Accurate prediction method for installation position of roadway wind speed sensor

Abstract

The invention belongs to the technical field of coal mine ventilation, and particularly relates to a roadway wind speed sensor mounting position accurate prediction method, which comprises the following steps: S1, analyzing a roadway wind flow state, S2, if the roadway wind flow state is a turbulent flow state, simulating a roadway section average wind speed circle to obtain a position from the average wind speed circle to the top of a roadway; s3, establishing a BP neural network prediction model, and outputting the position from the average wind speed circle to the top of the roadway; and S4, comparing the position, simulated in the step S2, from the average wind speed circle to the top of the roadway with the position, output by the BP neural network prediction model in the step S3, from the average wind speed circle to the top of the roadway, and if the error result is within 4%, taking the position, output by the BP neural network prediction model, from the average wind speed circle to the top of the roadway as the mounting position of a wind speed sensor. According to the method, the network prediction result and the simulation result are compared, if the error range is within 4%, the prediction result is used as the installation position of the wind speed sensor, and the installation accuracy of the wind speed sensor is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of coal mine ventilation, and in particular relates to a method for accurately predicting the installation position of a tunnel wind speed sensor. Background Art

[0002] Ventilation is the basis for ensuring safe and efficient production in coal mines. The precise measurement and control of underground air volume is an important basic component of intelligent mine ventilation management and control. The ability to effectively monitor underground air volume is one of the important conditions for the construction of intelligent mine ventilation. The traditional underground air volume measurement process is mainly through manual wind measurement and sensor wind measurement. The accuracy of manual wind measurement is greatly limited by the wind measurement level of the wind surveyor; the wind speed sensor can only monitor the wind speed at one point, and the relative error is large compared to the average wind speed in the tunnel.

[0003] Accurate monitoring of ventilation parameters such as tunnel wind speed, air volume, wind pressure, and temperature is an important prerequisite for real-time monitoring and calculation of mine ventilation networks. It is of great significance for regulating the operating conditions of major ventilators and saving electricity resources. At the same time, remote control of ventilation facilities during disasters to protect the lives and health of underground workers and reduce accident losses also requires accurate ventilation parameter support. "Accurate wind speed measurement" can be said to be an important foundation for building an intelligent mine ventilation system.

[0004] At present, there have been many studies on the wind flow distribution state of the tunnel cross section. For wind speed monitoring, on the one hand, numerical simulation software is mainly used to simulate the wind flow field of the tunnel underground, so as to find the average wind speed circle inside the tunnel, and sensors are arranged at the location of the average wind speed circle for wind speed monitoring; on the other hand, the wind speed of a point is monitored by a wind speed sensor, and then a wind speed compensation coefficient is given for correction to calculate the average wind speed of the tunnel. However, both of the above methods have certain inapplicability. First, the average wind speed circle of the tunnel is affected by the shape of the tunnel section, the cross-sectional area of ​​the tunnel, and the friction resistance coefficient of the tunnel. It is relatively time-consuming and laborious to use numerical simulation methods to find the average wind speed circle of each tunnel; secondly, since the wind speed compensation coefficient is generally determined by fitting the monitoring value and the artificial value, the wind speed compensation coefficients at different locations are different. Since the layout position of the wind speed sensor is relatively uncertain, the accuracy of the wind speed compensation correction coefficient is difficult to guarantee.

[0005] Based on this, it is particularly important to propose a method that can quickly and accurately determine the installation location of the tunnel wind speed sensor. Summary of the invention

[0006] The purpose of the present invention is to provide a method for accurately predicting the installation position of a tunnel wind speed sensor, so as to solve the technical problem that the traditional inaccurate wind speed sensor hanging position leads to inaccurate wind volume monitoring. In order to solve the above technical problems, the present invention provides a method for accurately predicting the installation position of a tunnel wind speed sensor, which performs a large amount of data simulation by means of numerical simulation to obtain the average wind speed circle position under different specifications and friction resistance coefficients, and then quickly predicts the position of the tunnel average wind speed circle by the BP neural network prediction method to guide the installation of the wind speed sensor and improve the accuracy of wind volume monitoring.

