature sensor, vibration sensor, noise sensor. The five sensors are usually installed on the conveying pipeline of the fluid system respectively, and the following problems exist in the practical engineering:
firstly, the installation of flow, pressure and temperature sensors usually needs to open holes on a conveying pipeline to enable a conveying medium to be in direct contact with the sensors so as to obtain sensing signals, and the open holes are called as source taking holes. The source taking hole can actually bring potential safety hazards to the fluid system, for example, medium leakage caused by burst of a welding seam at the position of the source taking hole and the sensor connecting pipe fitting is caused. However, the number of source extraction holes of the conventional conveying pipeline is too large, and each sensor needs to correspond to one source extraction hole.
Secondly, vibration and noise sensing signals are easily influenced by the external environment, and users often cannot read actual values of parameters of the fluid system, so that the users do not select to install the two sensors in many times. However, these two types of sensors are significant for determining and detecting a fluid system failure, such as: the abnormality of vibration reflects the possibility of mechanical failure of system components, the abnormality of noise reflects the possibility of medium leakage of system components, and the like.
The five sensors need to be checked after being used for a period of time, and the reasons are manifold and are two main reasons after being combined: the first is the reason of the accuracy of the sensor; the second is external environmental factors. For example, most sensors are checked after being used for one year according to the requirements of petrochemical industry standard SY/T6069-2005 oil and gas pipeline instrument and automatic system operation technical specification. However, in an actual process, a user often cannot find metering faults in time or check professional personnel and equipment places, so that the phenomenon that metering work is generally out of place exists, and passivity is brought to fluid management systems. Furthermore, we can consider that: the parameters of the sensors are correct within one year, faults can occur after one year, the faults are random, and the accuracy of quantitative fixed-point analysis data cannot be achieved in the prior art, so that whether the sensors have problems or not is judged.
And the prior artificial intelligence needs to utilize a cloud terminal to perform machine learning, so that the established data center has high energy consumption and causes environmental pollution.
At present, when fluid medium passes through the whole system from a pump or a high-water flow, the previous engineering experience is to determine the power of the required pump or the length of the conveying distance by calculating the water head loss and the pipeline pressure of each part of the system, but when the motor frequency of the pump changes or the opening of a valve in a pipeline changes, the conveying flow changes, and at the moment, the water head loss and the conveying flow of each part of the fluid system are different before changing, but no corresponding instrument is used for metering.
Disclosure of Invention
Aiming at the problems, the invention provides an artificial intelligence-based accurate metering method for a fluid conveying system, which realizes miniature artificial intelligence, adopts a matrix difference empowerment management method and develops an artificial intelligence algorithm for discriminating wrong data, is simple and reliable, ensures that the data can realize the evolution of services and functions without leaving equipment, and simultaneously ensures the privacy of user data.
The technical scheme adopted by the invention for solving the technical problems is as follows: an artificial intelligence-based accurate metering method for a fluid delivery system comprises the following steps:
the method comprises the following steps: installing a metering device capable of measuring flow Q, pressure P, temperature T, vibration V and noise I in a fluid system to obtain data of five basic parameters of flow Q, pressure P, temperature T, vibration V and noise I as metering model data;
step two: carrying out graphical processing on the metering model data: the method comprises the following steps:
2.1 metrology model learning phase:
the method comprises the following steps of operating a fluid system at a maximum flow point and a minimum flow point, and then obtaining five parameters of n points according to daily operation of the fluid system, wherein the n points comprise the maximum point and the minimum point;
2.2 metrology model reduction stage:
normalizing and sorting 5n parameters of n points in the metering model to form a 5 x n matrix;
2.3, dimension reduction treatment of the metering model:
and (3) performing dimensionality reduction on the 5 × n matrix in the step 2.2, forming 10 two-dimensional graphs at each point, and connecting the n points on the two-dimensional graphs by using curves to form a two-dimensional metering model, so that: the upper layer of the model is a classification array of n matrixes, and the lower layer of the model is a regression model of a continuous curve;
step three: setting the value of the measured data of the subsequent measurement as Dt ═ (Qt Pt Tt Vt At), and comparing the value with the metering model to verify so as to determine whether the value result is correct;
if the value is correct, directly outputting data;
if the value is abnormal, the number of error data in the value is required to be obtained, and the data is output after the parameter of the abnormal value is adjusted.
Preferably, the third step comprises the following specific steps:
3.1 the measured data integration stage: normalizing Dt to (Qt Pt Tt Vt At) and forming a homotype matrix with the metrology model;
3.2 identification stage of measured data: calculating the difference between the actually measured data and the model data of the metering model to obtain the sum of absolute values of the difference between the two matrix single-row data, and performing weighting management on the five types of data to obtain the minimum value minDk of the sum of absolute values of the difference between the two matrix single-row data;
3.3 allowable range learning of deviation: the allowable range of learned minDk is K;
if minDk ≦ K, it may be determined that the result of Dt ≦ K measured At this time (Qt Pt Tt Vt At) falls on the metrology model and is accurate, and thus data may be directly output;
if minDk > K, it can be determined that the result of Dt ═ Qt Tt Vt At measured At this time is not on the metrology model, and thus there is an anomaly, and the number of error data needs to be found At this time;
3.4 abnormal value of the key parameter: and (4) adjusting parameters of the error data by adopting correct data based on the metering model, and outputting the data after the adjustment is finished.
Preferably, the metering method can derive the instantaneous heat energy value J from Cmt as the fluid medium passes, where C is the specific heat of the medium, m is the mass flow rate per unit time, and t is the temperature.
Preferably, the metering method is based on
And obtaining an instantaneous kinetic energy value E, wherein m is the mass flow rate in unit time, S is the flow area of the metering device, and Q is the flow rate.