[0007] The present invention provides a method for accurately predicting the installation position of a tunnel wind speed sensor, which is characterized by comprising the following steps:

[0008] S1. Analyze the airflow state of the laneway and determine whether the airflow state of the laneway is laminar or turbulent;

[0009] S2. If the wind flow state of the tunnel is turbulent, numerical simulation is performed on the average wind speed circle of the tunnel section to obtain the distance between the average wind speed circle and the top of the tunnel;

[0010] S3, establish a BP neural network prediction model to output the position of the average wind speed circle from the top of the tunnel;

[0011] S4. Compare the position of the average wind speed circle simulated in step S2 from the top of the tunnel with the position of the average wind speed circle output by the BP neural network prediction model in step S3 from the top of the tunnel. If the error result is within 4%, the position of the average wind speed circle output by the BP neural network prediction model from the top of the tunnel is used as the installation position of the wind speed sensor.

[0012] Further, in step S1, the average velocity of the circular channel fluid, m / s:

[0013]

[0014] Average velocity of fluid in non-circular tunnel, m / s:

[0015]

[0016] Among them, R e is the Reynolds number; r 0 is the diameter of the circular pipe, m; α is the kinematic viscosity coefficient of the fluid, U represents the circumference of the tunnel section, and S represents the cross-sectional area of ​​the tunnel section.

[0017] Furthermore, the average velocity v of the fluid in the tunnel is obtained according to formula (1) and formula (2). If v is greater than 0.15 m / s, the wind flow state in the tunnel is determined to be a turbulent state.

[0018] Furthermore, in step S2, the following steps are also included:

[0019] S21. According to the actual situation of the lane, a three-dimensional coordinate model of the lane is established with the X axis as the lane length, the Y axis as the lane height, and the Z axis as the lane width;

[0020] S22. Assumptions a. The mine airflow is incompressible; b. The inlet airflow is in a stable state, satisfying the Boussinesq assumption; c. The tunnel wall is adiabatically constant temperature, ignoring the heat generated by the friction between the airflow and the tunnel wall;

[0021] S23, using the front model processing software CFD to establish a three-dimensional coordinate model, and through Boolean operations, meshing the model;

[0022] S24, use the pressure implicit separation solver, select absolute speed as the speed, select the standard two-equation model as the turbulence model, select the Simple algorithm as the flow field solution calculation method, and simulate the position of the average wind speed circle from the top of the tunnel.

[0023] Further, in step S3, the BP neural network model includes: an input layer, a hidden layer and an output layer, and the tunnel section type, tunnel section height, tunnel section width, and tunnel friction resistance coefficient are used as the input layer, and the position of the average wind speed circle from the tunnel top is used as the output layer, wherein [X 1 ,X 2 ,…X m ] is the input sample, [Y 1 ,Y 2 ,…Y n ] is the output sample, and the output of each unit in the hidden layer is calculated as follows:

[0024] O j =f(∑W ij X i -θ j ) (3);

[0025] The output of the output layer is:

[0026] Y k =f(∑T JK O j -θ k ) (4);

[0027] Where: θ is the neuron threshold; f is the nonlinear function; W ij is the weight matrix of the hidden layer; T JK is the weight matrix of the output layer.

[0028] Furthermore, in step S4, it also includes:

[0029] Step S41, selecting the positions of some simulated average wind speed circles from the top of the tunnel in step S24 as test samples of the BP neural network model;

[0030] Step S42, taking the remaining portion of the simulated average wind speed circle from the top of the tunnel in step S24 as a predicted value;

[0031] Step S43, comparing the BP neural network prediction result with the predicted value, if the error result is within 4%, the position of the average wind speed circle output by the BP neural network prediction model from the top of the tunnel is used as the installation position of the wind speed sensor.

[0032] The method for accurately predicting the installation position of a tunnel wind speed sensor of the present invention has the following beneficial effects: the distance between the average wind speed circle position and the top plate predicted by the BP neural network is compared with the simulation result, and the error range of the prediction result is within 4%. The monitoring data for predicting and guiding the installation position of the wind speed sensor is basically consistent with the manual wind measurement data, which proves that the neural network model can solve the change of the average wind speed circle position under different conditions within the allowable error range, so as to determine the position of the wind speed sensor, has good applicability and feasibility, and has certain practical value.