The invention also provides a metering device capable of realizing the method, the metering device is communicated on a fluid system pipeline through a connecting device, the metering device comprises a measuring pipe and a shell sound insulation layer arranged outside the measuring pipe, the measuring pipe and the shell sound insulation layer are arranged in a vacuum mode, a pressure sensor used for measuring pressure, a temperature sensor used for measuring temperature, a vibration sensor used for measuring vibration, a sound intensity sensor used for measuring noise and a flow sensor used for measuring flow are arranged in the vacuum mode, the metering device further comprises a comprehensive processing device arranged on the wall of the measuring pipe, and the comprehensive processing device is used for carrying out data processing and output display on the collected pressure, temperature, vibration, noise and flow signals.
Preferably, the pressure sensor and the temperature sensor are in fluid communication with the measurement tube through the same source aperture.
Preferably, the outer end of the source taking hole is provided with a three-way valve used for connecting the pressure sensor and the temperature sensor.
Preferably, the vibration sensor and the sound intensity sensor are arranged outside the measuring pipe wall.
Preferably, the flow sensor may be an electromagnetic flow meter, a differential pressure flow meter, or an ultrasonic flow meter.
Compared with the prior art, the invention has the following beneficial effects:
1.1 size and cost optimization of the metering device:
the metering device of the invention centralizes a plurality of sensors on the same metering device, greatly reduces the installation size and cost of the metering device by adopting an integration means, and solves the technical problem of overlarge cost caused by arranging each sensor at different positions of each pipeline in the prior art;
secondly, the technical scheme of the invention utilizes the fluid mechanics principle to ensure that the whole metering of the fluid system is more perfect and the cost is lower.
1.2, improving the operation safety of a fluid system:
the electromagnetic flowmeter is used, and the three-way valve is adopted, so that the same source taking hole is connected with two sensors, the number of the source taking holes is reduced as much as possible, and the operation safety of a fluid system is improved.
1.3 reducing the checking times of the metering instrument (sensor), the metering precision is obviously improved:
according to the invention, data are mined in real time by an artificial intelligence method, data faults are found in time, and parameters are automatically compensated, so that on-line automatic checking of the instrument is realized, manual intervention is not required, checking work is reduced, and meanwhile, the metering precision is greatly improved. The problem of prior art actual process, the user often can not discover in time measurement trouble or check the inaccurate technique of data that needs professional personnel and equipment place to lead to is solved.
1.4 adding the function of a' kinetic energy meter
When the fluid medium passes through the metering device, the instantaneous kinetic energy and the accumulated kinetic energy can be obtained. The invention can calculate the kinetic energy of the medium at each position of the fluid system according to the principle of energy conservation, and know the kinetic energy distribution condition of the conveying medium, thereby having extremely important practical application value in the management (energy conservation) of the fluid system.
1.5 realize a miniature artificial intelligence
The upgrading of the service and the function can be realized without leaving the equipment, a cloud database is not needed, the high energy consumption and the environmental pollution of a large database are avoided, and the data privacy of a user is protected.
Drawings
FIG. 1 is a schematic view of the delivery principle of the fluid system of the present invention;
FIG. 2 is an enlarged schematic view of the metering device of the present invention;
FIG. 3 is a partial cross-sectional view of FIG. 2;
FIG. 4 is a cross-sectional view taken along A-A of FIG. 2;
FIG. 5 is a two-dimensional model of a fluid system when n is 8;
FIG. 6 is a schematic diagram showing the comparison of test points and model points in the present invention;
FIG. 7 is a diagram of a metrology model if minDk is within an allowed range K in the present invention;
FIG. 8 is a diagram of a decision tree with node (Q) as the root node.
Detailed Description
The present invention will be described in detail with reference to fig. 1 to 8, and the exemplary embodiments and descriptions of the present invention are provided herein to explain the present invention but not to limit the present invention.
The flow Q of the invention refers to the volume flow, and in the same fluid system, hardware or flow components can not be changed after a set of fluid conveying system is established and operated. The flow passage component refers to the surface of an object which is in contact with a conveying medium. The invention is applicable to open fluid systems, namely: the conveying medium is in contact with the external environment, and the pressure of the external environment is not changed. The fluid conveying system is called a fluid system for short. Such fluid systems comprise a very large part, for example: tap water systems, industrial cooling water systems, fresh air systems, boiler flue gas systems and the like. In actual engineering, due to design, installation, and characteristics of the conveying medium and external environmental factors, it is difficult to manually calculate the exact process and solution of the equation. The data measured by the sensor needs to be imaged by a computer in combination with manual identification.
The invention provides a metering device which can be used for a fluid system, as shown in figure 1, wherein arrows indicate the flowing direction of fluid, the fluid system comprises a water pool 10, a water supplementing mechanism 11 for supplying water to the water pool, a water pump 12 and a fluid system pipeline, the water pump is used for pumping the fluid in the water pool to the fluid system pipeline, the metering device 14 is communicated on the fluid system pipeline through a connecting device, and the fluid in the fluid system pipeline flows to water using equipment through the metering device. The connecting device can adopt a flange or a clamp and the like. The metering device comprises a measuring pipe and a shell sound insulation layer arranged outside the measuring pipe, the measuring pipe and the shell sound insulation layer are arranged in a vacuum mode, the shell sound insulation layer can be installed in a welding mode, a pressure sensor for measuring pressure, a temperature sensor for measuring temperature, a vibration sensor for measuring vibration, a sound intensity sensor for measuring noise and a flow sensor for measuring flow are arranged in the vacuum mode, and the working principle of the sound intensity sensor is based on the fact that sound waves are transmitted in fluid or solid; the device is characterized by further comprising a comprehensive processing device arranged on the wall of the measuring pipe, wherein the comprehensive processing device is used for carrying out data processing and output display on the collected pressure, temperature, vibration, noise and flow signals. The comprehensive processing device is a set of microcomputer data and graphic processing system, and can realize the communication among different devices with an industrial communication card, and comprises a central processing chip, a data acquisition system, a display screen and the like, and the specific software and hardware of the comprehensive processing device can adopt the devices in the prior art, and the detailed description is omitted.
The shell puigging and measure the intertube for vacuum environment, play the noise proof effect to noise sensor, protected other four kinds of sensors simultaneously again, improved the measurement accuracy and the life of all sensors. Vibration sensor and sound intensity sensor locate outside measuring the pipe wall, shell puigging and measuring intertube evacuation, and locate vibration sensor and sound intensity sensor in the vacuum environment, make each sensor operation not receive external disturbance, make the operation more stable, the precision is high.
And the pressure sensor and the temperature sensor are communicated with the fluid in the measuring pipe through the same source taking hole. And a three-way valve is arranged at the outer end of the source taking hole and is used for connecting the pressure sensor and the temperature sensor. The three-way valve enables the two sensors to share one source taking hole, reduces the number of the source taking holes and improves the operation safety of a fluid system. And a cut-off device is connected between the source taking hole and the sensor, so that the fluid system can adjust and check the two sensors when the production is stopped.
The flow sensor can be an electromagnetic flow meter, but is not limited to the electromagnetic flow meter, and includes all flow meters utilizing mechanical, electrical and acoustic measurement principles, such as differential pressure flow meters, ultrasonic flow meters and the like. The electromagnetic flowmeter comprises an excitation coil 71, an iron core 7 and an electrode 73, wherein the excitation coil and the iron core are both positioned in vacuum of the measuring pipe and a shell sound insulation layer, and the electrode is inserted into the measuring pipe. The metering device further comprises an antenna 74 for communication with the outside world and an antenna protective housing 75 provided at the antenna.
Before stating the metering method, the relationship among five parameters of flow, pressure, temperature, vibration and noise is discussed, and the five parameters are directly or indirectly linearly related in the same fluid system, and the discussion is as follows:
the relationship between flow (Q) and pressure (P) (law of conservation of energy and flow equation):
in the above formula:
p-pressure;
h-head height;
ρ -media density;
g-acceleration of gravity;
v-flow rate;
q-flow rate;
s-pipe flow area.
It is known that when the flow Q changes, the flow velocity v also changes, i.e. the pressure P changes, and vice versa.
It needs to be further explained that:
the fluid systems described herein are all open fluid systems, and when a medium is in contact with an external constant pressure environment, a pressure difference is formed between a medium pressure measurement position and the external environment pressure, and the external environment pressure is constant, so that the change of the pressure at a measurement point is the change of the pressure difference. Of course, closed fluid systems may also be used. For example, a central air-conditioning cold and warm water system is a typical closed fluid system, but the relative factors of the requirement for keeping the water replenishing pressure constant are complicated, so that the discussion is omitted.
B flow (Q) versus vibration (V) (differential equation of motion of the piping system):
for reasons of the fluidic system itself: natural frequency, pressure pulsation, medium vibration (turbulence), etc., the relationship of flow rate to vibration can be described graphically and regression equations established. The link between the following can be seen by the differential equation of motion of the piping system:
[M]X+[C]Y+[K]Z＝F
the formula is as follows:
[ M ] -total mass matrix;
[C] -a damping matrix;
[K] -a total stiffness matrix;
x-velocity vector;
y-acceleration vector;
z-displacement vector;
f-disturbance force vector.
When the flow rate changes, [ M ] changes, causing a change in the vibration.
C is the relationship between flow (Q) and temperature (T) (molar gas constant and convective heat transfer equation):
the relationship between flow and temperature needs to be classified and explained as follows:
when the medium is gas, according to a mole gas constant equation: PB ═ RT.
In the formula:
p-pressure;
b-volume of gas; (Note: the volume is denoted by V in the standard equation, but vibration is denoted by V herein, so B here denotes the gas volume)
R-a constant;
t-temperature.
When the medium is liquid or gas, according to a convection heat transfer equation: h iscSΔt1＝CmΔt2。
In the formula:
hc-surface heat transfer coefficient;
s-flow component area;
Δt1-temperature difference between the fluid medium and the flow-through component;
c-specific heat of the fluid medium;
m-mass flow rate of the fluid medium;
Δt2-the amount of change in the temperature of the fluid medium.
The whole fluid system can be regarded as a heat exchange device under a certain environment, and when one set of fluid system operates at any time, the following conditions can be considered: h isc、S、Δt1Since C is unchanged, the temperature of the fluid medium changes when the flow rate changes.
D, relation (sound pressure formula) of flow (Q) and noise (I):
when the medium flows through the pipeline, noise is generated due to vibration of mass points. When a set of fluid system is fixed (the device characteristics are not changed), if the flow in the pipeline is changed, the flow field is changed, the velocity amplitude of particle vibration is also changed, and the sound pressure is changed at the same time, according to the formula:
in the formula:
vZ-velocity amplitude of particle vibration;
PI-a sound pressure;
ρ -the density of the acoustic object;
c-speed of sound.
It can be seen that the sound velocity is also a fixed value when the density of the sound-transmitting object is constant. The velocity of particle vibration is linear with sound pressure.