[0033] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 A flow chart of a method for accurately predicting the installation position of a tunnel wind speed sensor according to the present invention;

[0036] Figure 2 Schematic diagram of tunnel simulation of the present invention, wherein (a) (b) are schematic diagrams of tunnel cross-section, and (c) (d) are schematic diagrams of tunnel three-dimensionally;

[0037] Figure 3 It is a simulated smooth distribution diagram of the semicircular arch tunnel of the present invention;

[0038] Figure 4 It is a simulation smooth distribution diagram of a rectangular lane of the present invention;

[0039] Figure 5 It is a schematic diagram of the BP neural network prediction model of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] At present, the use of wind speed sensors to monitor wind volume underground is a highly effective means, with the advantages of long-term online monitoring and real-time analysis. However, due to the characteristics of the wind flow itself in fluid mechanics and the specific underground environment, the accuracy of the monitoring data is low. Through numerical simulation, we can conclude that in the process of wind flow field changes on the tunnel section, there is an average wind speed circle in the tunnel, and the wind speed in this area can be used as the average wind speed of the tunnel section for reference. Therefore, the sensor can be installed at this position to improve the accuracy of wind speed monitoring. Based on this problem, the present invention provides a neural network prediction tool that can quickly determine the position of the average wind speed circle in the tunnel, and quickly calculate the position of the average wind speed circle to guide the installation position of the wind speed sensor, which can further improve the monitoring accuracy of the tunnel wind speed.

[0042] like Figure 1 As shown, the present invention provides a method for accurately predicting the installation position of a tunnel wind speed sensor, comprising the following steps: S1, analyzing the wind flow state of the tunnel to determine whether the wind flow state of the tunnel is a laminar flow state or a turbulent flow state;

[0043] The airflow in mine tunnels is a continuous and uniform medium. During the normal ventilation period of the mine, the underground airflow is regarded as a stable fluid. There are two types of airflow states: laminar flow and turbulent flow. Laminar flow means that the particles in each layer of the fluid do not mix with each other, and the trajectory of the particle flow is a straight line or a regular smooth curve, and is basically parallel to the axis of the pipeline. Turbulent flow means that the particles of the fluid are strongly mixed with each other, and the flow trajectory of the particles is extremely irregular. In addition to displacement along the general direction of flow, there is also displacement perpendicular to the general direction of flow, and there are vortices inside the fluid that appear and disappear from time to time.

[0044] The flow state of the fluid is affected by factors such as the fluid velocity, viscosity and pipe size, and can be expressed as a dimensionless parameter R e (Reynolds number) is used to represent the fluid flow state, the average velocity of the fluid in the circular tunnel, m / s:

[0045]

[0046] Average velocity of fluid in non-circular tunnel, m / s:

[0047]

[0048] Among them, R e is the Reynolds number; r 0 is the diameter of the circular pipe, m; α is the kinematic viscosity coefficient of the fluid, U represents the circumference of the tunnel section, and S represents the cross-sectional area of ​​the tunnel section.

[0049] The Reynolds number takes the critical value of 2000, and the tunnel cross-sectional area S = 14.2m 2 , circumference U = 15.8m, the kinematic viscosity coefficient of the wind flow is 14.4*10-6m 2 / s, the average wind speed v = (2000*15.8*14.4*10-6) / (4*14.2) = 0.008 (m / s), that is, the average wind speed of the tunnel airflow transitioning from laminar flow to turbulent flow is 0.008m / s. The "Coal Mine Safety Regulations" require that the minimum allowable wind speed in the shaft and tunnel be 0.15m / s, which is much larger than the calculated result. Therefore, it can be considered that the wind flow in most shafts and tunnels is in a turbulent state.

[0050] S2. If the wind flow state of the tunnel is turbulent, numerical simulation is performed on the average wind speed circle of the tunnel section to obtain the distance between the average wind speed circle and the top of the tunnel;

[0051] According to the actual situation underground in the coal mine, two different types of tunnels, semicircular arch and rectangular, were designed. Based on three different support methods, anchor shotcrete, anchor mesh and masonry, a large number of numerical simulations were carried out under different tunnel cross-sectional areas and different wind speeds.

[0052] According to the actual situation of the lane, the X-axis is the lane length, the Y-axis is the lane height, and the Z-axis is the lane width. Figure 2 The three-dimensional model of the roadway shown, Figure 2 (a) (b) are schematic diagrams of tunnel cross-sections, (c) (d) are three-dimensional schematic diagrams of tunnels, and the tunnel-related parameter settings are shown in Table 1.

[0053] Table 1 Geometric model related parameters

[0054]

[0055] For the convenience of calculation, the following assumptions are made: (1) The mine airflow is incompressible; (2) The inlet airflow is in a stable state, satisfying the Boussinesq assumption; (3) The tunnel wall is adiabatically constant temperature, ignoring the heat generated by the friction between the airflow and the tunnel wall; Using the pre-model processing software CFD, according to the needs of each scheme, a three-dimensional geometric model of corresponding size is established, and after Boolean operations, the model is meshed in units of 0.02mm. During the simulation process, the pressure implicit separation solver is used, the speed is selected as absolute speed, the turbulence model is selected as the standard two-equation model, and the flow field solution calculation method is selected as the Simple algorithm.