In summary, five parameters of the fluid system are flow (Q), pressure (P), temperature (T), vibration (V), and noise (I) related to each other. We can consider these five parameters as five characteristics of the fluid system, and when one characteristic changes, the other characteristics will be functionally related to this changed characteristic, namely:
in actual engineering, due to design, installation, and characteristics of the conveying medium and external environmental factors, it is difficult to manually calculate the exact process and solution of the equation. The data measured by the sensor needs to be imaged by a computer in combination with manual identification.
The invention also provides a method for realizing accurate metering of the fluid conveying system based on artificial intelligence by using the metering device. The method comprises the following steps:
the method comprises the following steps: installing the metering device capable of measuring flow rate (Q), pressure (P), temperature (T), vibration (V) and noise (I) in a fluid system to obtain data of five basic parameters of flow rate (Q), pressure (P), temperature (T), vibration (V) and noise (I) as metering model data;
step two: carrying out graphical processing on the metering model data: the method comprises the following steps:
2.1 metrology model learning phase:
at the early stage of the installation of the metering device of the invention on a fluid system, all sensors are accurate, this time being about one year or so, and the following models are established on the basis of the data measured by each sensor: flow, pressure, temperature, vibration, noise matrix model (Q-P-T-V-I)
Obtaining five parameters of n points according to daily operation of a fluid system, forming a column of a matrix once data is collected, and obtaining a 5 x n matrix if n different operation points are collected:
in order to perfect the model in the process, the fluid system is required to operate at a maximum flow point and a minimum flow point, wherein D1＝(Q1 P1 T1 V1 A1) Corresponding to a minimum flow point, Dn＝(Qn Pn Tn Vn An) The n points include a maximum point and a minimum point corresponding to the maximum flow point. This is easily done in practice because there is a range of flow rates for each set of fluid systems and the system needs to be tuned within this range. In the process, parameters which are greatly influenced by the external environment need to be manually marked or classified and predicted, and vector filling is carried out after the parameters are deleted. For example, if other data is accurate, it is easier to perform vector filling on only one type of data, and the filling of this data can also be based on other correct data of the same type。
2.2 metrology model reduction stage: 5n parameters of n points in the metering model are calculated according to a formula
Carrying out normalization arrangement to form a 5 xn matrix; where min and max represent the minimum and maximum values of the model data, respectively, so that the model data becomes a dimensionless number. (to make the description process less complicated, this dimensionless number matrix is still denoted by D.)
2.3, dimension reduction treatment of the metering model:
and (3) performing dimensionality reduction on the 5 × n matrix in the step 2.2, forming 10 two-dimensional graphs at each point, and connecting the n points on the two-dimensional graphs by using curves to form a two-dimensional metering model, so that: the upper layer of the model is a classification array of n matrixes, and the lower layer of the model is a regression model of a continuous curve;
the flow, pressure, temperature, vibration, and noise can form ten two-dimensional metering models (Q-P, Q-T, Q-V, Q-I, P-T, P-V, P-I, T-V, T-I, V-I). The five types of data can form a continuous functional relation, and meanwhile, according to the continuity principle of the fluid system, on the condition that a plurality of correct data with one characteristic are known, the data are connected into a smooth curve to obtain a two-dimensional metering model shown in the figure 5:
fig. 5 shows a two-dimensional metrology model for a fluid system when n is 8.
The above-mentioned process is the learning stage of the metering model, and needs to be completed under the condition that the sensor data can not have large-area errors, because the sensor has measurement deviation which is getting bigger and bigger after reaching a time node, and the data directly collected at this moment may have errors, so that the fluid system is mostly operated within one year in the actual operation at this time.
Step three: setting the value of the measured data of the subsequent measurement as Dt ═ Qt Pt Tt Vt At, comparing the value with the metering model to verify to judge whether the value result is correct,
if the value is correct, the data is directly output,
if the value is abnormal, the number of error data in the value is required to be obtained, and the data is output after the parameter of the abnormal value is adjusted.
After the metrology model learning is complete, the subsequently collected error data is categorized and compared less well with conventional statistical methods. Since there are five data (features) in one measurement, it is highly likely that the five data (features) fall on the two-dimensional metrology model curve respectively, but it is not clear which of the five data (features) are erroneous. At this time, the model needs to be retrained by adopting a machine learning method.
The third step comprises the following specific steps:
3.1 the measured data integration stage: let a set of measured data be Dt＝(Qt Pt Tt Vt At) Normalizing Dt ═ (Qt Pt Tt Vt At) and forming a homotypic matrix with the metrology model; normalization was performed in the same manner. Where min and max represent the minimum and maximum values of the model data, respectively, such that the model data becomes a dimensionless number (this dimensionless number matrix is still represented by Dt).
3.2 identification stage of measured data: calculating the difference between the actually measured data and the model data of the metering model to obtain the sum of absolute values of the difference between the two matrix single-row data, and performing weighting management on the five types of data to obtain the minimum value minDk of the sum of absolute values of the difference between the two matrix single-row data; the method specifically comprises the following steps: firstly, calculating the difference value of two dimensionless matrixes:
transformation of Dt into D isomorphism matrix DtzThen find D-Dtz：
The sum of the absolute values of the single-column matrices can be obtained:
Dk＝|Qk-Qt|+|Pk-Pt|+|Tk-Tt|+|Vk-Vt|+|Ik-It|(k＝1,2,3,...,n)
in this case, D must be presentkFor the minimum value of (2), then for the dimensionless matrix minDkEach feature is as followsPerforming entitlement processing to obtain:
minDk＝α|Qk-Qt|+β|Pk-Pt|+γ|Tk-Tt|+λ|Vk-Vt|+μ|Ik-It(the weighted values α, β, γ, λ, μ represent information gains of different data, respectively, and then see calculation methods). Special cases, when | Qk-Qt|＝|Pk-Pt|＝|Tk-Tt|＝|Vk-Vt|＝|Ik-ItWhen | ═ 0, i.e., this measurement point falls completely on the metrology model, all five measured data (features) are correct.
3.3 allowable range learning of deviation: the allowable range of learned minDk is K;
if minDk ≦ K, it may be determined that the result of Dt ≦ K measured At this time (Qt Pt Tt Vt At) falls on the metrology model and is accurate, and thus data may be directly output;
if minDk is within the allowed range K, then the test point values are also allowed. The metrology model can now be visualized as a map 7:
minDk represents the deviation of the measured data and the matrix model, and if the deviation is within the allowed range K, the measured data falls into the two-dimensional curve metering model layer and corresponds to each point on the two-dimensional curve metering model. The K value does not represent the distance between the two matrices and needs to be trained on different fluid system characteristics.
If minDk > K, it can be determined that the result of the measured Dt ═ is not on the metrology model (Qt Pt Tt Vt At), and therefore there is an anomaly, and an artificial intelligence algorithm is needed to find out the number of error data;
3.4 abnormal value of the key parameter: and (4) adjusting parameters of the error data by adopting correct data based on the metering model, and outputting the data after the adjustment is finished.
If the obtained minDkIf the deviation is not within the allowable range, the deviation is too large. The types of data with errors are five or less, and the types are shared
And (4) carrying out the following steps. Based on the possibility, an algorithm model is constructed by adopting a machine learning method (the data is subjected to normalized dimensionless processing in model expression):
inputting: training set M { (Q)1,P1,T1,V1,I1),(Q2,P2,T2,V2,I2),...,(Qm,Pm,Tm,Vm,Im)}；
Attribute set
(the algorithm is intended to find erroneous metering data, for
Measurement error value representing attribute X)
The process is as follows: function TreeGenerator (M, omega)
Generating nodes
(Q)∪(P,T,V,I)∈M
2:if minDk(P,T,V,I)≤k1(ii) a (k1 represents the measured value after Q is removed and the allowable value of the four-dimensional vector of the measurement model P-T-V-I, and similar processes are not described below)
3 marking note as C1Class nodes;
4：end if
5:if minDk(P,T,V,I)＞k1；
6 generating a branch for note
7：end if
8：if minDk(Q,T,V,I)≤k2；
9 marking note as C2Class nodes;
10:end if
11:if minDk(Q,T,V,I)＞k2
12-generate a branch for note
13：end if
14:if minDk(Q,P,V,I)≤k3；
15 marking note as C3Class nodes;
16:end if
17：if minDk(Q,P,V,I)＞k3
18: generating a branch for note
19:end if
20:if minDk(Q,P,T,I)≤k4；
21 note of note C4Class nodes;
22:end if
23:if minDk(Q,P,T,I)＞k4
24 generating a branch for note
25:end if
26:if minDk(Q,P,T,V)≤k5；
27 note C5Class nodes;
28:end if
29:if minDk(Q,P,T,V)＞k5
30-generate a branch for note
31:end if
32:if minDk(T,V,I)≤k6(ii) a (k6 represents the measured value after Q, P are removed and the allowable value of T-V-I three-dimensional vector of the metrology modelIn the following, similar processes are not described one by one
33 marking note as C6Class nodes;
34:end if
35:if minDk(T,V,I)＞k6
36-generate a branch for note
37:end if
38 (two of the five attributes are respectively taken out, the circulation is similar to the process of 32-37, and eight branches are also arranged in the process.)
39: selecting optimal partition attributes from omega
40:forω*Each value of (1)
Empowerment
41, generating a branch for note; let MvDenotes the value in M at ω*Up value of
A subset of samples of (a);
42:end for
and (3) outputting: with a node
A decision tree for the root node can be represented by fig. 8:
the above step 8 is based on the attribute
Information gain principle:
c ═ 1,2 indicates the possibility of some sensor measurement (1 ═ correct, 2 ═ wrong); it can be derived that:
Gain(M,ω)＝Ent(M)-Ent(MC)
in the information entropy formula: assume the probability p of occurrence of class t sample measurement in sample Mt(t ═ 1,2, ·, | y |); the algorithm just judges the result as being correct, and obviously, y is 2 (one possibility is correct and the other possibility is incorrect). And calculating five parameters of the flow (Q), the pressure (P), the temperature (T), the vibration (V) and the noise (I) of the fluid system according to the information entropy and the information gain. The smaller the entropy of information, the higher the purity of the sample. The greater the information gain, the greater the "purity improvement" resulting from using the attribute ω for the partitioning.
The combination of the above shows that:
when an error occurs in one of the sensors, the error rate in the sample is 0.2, the accuracy rate is 0.8, and end (m) is 0.7216.
When two sensors have errors, the error rate in the sample is 0.4, the accuracy rate is 0.6, and end (m) is 0.9702.
We can do the attribution of the attributes when the two cases happen. (of course, when four types of sensors have errors, there is also Ent (M) ═ 0.7216 in the case of five sensors, but in practice, only one sensor can be damaged and then four sensors can be damaged, and no case exists that four sensors are damaged and then one sensor is damaged)
From the above algorithm 1 to 29, we can find out all cases where data is wrong. This is also the most true in practice: after one sensor fails, the other sensors fail one by one. But we then correct the parameters and the situation is restored to before it is bad.