[0056] The inlet of the simulated tunnel model is the airflow inlet, the outlet is the airflow outlet, and the rest are walls. The airflow inlet type is velocity-inlet. The airflow velocity at the airflow inlet is given according to the actual situation, which are 2m / s, 3m / s, 4m / s, 5m / s, 6m / s, 7m / s, 8m / s, and 9m / s respectively. The inlet wind speed is assumed to be uniform, and the pressure gradient perpendicular to the wall is zero; the airflow outlet type is pressure-outlet; the boundary adopts a no-slip boundary condition, and the wall is assumed to be adiabatic. At the same time, the friction resistance coefficient of different support conditions of the actual tunnel is converted into Roughness Height of 0.08m, 0.05m, and 0.02m respectively, which is used as the friction resistance coefficient of the tunnel. The schematic diagram of the simulation results is shown as follows Figure 3 and Figure 4 The simulation data are shown in Table 2.

[0057] Table 2 BP neural network learning and testing samples

[0058]

[0059]

[0060] S3, establish a BP neural network prediction model to output the position of the average wind speed circle from the top of the tunnel;

[0061] BP neural network is a multi-layer feedforward network that uses the gradient descent method. Its focus is to achieve back propagation training of errors. Network learning and training have two propagation processes, positive and negative. From the perspective of the forward propagation process, after the hidden layer receives the input signal, it acts on the next level through the corresponding transfer function. After a series of transformations, the output signal we need is generated. The expected value is the standard. When the result is significantly different from it, the error will propagate in the opposite direction. Back propagation is to transmit the error through the hidden layer to the input layer in reverse, so as to achieve the purpose of continuously updating the weights of each unit. With the learning and training of the neural network, the weights are also continuously iterated and updated. The three-layer network structure is as follows: Figure 5 As shown, where [X 1 ,X2 ,…X m ] is the input sample, [Y 1 ,Y 2 ,…Y n ] is the output sample.

[0062] Then establish the BP neural network prediction model, such as Figure 5 As shown in the figure, the tunnel section type, tunnel section height, tunnel section width, and tunnel friction coefficient are used as the input layer, and the position of the average wind speed circle from the top of the tunnel is used as the output layer. The output of each unit in the hidden layer is calculated as follows:

[0063] O j =f(∑W ij X i -θ j )

[0064] The output of the output layer is:

[0065] Y j =f(∑T JK O j -θ k )

[0066] Where: θ is the neuron threshold; f is the nonlinear function; W ij is the weight matrix of the hidden layer; T JK is the weight matrix of the output layer.

[0067] S4. Compare the position of the average wind speed circle simulated in step S2 from the top of the tunnel with the position of the average wind speed circle output by the BP neural network prediction model in step S3 from the top of the tunnel. If the error result is within 4%, the position of the average wind speed circle output by the BP neural network prediction model from the top of the tunnel is used as the installation position of the wind speed sensor.

[0068] The data samples with serial numbers 192 to 288 in Table 2 are selected as test samples of the neural network, where the simulation value is the result of numerical simulation, the prediction value is the prediction result of the BP neural network with data numbers 1 to 191 as samples, and the relative error is the difference between the two. If the error result is within 4%, the position of the average wind speed circle output by the BP neural network prediction model from the top of the tunnel is used as the installation position of the wind speed sensor.

[0069] The network performance is evaluated by comparing the prediction results with the simulation results. The smaller the relative error, the better the network performance. The results are shown in Table 3.

[0070] Table 3 Comparison of network prediction value and simulation value results

[0071]

[0072] In order to prove the engineering value of BP neural network in predicting the average wind speed circle, an industrial test was conducted in a mine in Shanxi Province. Three tunnels with different cross-sectional areas were selected, and the positions of their average wind speed circles were predicted by the prediction model. Wind speed sensors were installed at their positions to monitor the wind speed and compared with the manual wind measurement at the same position. The results are shown in Table 4 below. According to the comparison between the monitoring results of the wind speed sensor and the manual wind measurement results, the wind speeds on both sides are basically compatible, and the relative error is less than 2%, indicating that the prediction model based on BP neural network can provide support and reference for the installation location of underground wind speed sensors.