From the above algorithm 30 th to 38 th steps, we can find out all cases where both data are in error at the same time. The possibility of simultaneous failure of the three sensors is too small to be considered in this contextConsideration is made. According to the formula: minDk＝α|Qk-Qt|+β|Pk-Pt|+γ|Tk-Tt|+λ|Vk-Vt|+μ|Ik-ItWhen | performing ownership on different attributes, one practical situation needs to be considered: if the vibration (V) and noise (I) sensors are abnormal all the time, the possibility of system mechanical failure or pipeline leakage exists, and the failure point needs to be eliminated and then confirmed. According to the formula: gain (M, ω) ═ Ent (M) -Ent (M)C) Information gain values with different attributes can be obtained, and the values are normalized to make alpha + beta + gamma + lambda + mu equal to 1, so as to obtain five weighting values of alpha, beta, gamma, lambda and mu. This value can be intuitively understood as: depending on the system, the more stable the measurement of which attribute parameter (lower error rate), the greater the weighting value.
And (3) adjusting abnormal values:
the number or the type of the abnormal data can be obtained through the steps. And in the generated two-dimensional metering model, adjusting the coordinates of the parameter of the error data according to the coordinates of the correct data until the coordinates are accurate. As shown in fig. 6, such as in the metrology model (Q-P): the model point is (Q)m,Pm) The test point is (Q)t,Pm) Then: with PmAdjusting the flow rate (Q) on a reference basis such that Q ism＝ζQtζ is an adjustment parameter (ζ not only represents a value but may also be a function).
In the invention, when the fluid medium passes through, the metering method can obtain an instantaneous heat energy value J according to Cmt, wherein C is the specific heat of the medium, m is the mass flow rate per unit time, and t is the temperature.
The metering method may be based on when the fluid medium passes
And obtaining an instantaneous kinetic energy value E, wherein m is the mass flow rate in unit time, S is the flow area of the metering device, and Q is the flow rate.
According to the output:
the data of the invention are output as follows:
the invention comprises a comprehensive detection device and a comprehensive solution of a set of fluid system, and parameter setting and model adjustment are required to be carried out on different fluid systems. The method is an integrated and artificial intelligence measure of the fluid system metering, and mainly can achieve the following technical effects:
the equipment installation size is reduced, and the equipment installation and manufacturing cost is reduced.
And secondly, the number of source taking holes is reduced, and the operation safety of the system is improved.
The metering precision is improved, the automatic adjustment of the parameters of the meter is realized according to different characteristics of various meters and the characteristics of a fluid system, and the production efficiency is improved. The improvement of the metering precision is mainly reflected in two aspects: 1. The data collection and analysis may be performed in ten minutes, whereby possible fault values will also be retrieved once in ten minutes. 52560 times can be carried out continuously in one year, and compared with the traditional manual one-year-one-time meter calibration, the meter calibration is improved by several orders of magnitude.
2. In the artificial intelligence algorithm, there is a minimal error in the elimination of erroneous data: for example, in step 1 to step 29, all the possibilities of data error are excluded, and one of the factors of error is: it is possible that this excluded data is correct and the other four are erroneous. According to the "matrix difference value empowerment management" method, all four points should fall on the four-dimensional metrology model. Then, the probability that the four error data may fall on the four-dimensional metrology model at the same time is (set K to 3):
52560 times are measured continuously for one year, about 0.81 times of errors may occur in 19 years. K3 may be considered reasonable because the error of all sensors may be less than 3％。
The new functionality of the kinetic energy meter is added, the energy distribution situation of the fluid system is clear, the energy consumption situation is judged through the pressure loss of the pipeline in the past industry, but the result can be directly read through the kinetic energy meter, so that the energy distribution situation of the fluid system is clear at a glance, and a basis is provided for scientific management of the system.
The accurate five parameters of a certain position of the system are obtained, the device can be arranged on other parts of the system, the flow measurement is only carried out on the main pipe according to the energy conservation and the fluid continuity principle, and when the device is arranged on other parts of the system, all five data of the position of the device can be obtained without arranging a flow meter.
The technical solutions provided by the embodiments of the present invention are described in detail above, and the principles and embodiments of the present invention are explained herein by using specific examples, and the descriptions of the embodiments are only used to help understanding the principles of the embodiments of the present invention; meanwhile, for a person skilled in the art, according to the embodiments of the present invention, there may be variations in the specific implementation manners and application ranges, and in summary, the content of the present description should not be construed as a limitation to the present invention.
Claims (8)
1. An artificial intelligence-based accurate metering method for a fluid delivery system is characterized by comprising the following steps: the method comprises the following steps:
the method comprises the following steps: installing a metering device capable of measuring flow rate (Q), pressure (P), temperature (T), vibration (V) and noise (I) in a fluid system to obtain data of five basic parameters of flow rate (Q), pressure (P), temperature (T), vibration (V) and noise (I) as metering model data;
step two: carrying out graphical processing on the metering model data: the method comprises the following steps:
2.1 metrology model learning phase:
the method comprises the following steps of operating a fluid system at a maximum flow point and a minimum flow point, and then obtaining five parameters of n points according to daily operation of the fluid system, wherein the n points comprise the maximum point and the minimum point;
2.2 metrology model reduction stage:
normalizing and sorting 5n parameters of n points in the metering model to form a 5 x n matrix;
2.3, dimension reduction treatment of the metering model:
and (3) performing dimensionality reduction on the 5 × n matrix in the step 2.2, forming 10 two-dimensional graphs at each point, and connecting the n points on the two-dimensional graphs by using curves to form a two-dimensional metering model, so that: the upper layer of the model is a classification array of n matrixes, and the lower layer of the model is a regression model of a continuous curve;
step three: setting the value of the measured data of the subsequent measurement as Dt ═ (Qt Pt Tt Vt At), and comparing the value with the metering model to verify so as to determine whether the value result is correct;
if the value is correct, directly outputting data;
if the value is abnormal, the number of error data in the value needs to be obtained, and the data is output after the parameter of the abnormal value is adjusted;
the third step comprises the following specific steps:
3.1 the measured data integration stage: normalizing Dt to (Qt Pt Tt Vt At) and forming a homotype matrix with the metrology model;
3.2 identification stage of measured data: calculating the difference between the actually measured data and the model data of the metering model to obtain the sum of absolute values of the difference between the two matrix single-row data, and performing weighting management on the five types of data to obtain the minimum value minDk of the sum of absolute values of the difference between the two matrix single-row data;
3.3 allowable range learning of deviation: the allowable range of learned minDk is K;
if minDk ≦ K, it may be determined that the result of Dt ≦ K measured At this time (Qt Pt Tt Vt At) falls on the metrology model and is accurate, and thus data may be directly output;
if minDk > K, it can be determined that the result of Dt ═ Qt Tt Vt At measured At this time is not on the metrology model, and thus there is an anomaly, and the number of error data needs to be found At this time;
3.4 abnormal value of the key parameter: and (4) adjusting parameters of the error data by adopting correct data based on the metering model, and outputting the data after the adjustment is finished.
2. The artificial intelligence based fluid delivery system precision metering method of claim 1, wherein: and in the third step, if the value is abnormal, obtaining the number of error data through the implementation of an artificial intelligence algorithm, and outputting the data after parameter adjustment of the abnormal value.
3. The artificial intelligence based fluid delivery system precision metering method of claim 1, wherein: the metering method may be based on when the fluid medium passes
And obtaining an instantaneous kinetic energy value E, wherein m is the mass flow rate in unit time, S is the flow area of the metering device, and Q is the flow rate.
4. A metering device using the metering method according to any one of claims 1 to 3, characterized in that: link up on fluid system pipeline through connecting device (1), metering device is including surveying buret (2) and locating outer shell puigging (3) of survey buret outside, set up for the vacuum between survey buret and the shell puigging, and be equipped with in the vacuum and be used for measuring pressure's pressure sensor, be used for measuring temperature sensor (4), be used for measuring vibration sensor (5), be used for measuring noise intensity sensor (6), be used for measuring flow's flow sensor (7), still including locating integrated processing apparatus (8) of measuring the pipe wall, integrated processing apparatus is used for carrying out data processing and output show to pressure, temperature, vibration, noise, the flow signal who gathers.
5. The metering device of the metering method according to claim 4, wherein: the pressure sensor and the temperature sensor are communicated with fluid in the measuring pipe through the same source taking hole (9).
6. The metering device of the metering method according to claim 5, wherein: and a three-way valve is arranged at the outer end of the source taking hole and is used for connecting the pressure sensor and the temperature sensor.
7. The metering device of the metering method according to claim 4, wherein: the vibration sensor and the sound intensity sensor are arranged outside the measuring pipe wall.
8. The metering device of the metering method according to claim 4, wherein: the flow sensor can adopt an electromagnetic flowmeter, a differential pressure flowmeter and an ultrasonic flowmeter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010208203.5A CN111504366B (en) | 2020-03-23 | 2020-03-23 | Artificial intelligence-based accurate metering method and metering device for fluid conveying system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010208203.5A CN111504366B (en) | 2020-03-23 | 2020-03-23 | Artificial intelligence-based accurate metering method and metering device for fluid conveying system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111504366A CN111504366A (en) | 2020-08-07 |
| CN111504366B true CN111504366B (en) | 2022-01-25 |
Family
ID=71874175
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010208203.5A Active CN111504366B (en) | 2020-03-23 | 2020-03-23 | Artificial intelligence-based accurate metering method and metering device for fluid conveying system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111504366B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112880778B (en) * | 2021-01-20 | 2021-10-15 | 中国水利水电科学研究院 | Measuring tank based detection device and measuring tank flow detection method |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103674326A (en) * | 2012-09-25 | 2014-03-26 | 爱知时计电机株式会社 | Temperature meter, flow meter and temperature metering method |
| CN105486363A (en) * | 2016-01-21 | 2016-04-13 | 成都声立德克技术有限公司 | Ultrasonic gas flowmeter and measuring method |
| CN106777622A (en) * | 2016-12-06 | 2017-05-31 | 山东瀚岳智能科技股份有限公司 | The method and system of the electromechanical equipment on-line fault diagnosis based on artificial intelligence |
| CN107064758A (en) * | 2017-04-18 | 2017-08-18 | 江西明正变电设备有限公司 | A kind of dry-type transformer on-Line Monitor Device and monitoring method |
| CN108168659A (en) * | 2018-03-21 | 2018-06-15 | 合肥科迈捷智能传感技术有限公司 | A kind of vortex-shedding meter intelligent diagnosing method |
| CN108805747A (en) * | 2018-06-13 | 2018-11-13 | 山东科技大学 | A kind of abnormal electricity consumption user detection method based on semi-supervised learning |
| CN108981796A (en) * | 2018-06-06 | 2018-12-11 | 江苏大学 | A kind of five in one hydraulic method for diagnosing faults |
| CN109130550A (en) * | 2018-08-21 | 2019-01-04 | 嘉兴学院 | A kind of artificial intelligent control system of electrohydrodynamics spray printing and control method |
| CN109545408A (en) * | 2017-09-19 | 2019-03-29 | 无锡物讯科技有限公司 | Judgement and prediction technique based on artificial intelligence technology to nuclear power plant reactor fuel element cladding breakage |
| CN110400231A (en) * | 2019-06-06 | 2019-11-01 | 湖南大学 | A kind of electric energy measuring equipment crash rate predictor method weighting Nonlinear Bayesian |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2010109644A (en) * | 2010-03-15 | 2011-09-20 | Владимир Анатольевич Иванов (RU) | OIL PRODUCING WELL WITH ARTIFICIAL INTELLIGENCE |