[0073] Table 4 Average wind speed sensor monitoring and wind meter measured data results

[0074]

[0075] In summary, the invention provides a method for accurately predicting the installation position of a tunnel wind speed sensor. The method compares the distance between the average wind speed circle position and the top plate predicted by the BP neural network with the simulation result. The error range of the prediction result is within 4%. The monitoring data of the wind speed sensor installation position guided by the prediction is basically consistent with the manual wind measurement data, which proves that the neural network model can solve the change of the average wind speed circle position under different conditions within the allowable error range, so as to determine the position of the wind speed sensor. The method has good applicability and feasibility, and has certain practical value.

[0076] The various devices selected in this application are all universal standard parts or parts known to those skilled in the art, and their structures and principles can be known to those skilled in the art through technical manuals or conventional experimental methods.

[0077] In the description of the embodiments of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or a connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0078] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0079] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A method for accurately predicting the installation position of a tunnel wind speed sensor, characterized in that: The following steps are involved: S1. Analyze the airflow state of the laneway and determine whether the airflow state of the laneway is laminar or turbulent; S2. If the wind flow state of the tunnel is turbulent, numerical simulation is performed on the average wind speed circle of the tunnel section to obtain the distance between the average wind speed circle and the top of the tunnel; S3, establish a BP neural network prediction model to output the position of the average wind speed circle from the top of the tunnel; S4. Compare the position of the average wind speed circle simulated in step S2 from the top of the tunnel with the position of the average wind speed circle output by the BP neural network prediction model in step S3 from the top of the tunnel. If the error result is within 4%, the position of the average wind speed circle output by the BP neural network prediction model from the top of the tunnel is used as the installation position of the wind speed sensor.

2. The method for accurately predicting the installation position of a tunnel wind speed sensor according to claim 1, characterized in that: In step S1, the average velocity v of the circular channel fluid, m / s: Average velocity v of fluid in non-circular tunnel, m / s: Among them, R e is the Reynolds number; r0 is the diameter of the circular pipe, m; α is the kinematic viscosity coefficient of the fluid, U represents the circumference of the tunnel section, and S represents the cross-sectional area of ​​the tunnel section.

3. The method for accurately predicting the installation position of a tunnel wind speed sensor according to claim 2, characterized in that: According to formula (1) and formula (2), the average velocity v of the fluid in the tunnel is obtained. If v is greater than 0.15 m / s, the wind flow state in the tunnel is determined to be a turbulent state.

4. The method for accurately predicting the installation position of a tunnel wind speed sensor according to claim 1, characterized in that: In step S2, the following steps are also included: S21. According to the actual situation of the lane, a three-dimensional coordinate model of the lane is established with the X axis as the lane length, the Y axis as the lane height, and the Z axis as the lane width; S22. Assumptions a. The mine airflow is incompressible; b. The inlet airflow is in a stable state, satisfying the Boussinesq assumption; c. The tunnel wall is adiabatically constant temperature, ignoring the heat generated by the friction between the airflow and the tunnel wall; S23, using the front model processing software CFD to establish a three-dimensional coordinate model, and through Boolean operations, meshing the model; S24, use the pressure implicit separation solver, select absolute speed as the speed, select the standard two-equation model as the turbulence model, select the Simple algorithm as the flow field solution calculation method, and simulate the position of the average wind speed circle from the top of the tunnel.

5. The method for accurately predicting the installation position of a tunnel wind speed sensor according to claim 4, characterized in that: In step S3, the BP neural network model includes: an input layer, a hidden layer and an output layer. The tunnel section type, tunnel section height, tunnel section width and tunnel friction coefficient are used as the input layer, and the position of the average wind speed circle from the top of the tunnel is used as the output layer. m ] is the input sample, [Y1,Y2,…Y n ] is the output sample, and the output of each unit in the hidden layer is calculated as follows: The j =f(∑W ij X i -θ j ) (3); The output of the output layer is: Y j =f(∑T JK The j -θ k ) (4); Where: θ is the neuron threshold; f is the nonlinear function; W ij is the weight matrix of the hidden layer; T JK is the weight matrix of the output layer.

6. The method for accurately predicting the installation position of a tunnel wind speed sensor according to claim 5, characterized in that: In step S4, it also includes: Step S41, selecting the positions of some simulated average wind speed circles from the top of the tunnel in step S24 as test samples of the BP neural network model; Step S42, taking the remaining portion of the simulated average wind speed circle from the top of the tunnel in step S24 as a predicted value; Step S43, comparing the BP neural network prediction result with the predicted value, if the error result is within 4%, the position of the average wind speed circle output by the BP neural network prediction model from the top of the tunnel is used as the installation position of the wind speed sensor.

Images