| CA2797794C (en) * | 2010-05-05 | 2019-01-15 | Greensleeves, LLC | Energy chassis and energy exchange device |
| US8834389B2 (en) * | 2011-11-25 | 2014-09-16 | Tepsync | Temperature based fertility monitoring system and related method |
| CN105807736A (en) * | 2014-12-27 | 2016-07-27 | 上海麦杰环境科技有限公司 | Pollutant intelligent accounting method |
| CN105808902B (en) * | 2014-12-27 | 2020-09-18 | 上海麦杰环境科技有限公司 | Qualitative method for analyzing operation condition of wet desulphurization system |
| CN108592165B (en) * | 2018-03-13 | 2019-11-19 | 沈阳久沃能源科技有限公司 | A kind of heat exchange station system optimal control method |
| IT201800005672A1 (en) * | 2018-05-24 | 2019-11-24 | SCALABLE BIOMETRIC ALARM SYSTEM WITH ARTIFICIAL INTELLIGENCE | |
| US11221353B2 (en) * | 2018-07-06 | 2022-01-11 | Schneider Electric USA, Inc. | Systems and methods for analyzing power quality events in an electrical system |
| CN110851957A (en) * | 2019-10-15 | 2020-02-28 | 南京航空航天大学 | Atmospheric data sensing system resolving method based on deep learning |
-
2020
- 2020-03-23 CN CN202010208203.5A patent/CN111504366B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103674326A (en) * | 2012-09-25 | 2014-03-26 | 爱知时计电机株式会社 | Temperature meter, flow meter and temperature metering method |
| CN105486363A (en) * | 2016-01-21 | 2016-04-13 | 成都声立德克技术有限公司 | Ultrasonic gas flowmeter and measuring method |
| CN106777622A (en) * | 2016-12-06 | 2017-05-31 | 山东瀚岳智能科技股份有限公司 | The method and system of the electromechanical equipment on-line fault diagnosis based on artificial intelligence |
| CN107064758A (en) * | 2017-04-18 | 2017-08-18 | 江西明正变电设备有限公司 | A kind of dry-type transformer on-Line Monitor Device and monitoring method |
| CN109545408A (en) * | 2017-09-19 | 2019-03-29 | 无锡物讯科技有限公司 | Judgement and prediction technique based on artificial intelligence technology to nuclear power plant reactor fuel element cladding breakage |
| CN108168659A (en) * | 2018-03-21 | 2018-06-15 | 合肥科迈捷智能传感技术有限公司 | A kind of vortex-shedding meter intelligent diagnosing method |
| CN108981796A (en) * | 2018-06-06 | 2018-12-11 | 江苏大学 | A kind of five in one hydraulic method for diagnosing faults |
| CN108805747A (en) * | 2018-06-13 | 2018-11-13 | 山东科技大学 | A kind of abnormal electricity consumption user detection method based on semi-supervised learning |
| CN109130550A (en) * | 2018-08-21 | 2019-01-04 | 嘉兴学院 | A kind of artificial intelligent control system of electrohydrodynamics spray printing and control method |
| CN110400231A (en) * | 2019-06-06 | 2019-11-01 | 湖南大学 | A kind of electric energy measuring equipment crash rate predictor method weighting Nonlinear Bayesian |
Non-Patent Citations (1)
| Title |
|---|
| 人工智能在流体传动中的应用;张立军 等;《人工智能在流体传动中的应用》;20090228;第37卷(第2期);第96-98页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111504366A (en) | 2020-08-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104685324A (en) | A flow meter system | |
| CN111684238B (en) | System and method for real-time discovery and resolution of wet gas venturi meter problems | |
| US20080127712A1 (en) | In-Situ Calibration Verification Device and Method for Electromagnetic Flowmeters | |
| CN105431715B (en) | Consumption meter with optional legal register | |
| CN113551744A (en) | Ultrasonic flowmeter performance online monitoring method and system | |
| AU2018437115B2 (en) | Detecting a change in a vibratory meter based on two baseline meter verifications | |
| CN111504366B (en) | Artificial intelligence-based accurate metering method and metering device for fluid conveying system | |
| US20200173836A1 (en) | Method and a system for configuring an electromagnetic flowmeter | |
| CN102147625B (en) | Detection and control device for density of regenerative acid and method | |
| CN115876288A (en) | Electronic instrument fault analysis method and system based on big data | |
| Hardy et al. | Flow measurement methods and applications | |
| CN107462305B (en) | A kind of flowmeter inline diagnosis method | |
| US20220205890A1 (en) | Device and method for fluid and equipment monitoring | |
| CN116685833A (en) | System and method for quantitative verification of flow measurements | |
| CN114165216A (en) | Single-well continuous multiphase flow metering system, storage medium and computer equipment | |
| CN114811448B (en) | Method for pipeline leakage detection, leakage flow velocity estimation and leakage positioning under flowing condition | |
| CN105841760A (en) | Insert-type electromagnetic flowmeter | |
| CN209025662U (en) | Three pressure value meters and system | |
| US3063287A (en) | Densitometer | |
| CN101813545A (en) | Precise locating method for leakage of pressure bearing pipe of three-dimensional quaternary array power station boiler | |
| CN213120667U (en) | Oil flow monitoring device | |
| Walski | Avoiding Pitfalls in Calibrating a Water Distribution System Model | |
| Khadrawy et al. | Towards high accuracy measurements for gas flow rate using ultrasonic devices in oil fields | |
| Konu et al. | Non-Intrusive Passive Acoustic Sensing for Accurate Gas Flow Measurement in Gas Pipelines | |
| CN112113622A (en) | Oil flow monitoring device and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |