Bypass self-supplying bend structure for pneumatic conveying and its air supply control method

Abstract

This invention discloses a bypass self-supplying bend structure and its air supply control method for pneumatic conveying. The structure includes a straight inlet pipe section, a bend section, and a straight outlet pipe section connected in sequence, and a bypass pipe bridging the side wall of the straight inlet pipe section and the outer arc side of the bend section. The outlet end of the bypass pipe corresponds to the main particle impact zone, and the return airflow is tangential to the inner wall of the outer arc side of the bend section. The inner diameter of the bend section is smaller than that of the straight inlet and outlet pipe sections. By utilizing the static pressure difference created by the reduced diameter, bypass self-supply is achieved without an external compressed air source. A regulating valve is installed on the bypass pipe, which can quickly correct the valve position through a pressure feedback control channel and periodically update the nominal valve position through an operating point optimization control channel, achieving dual-time-scale closed-loop control. This invention can significantly reduce the particle-pipe wall collision intensity, improve the localized concentrated distribution of collisions, reduce the risk of bulk grain particle breakage and bend erosion and abrasion, while also ensuring conveying pressure safety.

Description

Technical Field

[0001] This invention relates to the field of pneumatic conveying technology, and in particular to a bypass self-supplying bend structure for pneumatic conveying and its air supply control method. Background Technology

[0002] In the grain storage and transportation industry, pneumatic conveying is a technology that uses pressure differentials to create a directional gas-solid two-phase flow to transport bulk grain. While it enables efficient, sealed, and continuous conveying, it also leads to problems such as increased breakage rates, higher energy consumption, and pipe wall wear. When grain enters or leaves the storage facility, the gas-solid two-phase flow changes direction as it passes through bends in the pipes. Grain grains, due to inertia, tend to deflect towards the outer arc of the bend and violently impact the pipe wall, making the bend a primary area for grain breakage, energy loss, and pipe wall wear. Therefore, optimizing the local airflow in the bends and adjusting their operating conditions are crucial for improving bend performance and achieving low-loss, stable operation of the grain pneumatic conveying system.

[0003] Existing technologies regulate local airflow in bends from three aspects: bypass air adjustment, external airflow cushion, and pressure feedback control. Patent document CN113307037A uses a bypass pipeline to divert part of the airflow at the bend inlet to the outlet to form a vortex, thus improving particle movement within the bend. However, its return air inlet is located on the bend outlet side, failing to act on the main particle impact zone. Furthermore, it employs a fixed bypass structure, meaning the bypass air supply cannot be actively adjusted according to the conveying conditions, limiting its ability to improve the main particle impact zone. It also lacks a control method suitable for bulk grain handling. Patent document CN211920164U uses an external air replenishment and pressurization device to input compressed air to the outer arc side of the bend, forming a stable airflow. While the system protects the pipe wall of the bend by providing air cushions, relying on external air supply to form these cushions increases additional energy consumption for air supply, and the amount of air supplied is not adjusted in a closed loop based on the actual collision conditions. Patent document CN111377243B employs a control method that combines main pipe frequency conversion constant pressure feedback with branch pipe PID constant pressure regulation, and maintains the air supply pressure difference between the main pipe and branch pipe through pressure monitoring rules. However, its control objective is mainly to maintain pipeline pressure, and it does not incorporate bend vibration and the collision frequency and energy it represents into the feedback and operating point optimization, and it ignores the blockage risk caused by changes in pipeline pressure.

[0004] In summary, existing solutions fail to achieve adaptive and coordinated control based on the particle-pipe wall collision state while also considering particle breakage and pipe wear. Furthermore, existing bypass gas regulation solutions either cannot operate within the main particle impact zone, cannot actively adjust, or rely on an external compressed air source, thus increasing energy consumption. Therefore, it is necessary to provide a pipe structure and corresponding control method that does not require an external compressed air source, can operate within the main particle impact zone, and can actively adjust the bypass gas supply according to changes in conveying conditions. Summary of the Invention

[0005] To overcome the above-mentioned shortcomings of existing pneumatic conveying bend technology, the technical problem to be solved by the present invention is to provide a bypass self-supplying bend structure and its air supply control method for pneumatic conveying that does not require an external compressed air source, can actively adjust the bypass air supply intensity according to the conveying conditions, takes into account particle crushing and bend wear, and ensures safe conveying pressure.

[0006] The technical solution adopted by this invention to solve its technical problem is:

[0007] A bypass self-supplying bend structure for pneumatic conveying includes an inlet straight pipe section, a bend section, and an outlet straight pipe section connected in sequence, and a bypass pipe. The inlet end of the bypass pipe is located on and connected to the side wall of the inlet straight pipe section, and the outlet end of the bypass pipe is located on and connected to the outer arc side of the bend section. The intersection point of the axis of the inlet straight pipe section and the inner wall of the outer arc side of the bend section is located in the middle of the intersection area of ​​the bypass pipe and the bend section. The airflow entering the bend section from the outlet end of the bypass pipe is tangential to the inner wall of the outer arc side of the bend section. The inner diameters of the inlet straight pipe section and the outlet straight pipe section are the same, and the inner diameter of the bend section is smaller than the inner diameters of the inlet straight pipe section and the outlet straight pipe section. The connection between the inner wall of the bend section and the inner walls of the inlet straight pipe section and the outlet straight pipe section is achieved by a bevel transition.

[0008] Furthermore, the bypass pipeline consists of a smoothly transitioning variable diameter section, an arc section, and a straight section from the air inlet to the air outlet. The inner diameter of the variable diameter section gradually decreases from the air inlet to the arc section, and the inner diameters of the arc section and the straight section are the same as the inner diameter of the small end of the variable diameter section.

[0009] Furthermore, the section of the bend that connects to the outlet of the bypass pipe has a reduced diameter section, the inner diameter of which is smaller than the inner diameter of the rest of the bend.

[0010] Furthermore, the inner diameter of the reduced-diameter section in the bend is... The inner diameter of the main pipe in the remaining parts of the bend section is The inner diameter of the straight section of the bypass pipe is The distance between the center projections of the inlet and outlet ends of the bypass pipe onto the axis of the straight inlet pipe section is... It satisfies the following relationship:

[0011] ,

[0012] in, The value is 0.75~0.85. It is 2.5~3.5. The value is 2.5~5.5.

[0013] Furthermore, It is 0.85. It is 3.22. It is 2.98.

[0014] Furthermore, a regulating valve is provided on the bypass pipeline.

[0015] Furthermore, a pressure sensor is installed at the air inlet end of the straight inlet pipe section opposite to the bypass pipe, and multiple vibration sensors are installed on the outer and inner arc sides of the bend section from the part connected to the bypass pipe section to the part connected to the straight outlet pipe section; the regulating valve is an electric ball valve, and the pressure sensor and all vibration sensors are electrically connected to the controller of the electric ball valve.

[0016] Furthermore, the controller of the electric ball valve is configured to include a pressure feedback control channel based on a pressure sensor, an operating point optimization control channel based on a vibration sensor, and an instruction synthesis and execution channel; the pressure feedback control channel uses a sampling period... During operation, based on the deviation between the pressure setpoint and the pressure feedback value from the pressure sensor, a dynamic valve position compensation value is generated; the operation point optimization control channel optimizes the update cycle. run, The collision state evaluation index and closed-loop performance index are calculated based on the pressure feedback value and the vibration signal from the vibration sensor, and the nominal valve position is optimized and updated. The command synthesis and execution channel synthesizes the nominal valve position and the valve position dynamic compensation amount into a synthesized valve position command to drive the actuator of the electric ball valve.

[0017] In addition, the present invention also provides a method for controlling the air supply of a bypass self-supplying bend structure for pneumatic conveying, which employs the above-mentioned bypass self-supplying bend structure for pneumatic conveying and includes the following steps:

[0018] S1: Set the pressure setpoint, sampling period, optimization update period, allowable valve position set and pressure control deviation upper limit, select the initial nominal valve position from the known safety valve positions, and establish the initial Gaussian process proxy model of the closed-loop performance index and pressure constraint function;

[0019] S2: After each sampling cycle, read the pressure feedback value of the pressure sensor and calculate the pressure control deviation. Generate the valve position dynamic compensation amount by the filter over-torsion control, and combine it with the current nominal valve position to form a composite valve position command.

[0020] S3: Implements safety supervision, mode switching, valve position command amplitude limiting, and valve position command rate limitation on the synthesized valve position command, generates a restricted valve position command and drives the electric ball valve to reach the actual valve position, changes the flow rate of the bypass pipeline, and corrects the pressure feedback value.

[0021] S4: Record the vibration signals of each vibration sensor. After each optimization update cycle, calculate the collision event count index, vibration energy index, closed-loop performance index and pressure constraint function, and add the observation results of this cycle to the Gaussian process sample set.

[0022] S5: Update the objective function proxy model, pressure constraint proxy model, safety set, and reachable safety set, and select the nominal valve position for the next optimization update cycle within the reachable safety set;

[0023] S6: Repeat steps S3 to S5. When the pressure control deviation, nominal valve position change, and closed-loop performance index change continuously meet the convergence conditions, it is determined that the system has entered steady-state operation. When the operating conditions change, the nominal valve position is updated again and converged to a new safe operating point.

[0024] Furthermore, in step S2, the pressure control deviation is defined as:

[0025] ,

[0026] In the formula: For the first Pressure control deviation at each sampling time, in units of ; Pressure setpoint, unit: ; For the pressure sensor in the first Pressure feedback values ​​obtained at each sampling time, in units of ;

[0027] A filtered over-twist controller with a first-order filtering state is used to generate the valve position dynamic compensation quantity, which is in continuous form according to the sampling period. Discretize it, and write it as:

[0028] ,

[0029] In the formula: In the filtering state, The rate at which the filter state changes over time; For filter gain, ; It is a symbolic function;

[0030] The valve position dynamic compensation is generated by the following formula:

[0031] ,

[0032] In the formula: This is the dynamic compensation amount for the valve position; This is a super-distorted integral state. It is an integral state. The rate of change over time; and To control the gain, and , The values ​​of each gain are determined based on the actual valve position's effect on the pressure feedback value and the dynamic response.

[0033] Limit the dynamic compensation amount of the valve position:

[0034] ,

[0035] In the formula: This is the dynamic compensation amount for the valve position after amplitude limiting; Valve position compensation limit; This indicates that the input is restricted to a range. Inside;

[0036] In the Within each optimization update cycle, the synthesized valve position command is:

[0037] ,

[0038] In the formula: This is a synthesized valve position command; For the first The nominal valve position used in each optimization update cycle;

[0039] In step S3, the synthesized valve position command is sequentially subjected to valve position command amplitude limiting and valve position command rate limiting:

[0040] ,

[0041] ,

[0042] In the formula: For valve position commands with limited amplitude; and These are the minimum and maximum allowable valve positions, respectively. This is a restricted valve position command; The maximum permissible rate of valve position change, in units of ; The sampling period is expressed in units of 10 ... Electric ball valve according to Make the actual valve position follow the restricted valve position command;

[0043] In step S4, the first The optimization update cycle is:

[0044] ,

[0045] In the formula: For the first The time range of each optimization update cycle;

[0046] Four vibration sensors were installed at selected measuring points on the bend section. The operation point optimization control channel processed the signals from each vibration sensor within this cycle. Filtered vibration signal at each measuring point To perform event identification, the collision event count metric is defined as follows:

[0047] ,

[0048] In the formula: For the first Within the first optimization update cycle, the first Collision event count index at the measuring point; The number of candidate vibration events identified; For the first The characteristic amplitude of each candidate event; For the first Event recognition threshold for measurement points; It is an indicator function; It only indicates the number of vibration events that meet the criteria, and is not equivalent to the count of actual collisions that are not calibrated;

[0049] No. The vibration energy index of the measuring point is defined as follows:

[0050] ,

[0051] In the formula: It is a dimensionless vibrational energy index; This is a reference value for vibration acceleration, in units of... ; For the pre-processed first Vibration acceleration at the measuring point, in units of This index is used to characterize the vibration intensity at the measuring point and is not equivalent to the contact energy in the discrete element method.

[0052] The four measurement point indices are synthesized after normalization as follows:

[0053] ,

[0054] In the formula: and These are the normalized collision event count index and the normalized vibration energy index, respectively. and The first Reference values ​​for measuring points; For the first The weight of the measurement points ;

[0055] Closed-loop performance index is defined as:

[0056] ,

[0057] In the formula: For the first The closed-loop performance index of each optimization and update cycle; the smaller the value, the better the overall performance. , and These are the weights for the collision event count, vibration energy, and pressure tracking term, respectively. This is a normalized reference value for pressure, in units of ;

[0058] The pressure constraint function is defined as follows:

[0059] ,

[0060] In the formula: This is the pressure constraint function; The upper limit of the allowable pressure control deviation, in units of ;when When this occurs, it indicates that the optimization update cycle meets the pressure constraint;

[0061] The run-point optimization control channel employs one-dimensional safe Bayesian run-point optimization, establishing Gaussian process surrogate models for both closed-loop performance indices and pressure constraint functions; let Let represent the objective function and the constraint function, respectively. Then their posterior mean and posterior variance are:

[0062] ,

[0063] ,

[0064] In the formula: For the first Each optimization update cycle is at the valve position. The posterior mean at point ; This represents the number of existing samples; Valve position The kernel function vector between the kernel function and existing sample points; The kernel matrix of the existing sample points; To observe the noise variance; It is the identity matrix; This is a vector of existing observations; The posterior standard deviation; Valve position The prior kernel function value at that position; the posterior standard deviation is used to characterize the prediction uncertainty of the surrogate model at that valve position;

[0065] In step S5, the confidence bounds of the objective function and the constraint functions are defined as follows:

[0066] ,

[0067] In the formula: and These are the lower confidence bound and the upper confidence bound, respectively. For the first Confidence coefficient for each optimization update cycle;

[0068] Take in the constraint direction Under safe conditions, the safe set is defined as:

[0069] ,

[0070] In the formula: For the first A safety set for each optimization and update cycle; valve position set ; The upper confidence bound is used for the pressure constraint proxy model; by using the upper confidence bound judgment, candidate valve positions that meet the pressure constraints can be screened while considering the model's uncertainty.

[0071] The security correction set and the reachable security region are defined as follows:

[0072] ,

[0073] ,

[0074] In the formula: For the first A set of security fixes for each optimization and update cycle; This is the candidate nominal valve position correction amount; Optimize the maximum allowable nominal valve position change during the update cycle; It is the safe reachable range, which is the set of adjustable safety valve positions; and The meaning is the same as before;

[0075] Within the reachable safety set, using the lower confidence bound of the objective function as the acquisition function, the next nominal valve position can be determined by the following formula:

[0076] ,

[0077] In the formula: The nominal valve position for the next optimization update cycle; This indicates the valve position at which the acquisition function within the parentheses reaches its minimum value; and These are the posterior mean and posterior standard deviation of the closed-loop performance index proxy model, respectively; this control method balances the current performance prediction and the exploration value of unknown valve positions within a safe range; if If no candidate valve position meets the update conditions, then maintain... .

[0078] The beneficial effects of this invention are:

[0079] 1. This invention provides a bypass pipe connecting the inlet straight pipe section and the bend pipe section on the outer arc side of the bend pipe section, and sets the inner diameter of the bend pipe section to be smaller than the inner diameters of the inlet straight pipe section and the outlet straight pipe section. This reduces the flow cross-sectional area of ​​the bend pipe section, adjusts the static pressure distribution between the inlet and outlet ends of the bypass pipe, and forms a bypass driving pressure difference. This allows some gas to return to the bend pipe section through the bypass pipe without relying on an external compressed gas source, achieving bypass self-supply. This can form an airflow-particle mixing cushion layer in the main particle impact area, thereby reducing the collision energy between grain particles and between particles and the inner wall of the bend pipe section, thus simultaneously reducing the risk of particle breakage and the risk of pipe wall erosion and abrasion.

[0080] 2. By installing a regulating valve on the bypass pipeline, and installing pressure sensors and vibration sensors electrically connected to the regulating valve on the straight and bend sections of the pipeline, the regulating valve can change the air flow rate of the bypass pipeline according to the conveying conditions. This overcomes the problem that the fixed bypass structure cannot actively adjust the return air intensity, and makes the bypass airflow adapt to the main pressure and particle movement state, thereby achieving coordinated control of pipeline conveying pressure and particle collision intensity.

[0081] 3. The air flow rate of the bypass pipeline is regulated by a dual-time-scale closed-loop control system. The pressure feedback control channel suppresses pressure disturbances with a shorter sampling period, while the operating point optimization control channel reduces collision state evaluation indicators with a longer optimization update period. The dual-time-scale closed-loop control can take into account both the dynamic stability of pressure and the requirements for long-term low-loss operation, thereby improving the stability of grain pellet conveying. Attached Figure Description

[0082] Figure 1 A cross-sectional view of the bypass self-supplying air bend structure;

[0083] Figure 2 This is a cross-sectional view of the control valve;

[0084] Figure 3 This is a block diagram of a dual-time-scale closed-loop control system.

[0085] Figure 4 The flowchart shows a dual-time-scale closed-loop control method.

[0086] Figure 5 This is a diagram showing the trajectory and velocity distribution of particles inside a typical bend in a pipe.

[0087] Figure 6 This is a diagram showing the particle movement trajectory and velocity distribution within the bypass self-supplying bend of the present invention;

[0088] Figure 7 This is a comparison diagram of the average normal contact force between particles and the pipe wall in a conventional bend and the novel bend of this invention.

[0089] Figure 8 This is a comparison diagram of the average tangential contact force between particles and the pipe wall in a conventional bend and the novel bend of this invention.

[0090] Figure 9 A comparison diagram of the maximum cumulative contact energy between a conventional bend and the novel bend of this invention;

[0091] Figure 10 The diagram shows the energy loss distribution of a single collision between a conventional bend and the novel bend of this invention.

[0092] Figure 11 for hour , Response surface plot with respect to maximum cumulative contact energy;

[0093] Figure 12 for hour , Response surface plot with respect to maximum cumulative contact energy;

[0094] Figure 13 for hour , Response surface plot with maximum cumulative contact energy.

[0095] The markings in the diagram are as follows: 1. Inlet straight pipe section; 2. Bend section; 21. Reduction section; 3. Outlet straight pipe section; 4. Bypass pipe; 41. Air inlet; 42. Reduction section; 43. Arc section; 44. Straight section; 45. Air outlet; 5. Control valve; 51. Electric actuator; 52. Valve stem; 53. Valve body; 54. Ball; 6. Pressure sensor; 7. Vibration sensor. Detailed Implementation

[0096] The invention will be further described below with reference to the accompanying drawings.

[0097] It should be noted that if this invention uses directional terms such as up, down, left, right, front, and back, these are for describing the relative positions of components and are not specific references to the absolute positions of related components or the relationships between them. They are only used to explain the relative positional relationships and movements of components in a specific posture. If the specific posture changes, the directional indication will also change accordingly. If this invention uses terms related to quantity such as "many," "multiple," or "several," these specifically refer to two or more.

[0098] like Figure 1 As shown, the present invention provides a bypass self-supplying bend structure for pneumatic conveying, comprising a straight inlet pipe section 1, a bend section 2, and a straight outlet pipe section 3 connected in sequence. These three sections can be integrally formed, welded, or flanged to form a main pipe for gas-solid two-phase flow. It also includes a bypass pipe 4, with its inlet end 41 located on and connected to the side wall of the straight inlet pipe section 1, and its outlet end 45 located on and connected to the outer arc side of the bend section 2. The intersection point of the axis of the straight inlet pipe section 1 and the inner wall of the outer arc side of the bend section 2 is located in the middle of the intersection area between the bypass pipe 4 and the bend section 2. The bypass pipe 4 can be welded to the straight inlet pipe section 1 and the bend section 2 or integrally cast. When the gas-solid two-phase flow passes through the bend section 2, the flow direction changes. The corn kernels are deflected towards the outer arc side of the bend section 2 due to inertia. The intersection point corresponds to the main area where the kernels impact the inner wall of the outer arc side. The middle of the intersection area between the bypass pipe 4 and the bend section 2 is set at this intersection point, so that the bypass airflow directly acts on the main impact area of ​​the kernels. The airflow entering the bend section 2 from the outlet 45 of the bypass pipe 4 can be along the radial direction of the bend section 2 or at a certain angle to the radial direction, as long as it can act on the grain kernels in the main impact area. However, the preferred solution is to be as tangential to the inner wall of the outer arc side of the bend section 2 as possible. The bypass airflow entering tangentially flows along the inner wall of the outer arc side of the bend section 2, which can form a good airflow-particle mixing cushion layer in the main impact area of ​​the kernels, thereby buffering and guiding the grain kernels that collide with the wall.

[0099] The inner diameters of the inlet straight pipe section 1 and the outlet straight pipe section 3 are the same, while the inner diameter of the bend section 2 is smaller than that of both the inlet straight pipe section 1 and the outlet straight pipe section 3. The connection between the bend section 2 and the inlet straight pipe section 1 and the outlet straight pipe section 3 is made by a bevel transition to avoid local eddies and particle retention caused by abrupt changes in cross-section. The relatively smaller inner diameter of the bend section 2 accelerates the main airflow and reduces the static pressure within it, thereby adjusting the static pressure distribution between the inlet end 41 and the outlet end 45, establishing a bypass driving pressure difference. This allows some gas to enter the bypass pipe 4 and return to the bend section 2 without relying on an external compressed air source, achieving bypass self-supply. The self-supply process is established synchronously with the main pipe delivery.

[0100] The bypass pipe 4 consists of a smoothly transitioning variable diameter section 42, an arc section 43, and a straight section 44, from the inlet end 41 to the outlet end 45. The inlet end 41 is connected to the straight pipe section 1 and is used to draw part of the gas from the main pipe. The inner diameter of the variable diameter section 42 gradually decreases from the inlet end 41 to the arc section 43, so that the bypass airflow gradually accelerates within the variable diameter section 42, avoiding excessive friction loss caused by abrupt changes in cross-section. The arc section 43 changes the direction of the bypass airflow, directing it to the outer arc side of the bend section 2. The straight section 44 is connected to the outlet end 45, so that the accelerated bypass airflow enters the bend section 2 tangentially through the outlet end 45 in a stable flow state, acting on the main particle impact zone of the bend section 2.

[0101] In addition, to increase the flow rate and velocity of the bypass airflow, a reduced-diameter section 21 is provided at the connection point between the bend section 2 and the outlet end 45 of the bypass pipe 4. The inner diameter of the reduced-diameter section 21 is smaller than the inner diameter of the rest of the bend section 2. The reduced-diameter section 21 can be integrally formed with the bend section 2, or a detachable reduced-diameter sleeve, a tapered pipe section, or an annular throttling component can be used to achieve the reduced-diameter effect. By further reducing the local flow cross-sectional area of ​​the main pipe, the reduced-diameter section 21 further reduces the static pressure near the outlet end 45, thereby increasing the bypass driving pressure difference between the inlet end 41 and the outlet end 45, ensuring that the bypass airflow has sufficient flow rate and velocity, and forming a stable airflow-particle mixing cushion layer in the main particle impact zone.

[0102] To achieve parametric design of the pipeline structure, this invention defines three dimensionless structural parameters: the inner diameter of the reduced diameter section 21 in the bend section 2 is... The inner diameter of the main pipe in the remaining parts of bend section 2 is The inner diameter of the straight section 44 of the bypass pipe 4 is The distance between the center projections of the inlet end 41 and outlet end 45 of the bypass pipe 4 onto the axis of the straight inlet pipe section 1 is... It satisfies the following relationship:

[0103] ,

[0104] in, For the diameter reduction ratio, This is the ratio of the inner diameter of the reduced-diameter section to the straight section of the bypass pipe. This is the ratio of the axial spacing of the bypass connection to the inner diameter of the main pipe; based on extensive experimental analysis, the optimal range for the above parameters is: The value is 0.75~0.85. It is 2.5~3.5. The value is 2.5~5.5.

[0105] The optimal range of the above parameters is based on the following: from the perspective of structural function, The degree to which the reduced diameter section 21 adjusts the local static pressure distribution of the main pipe is determined. If the diameter reduction is too small, the local resistance of the main pipe will increase, energy consumption will rise, and particles will easily become stuck at the 21-point diameter reduction section. If the diameter is too large, the narrowing effect is insufficient, making it difficult to establish a sufficient bypass drive pressure difference; The ability of the bypass duct airflow to interfere with the main duct airflow is determined. When the diameter of the straight section 44 is too large, the bypass flow rate is too high, and the impact of the return gas on the main flow field is too strong. When the flow rate is too large, the bypass channel becomes too narrow, and the bypass airflow is insufficient to disperse the concentrated contact effect of the main impact zone of the particles. It affects the location of the bypass connection and the flow path of the bypass airflow. When the value is too large, the bypass loss along the way increases. When the inlet end 41 and outlet end 45 are too close, the bypass drive pressure differential is insufficient. Three parameters jointly determine the bypass drive pressure differential, bypass flow rate, and return gas effect intensity, and there is a strong interaction among them, requiring coordinated matching. This invention optimizes the above parameters using a three-factor, three-level Box-Behnken response surface methodology (see the embodiments described later). The preferred value is 0.75~0.85. Preferably 2.5~3.5 The preferred value is 2.5~5.5, which can take into account the bypass driving pressure difference, bypass flow rate and return gas intensity, so that the particle-pipe wall contact effect changes from a local concentrated state to a more dispersed state.

[0106] More preferred parameters are, It is 0.85. It is 3.22. The value is 2.98. This combination is the optimal parameter combination obtained by response surface optimization. Under the same simulation conditions, the maximum cumulative contact energy at the termination time of the bent pipe structure using this parameter combination is further reduced by about 4.4% compared with the known minimum value in the Box-Behnken design, and the collision frequency is maintained at a low level.

[0107] To better control the bypass flow rate of bypass pipeline 4, a regulating valve 5 is installed on bypass pipeline 4. The regulating valve 5 is used to adjust the effective flow area of ​​bypass pipeline 4, thereby actively adjusting the bypass flow rate according to the conveying conditions. Figure 2 As shown, in this embodiment, the regulating valve 5 is an electric ball valve, including an electric actuator 51, a valve stem 52, a valve body 53, and a ball 54. The electric actuator 51 receives a restricted valve position command sent by an external controller and drives the ball 54 to rotate through the valve stem 52, changing the effective flow area between the flow channel of the ball 54 and the bypass pipe 4, so that the actual valve position reaches the position corresponding to the restricted valve position command. The electric ball valve has a smooth flow channel and low flow resistance, and it is not easy for particles to accumulate when fully open, making it suitable for gas-solid two-phase flow applications containing solid particles.

[0108] In addition, a pressure sensor 6 is installed at the air inlet end 41 of the straight pipe section 1, which is directly opposite the bypass pipe 4. The pressure feedback value at this location can directly reflect the static pressure level at the air inlet end 41, i.e., the high-pressure side pressure of the bypass driving pressure difference. Multiple vibration sensors 7 are installed on the outer and inner arc sides of the bend section 2, from the part connected to the bypass pipe 4 to the part connected to the straight pipe section 3, to acquire the bend vibration signal characterizing the particle-pipe wall collision state. The pressure sensor 6 and all vibration sensors 7 are electrically connected to the controller of the electric ball valve. In this embodiment, there are four vibration sensors 7, namely three evenly spaced along the outer arc side of the bend section 2 and one located on the inner arc side near the straight pipe section 3, corresponding to four vibration measurement points, to cover the main particle impact area and its downstream area, and to comprehensively acquire the distribution information of collision frequency and collision energy along the bend section 2.

[0109] like Figure 3 As shown, the controller of the electric ball valve is configured to include a pressure feedback control channel based on pressure sensor 6, an operating point optimization control channel based on vibration sensor 7, and a command synthesis and execution channel. The pressure feedback control channel operates at a sampling period of... During operation, a dynamic valve position compensation value is generated based on the deviation between the pressure setpoint and the pressure feedback value from pressure sensor 6 to suppress transient disturbances; the operating point optimizes the control channel to optimize the update cycle. run, Based on the pressure feedback value and the vibration signal from vibration sensor 7, the collision state evaluation index and closed-loop performance index are calculated to optimize and update the nominal valve position. The nominal valve position determines the basic operating point of the bypass gas supply. The command synthesis and execution channel synthesizes the nominal valve position and the valve position dynamic compensation amount into a synthesized valve position command to drive the electric ball valve. The pressure feedback control channel forms a fast inner loop, and the operating point optimization control channel forms a slow outer loop. The two form a dual-time-scale closed-loop control.

[0110] This invention, in addition to the aforementioned bypass self-supplying bend structure for pneumatic conveying, also provides a gas supply control method, such as... Figure 4 As shown, the main steps include:

[0111] S1: Set the pressure setpoint, sampling period, optimization update period, allowable valve position set, and pressure control deviation upper limit. Select the initial nominal valve position from the known safety valve positions and establish an initial Gaussian process proxy model for the closed-loop performance index and pressure constraint function. In specific implementation, first determine the pressure setpoint based on the type and quantity of the conveyed material. upper limit of allowable pressure control deviation The sampling period is determined based on the dynamic characteristics of pressure disturbance and the statistical requirements of collision state. and optimize update cycle ,make The initial nominal valve position is selected from the known safety valve positions to ensure that the system is in a safe operating state during the initial stage of closed-loop commissioning.

[0112] S2: After the gas-solid two-phase flow enters the straight pipe section 1, the bend section 2 and the reduced diameter section 21 change the static pressure field distribution of the main pipe, forming a low static pressure zone on the outer arc side of the bend section 2. This creates a bypass driving pressure difference between the inlet end 41 and the outlet end 45, and some gas returns to the bend section 2 through the bypass pipe 4 and the regulating valve 5. The self-supply process is established synchronously with the main pipe delivery. After each sampling cycle... The pressure feedback value from pressure sensor 6 is read and the pressure control deviation is calculated. A dynamic valve position compensation value is generated by the filtered over-torsion control and combined with the current nominal valve position to form a composite valve position command. The pressure control deviation is defined as follows:

[0113] ,

[0114] In the formula: For the first Pressure control deviation at each sampling time, in units of ; Pressure setpoint, unit: ; For the pressure sensor in the first Pressure feedback values ​​obtained at each sampling time, in units of ;

[0115] The pressure feedback control channel uses a filtered over-torsion controller with a first-order filtering state to generate the valve position dynamic compensation quantity, which is in continuous form according to the sampling period. Discretize it, and write it as:

[0116] ,

[0117] In the formula: In the filtering state, The rate at which the filter state changes over time; For filter gain, ; The function is a sign function; this processing utilizes the filtering state to weaken the direct effect of pressure measurement noise on the control quantity, thus avoiding valve position jitter caused by noise.

[0118] The valve position dynamic compensation is generated by the following formula:

[0119] ,

[0120] In the formula: This is the dynamic compensation amount for the valve position; This is a super-distorted integral state. It is an integral state. The rate of change over time; and To control the gain, and , The values ​​of each gain are determined based on the actual valve position's effect on the pressure feedback value and the dynamic response. Filtered super-torsion control, while maintaining the finite-time convergence and strong disturbance rejection capability of the super-torsion algorithm, makes the control quantity continuously smooth through filtering, making it suitable for the continuous adjustment characteristics of electric ball valves.

[0121] To avoid excessive valve position correction caused by pressure transients, the dynamic compensation amount of the valve position needs to be limited:

[0122] ,

[0123] In the formula: This is the dynamic compensation amount for the valve position after amplitude limiting; Valve position compensation limit; This indicates that the input is restricted to a range. Inside;

[0124] In the Within each optimization update cycle, the synthesized valve position command is:

[0125] ,

[0126] In the formula: This is a synthesized valve position command; For the first The nominal valve position is used in each optimization update cycle. The nominal valve position determines the basic operating point of the bypass gas supply, and the valve position dynamic compensation is used to correct pressure disturbances near this operating point.

[0127] S3: Implements safety supervision, mode switching, valve position command amplitude limiting, and valve position command rate limiting on the synthesized valve position command, forming a restricted valve position command and driving the electric ball valve to reach the actual valve position, changing the flow rate of bypass line 4 and correcting the pressure feedback value. Specifically, the synthesized valve position command sequentially undergoes valve position command amplitude limiting and valve position command rate limiting:

[0128] ,

[0129] ,

[0130] In the formula: For valve position commands with limited amplitude; and These are the minimum and maximum allowable valve positions, respectively. This is a restricted valve position command; The maximum permissible rate of valve position change, in units of ; The sampling period is expressed in units of 10 ... Valve position command limiting ensures the valve position remains within the permissible range, while valve position command rate limiting prevents sudden valve position changes that could cause drastic fluctuations in mains pressure. Electric actuator 51, according to... The electric ball valve is driven so that the actual valve position follows the restricted valve position command. If pressure exceeds the limit or a critical channel fails, the system switches to safety mode and executes the preset safety valve position to ensure the safety of the conveying system.

[0131] S4: While the pressure feedback control continues to operate, record the vibration signals of each vibration sensor 7. After each optimization update cycle, calculate the collision event count index, vibration energy index, closed-loop performance index, and pressure constraint function, and add the observation results of this cycle to the Gaussian process sample set. Among these, the... The optimization update cycle is:

[0132] ,

[0133] In the formula: For the first The time range of each optimization update cycle;

[0134] The operation point optimization control channel processes signals from four vibration measurement points within this cycle, and performs the following operation on the first vibration measurement point. Filtered vibration signal at each measuring point To perform event identification, the collision event count metric is defined as follows:

[0135] ,

[0136] In the formula: For the first Within the first optimization update cycle, the first Collision event count index at the measuring point; The number of candidate vibration events identified; For the first The characteristic amplitude of each candidate event; For the first Event recognition threshold for measurement points; It is an indicator function; It only indicates the number of vibration events that meet the criteria, and is not equivalent to the count of actual collisions that are not calibrated;

[0137] No. The vibration energy index of the measuring point is defined as follows:

[0138] ,

[0139] In the formula: It is a dimensionless vibrational energy index; This is a reference value for vibration acceleration, in units of... ; For the pre-processed first Vibration acceleration at the measuring point, in units of , For integration variables, corresponding to the time range This index is used to characterize the vibration intensity at the measuring point and is not equivalent to the contact energy in the discrete element method.

[0140] The four measurement point indices are synthesized after normalization as follows:

[0141] ,

[0142] In the formula: and These are the normalized collision event count index and the normalized vibration energy index, respectively. and The first Reference values ​​for measuring points; For the first The weight of the measurement points ;

[0143] Closed-loop performance metrics are defined as follows:

[0144] ,

[0145] In the formula: For the first The closed-loop performance index of each optimization and update cycle; the smaller the value, the better the overall performance. , and These are the weights for the collision event count, vibration energy, and pressure tracking term, respectively. This is a normalized reference value for pressure, in units of The closed-loop performance index unifies collision frequency, collision energy, and pressure tracking performance into a single scalar evaluation index, which serves as the objective function for runtime optimization.

[0146] The pressure constraint function is defined as:

[0147] ,

[0148] In the formula: This is the pressure constraint function; The upper limit of the allowable pressure control deviation, in units of ;when When the pressure constraint is met, it indicates that the optimization update cycle satisfies the pressure constraint. The pressure constraint function introduces the safety of the delivery pressure as a hard constraint into the operation point optimization to prevent pressure exceeding the limit or even pipeline blockage due to the pursuit of low collision indicators.

[0149] The run-point optimization control channel employs one-dimensional safe Bayesian run-point optimization, establishing Gaussian process surrogate models for both closed-loop performance indices and pressure constraint functions; let Let represent the objective function and the constraint function, respectively. Then their posterior mean and posterior variance are:

[0150] ,

[0151] ,

[0152] In the formula: For the first Each optimization update cycle is at the valve position. The posterior mean at the location; This represents the number of existing samples; Valve position The kernel function vector between the kernel function and existing sample points; The kernel matrix of the existing sample points; To observe the noise variance; It is the identity matrix; This is a vector of existing observations; The posterior standard deviation; Valve position The prior kernel function value at the valve position; the posterior standard deviation is used to characterize the prediction uncertainty of the surrogate model at that valve position.

[0153] S5: Update the objective function surrogate model, the pressure constraint surrogate model, the safety set, and the reachable safety set. Select the nominal valve position for the next optimization update cycle within the reachable safety set. The confidence bounds for the objective function and constraint functions are defined as follows:

[0154] ,

[0155] In the formula: and These are the lower confidence bound and the upper confidence bound, respectively. For the first Confidence coefficient for each optimization update cycle;

[0156] Take in the constraint direction Under safe conditions, the safe set is defined as:

[0157] ,

[0158] In the formula: For the first A safety set for each optimization and update cycle; valve position set ; The upper confidence bound is used for the pressure constraint proxy model. By using the upper confidence bound judgment, candidate valve positions that meet the pressure constraints can be screened while taking into account the model's uncertainty.

[0159] The security correction set and the reachable security region are defined as follows:

[0160] ,

[0161] ,

[0162] In the formula: For the first A set of security fixes for each optimization and update cycle; This is the candidate nominal valve position correction amount; Optimize the maximum allowable nominal valve position change during the update cycle; It is the safe reachable range, which is the set of adjustable safety valve positions; , and The meaning is the same as before. The safety correction set limits the nominal valve position change within each optimization update cycle to... Within this range, large jumps in the nominal valve position are avoided; the reachable safety range is the intersection of the safety set and the reachable range of the safety correction, ensuring that the updated nominal valve position satisfies the pressure constraint and can be reached smoothly.

[0163] Within the reachable safety set, using the lower confidence bound of the objective function as the acquisition function, the next nominal valve position can be determined by the following formula:

[0164] ,

[0165] In the formula: The nominal valve position for the next optimization update cycle; This indicates the valve position at which the acquisition function within the parentheses reaches its minimum value; and These are the posterior mean and posterior standard deviation of the closed-loop performance index proxy model, respectively; this control method balances the current performance prediction and the exploration value of unknown valve positions within a safe range; if If no candidate valve position meets the update conditions, then maintain... The nominal valve position gradually shifts under safety monitoring, amplitude limiting, and valve position command rate constraints, while the pressure feedback control channel continues to suppress instantaneous pressure fluctuations during the shift.

[0166] S6: Repeat steps S3 to S5. When the pressure control deviation, nominal valve position change, and closed-loop performance index change continuously satisfy the convergence condition shown in the following formula, the system is determined to have entered steady-state operation:

[0167]

[0168] In the formula: The nominal valve position convergence threshold; This is the convergence threshold for the closed-loop performance index. , , and The meaning is the same as before. After entering steady state, the pressure feedback control channel still operates as before. The running point optimization control channel still follows the... Check the conveying conditions; when the conditions change, update the nominal valve position and converge to the new safe operating point.

[0169] The technical effects of the present invention will be further explained below with reference to simulation experiments.

[0170] This embodiment takes corn transportation as an example and adopts a three-factor Box-Behnken design. , and A three-level combination was performed, with a total of 15 structural simulations including 12 edge-midpoint scenarios and 3 center-point scenarios. The 12 edge-midpoint scenarios used a fixed particle generation sequence to reduce the impact of random particle generation on the simulation results; the 3 center-point scenarios were used to estimate the discreteness of repeated calculations, and independent random sequences were used when the particle generation process contained random factors. The ordinary bend was a comparative structure without bypass pipe 4, reduction section 21, and regulating valve 5, under the same main pipe inner diameter, bend centerline curvature radius, inlet straight pipe section, outlet straight pipe section, and boundary conditions. Compared to the ordinary bend, the bend structure of this invention is also called a novel bend. The simulation showed that the pressure drop, contact force, and other indicators fluctuated by less than 5% after 0.4s; therefore, 0.4s to 1s was used as the steady-state time window for data analysis. The factor levels of the structural simulation are shown in Table 1.

[0171] Table 1. Box-Behnken Design Factors and Levels

[0172]

[0173] To simplify the simulation process and focus on the particle-bend tube collision effect, referencing existing literature, corn kernels are equivalent to spherical particles of the same size. Detailed simulation parameters are shown in Table 2.

[0174] Table 2 Simulation Parameter Table

[0175]

[0176] The trajectory and velocity distribution of particles in a typical bend in a pipe are as follows: Figure 5 As shown, the particle motion trajectory and velocity distribution inside the novel curved pipe are as follows: Figure 6As shown, this diagram visually compares the aggregation and dispersion of particles within the bend section 2 and the straight section 3. Under the same conditions, particles in a conventional bend are significantly concentrated towards the outer arc sidewall due to inertia. This invention, through bypass airflow, causes the lower-velocity portions of the particles to shift earlier in bend section 2, while the higher-velocity portions interact with the airflow-particle mixing layer. This result demonstrates that the bypass self-supplying air structure can alter the localized concentrated collision state at the particle trajectory level.

[0177] The average normal contact force and average tangential contact force are the normal and tangential components of the contact force along the contact direction during particle-pipe wall interaction, and can well reflect the intensity of particle-pipe wall interaction. The average normal contact force is an important factor affecting bend erosion, and the average tangential contact force is an important factor affecting bend abrasion; both are positively correlated with corn breakage rate. In this experiment, the average normal contact force and average tangential contact force were statistically analyzed based on a single complete particle-pipe wall collision. The next collision has:

[0178] ,

[0179] In the formula: and The first The average normal contact force and average tangential contact force during the duration of the collision, in units of ; and These are the instantaneous normal contact force and the instantaneous tangential contact force, respectively. and These are the start and end times of the collision, respectively. The duration of the collision.

[0180] In the The output time of each result represents the time period during which the results were completed. The average value is then taken from each collision:

[0181] ,

[0182] In the formula: and The first The output results show the average normal and average tangential contact forces between the particle and the tube wall at each time point. This represents the number of particle-tube wall collisions that occurred during that period. Figure 7 , Figure 8Based on this, the contact force levels of ordinary bends and novel bends were compared within a steady-state statistical time window. Within the steady-state time window, compared to ordinary bends, the average normal contact force of the present invention was reduced by 41.56%, and the average tangential contact force was reduced by 45.55%, with several higher peak values ​​observed in ordinary bends being suppressed. This indicates that the bypass airflow not only reduces the contact effect when particles impact the pipe wall forward, but also weakens the tangential contact effect when particles slide along the pipe wall, thereby simultaneously reducing the risk of particle breakage and pipe wall erosion.

[0183] Cumulative contact energy characterizes the cumulative energy generated by normal impact and tangential friction during particle-to-pipe wall contact, i.e., the work done by the particle on the pipe wall. Under the same conditions, a larger cumulative contact energy on a pipe wall geometric unit indicates stronger cumulative impact and friction in that area, and can be used to characterize the relative trend of the maximum wear depth; simultaneously, this index can indirectly reflect the cumulative load and deformation damage level of particles during repeated collisions. Based on the relative wear model, the first... The cumulative contact energy on a pipe wall geometric element is defined as the sum of the normal cumulative contact energy and the tangential cumulative contact energy:

[0184]

[0185] In the formula: For a moment No. The cumulative contact energy on a single pipe wall geometric unit, in units of ; and The first The normal contact force and normal relative velocity acting on the geometric element at each time step satisfy the following conditions during the normal loading phase: ; and The first The tangential contact force and tangential relative velocity acting on the geometric unit at each time step; is the time step of the discrete element method.

[0186] Specify pipe wall statistics domain The maximum cumulative contact energy within is:

[0187]

[0188] In the formula: For a moment The maximum cumulative contact energy of each geometric unit within the statistical domain of the pipe wall, in units of ; This is the statistical domain of the pipe wall of bend 2. Figure 9 Comparison of two types of pipe bends Changes over time. From Figure 9It can be seen that the maximum cumulative contact energy of both types of bends increases approximately continuously over time, while that of the ordinary bend... The time was 24.57 The present invention is 5.99 at the same time. The decrease was 75.62%. The significant decrease in maximum cumulative contact energy indicates that the cumulative effect on the main impact zone of the particles was effectively weakened.

[0189] For the The total energy loss in a single, complete collision is defined as:

[0190]

[0191] In the formula: For the first The total energy loss from the collision, in units of ; and These are the normal and tangential dissipative contact forces, respectively, in units of ; and These are the normal and tangential contact deformation rates, respectively, in units of ; The duration of the collision; and These represent the energy losses in the normal and tangential directions, respectively.

[0192] The total energy loss in a single collision is divided into several intervals. The corresponding collision frequency is:

[0193]

[0194] In the formula: For the first Energy loss interval The collision frequency within the area, expressed in times; To count the total number of particle-tube wall collisions completed within the time window; and These are the lower and upper limits of the energy loss range, respectively. Figure 10 The frequency distribution of energy loss in the two types of bends was compared using a logarithmic ordinate. Figure 10As can be seen, the total number of collisions in this invention is 26,621, higher than the 17,838 in ordinary bends. However, the newly added collisions are mainly concentrated in the lowest energy loss range, with an increase of 10,928, greater than the overall increase of 8,783. Meanwhile, the collision frequency in most higher energy loss ranges is lower than that in ordinary bends. Therefore, this result should not be generalized as a decrease in the total number of collisions, but rather interpreted as the bypass airflow transforming localized high-intensity collisions into a larger number of contact processes with lower energy, causing the collision energy loss distribution to shift towards lower value ranges. When evaluating structural performance, the average contact force and maximum cumulative contact energy should be the primary criteria, rather than simply comparing the total number of collisions.

[0195] The termination time of 15 sets of structural simulations The maximum cumulative contact energy is used as the response value, and a quadratic regression model is fitted:

[0196] ,

[0197] In the formula: This represents the predicted response value of the quadratic regression model; The coefficients of the constant term; , and The coefficients of the linear terms; , and The coefficients of the interaction terms; , and The coefficient of the quadratic term.

[0198] Figure 11 for =4.00 and right The interaction effects. The response surface shows a clear reversal: when At a low level, Follow The magnitude of the change is small; with Increase, improve Able to make Significantly reduced. Conversely, when When at a low level, increase It will make Increase; when When at a higher level, increase Then make Decrease. The direction of the effect of the two factors changes with the level of the other factor, indicating that... and There is a strong tendency for interaction between them. The low-response region under this combination is mainly distributed in the higher... and higher On one side, where the diameter reduction is relatively weak and the straight section 44 of the bypass pipe 4 is narrower than the diameter reduction section 21, the cumulative contact effect of the main impact zone of the particles is smaller.

[0199] Figure 12 for =3.00 and right The interaction effects are most pronounced in this surface, particularly in its twisting and slope variations. When smaller, Follow Increases and decreases significantly, with the low response region moving towards high. ,Low directional extension; when After enlarging, right The direction of action changes, high Under the conditions On the contrary Increases and rises rapidly. Correspondingly, at low... Under the condition, change The resulting response changes are relatively slow; at high conditions, The slope of the curved surface in the direction increases significantly, and the contour lines become denser. The above results indicate that the effect of the axial spacing of the bypass connection depends on the diameter reduction ratio; both together determine the static pressure relationship between the inlet end 41 and the outlet end 45, as well as the bypass airflow loss along the path. (Higher) Must be with lower Only through proper coordination can a lower maximum cumulative contact energy be achieved; simply increasing... Without restriction It cannot consistently obtain a low response.

[0200] Figure 13 for =0.80 and right The interaction between the surfaces is significant. The surface is generally convex with an arc-shaped bulge, forming a relatively high response region in the middle, while the lowest response is located at a high level. ,Low One side. When When smaller, Follow Increase and decrease significantly; with Increase As the slope of the curved surface gradually decreases in the direction of travel, its lowering effect weakens. When At a high level, Follow Increase and rise; when When smaller, The resulting response change is relatively gradual. Therefore, the matching relationship between the inner diameter of the reduced diameter section 21 and the straight section 44 of the bypass pipe 4 needs to be determined in coordination with the axial spacing of the bypass connection. Larger... This is beneficial for achieving a lower response at shorter bypass connection axial spacing, but when If it is too large, this advantage will be weakened.

[0201] Compare Figures 11 to 13 It can be seen that, and The degree of surface torsion and the magnitude of response change are most prominent in the combination. and The combinations also show a strong tendency for interaction. and The surface changes of the combination are relatively gentle. The low response directions indicated by the three sets of response surfaces are consistent, that is... Overall and To a higher level It decreases towards lower level combinations. Because... , and The effect on the maximum cumulative contact energy exhibits significant nonlinearity, and the response surface shows significant torsion at different factor levels, indicating a strong interaction among structural factors. The single-control variable method for optimizing structural parameters cannot accurately reflect the cross-influence of various factors; therefore, the response surface optimization method is used to find the optimal parameters. The final optimized results for the three structural parameters are as follows: =0.85, =3.22, =2.98. Under the same conditions, simulation analysis was performed using the optimized structural parameters, and the results were as follows: =4.34J, which is the smallest known value in Box-Behnken designs. The collision frequency decreased further by 4.4% to 23,376 times, remaining at a low level.

Claims

1. A bypass self-supplying bend structure for pneumatic conveying, comprising a straight inlet pipe section (1), a bend pipe section (2), and a straight outlet pipe section (3) connected in sequence, characterized in that, It also includes a bypass pipe (4), the inlet end (41) of which is located on the side wall of the straight pipe section (1) and connected to it, and the outlet end (45) of which is located on the outer arc side of the bend section (2) and connected to it. The intersection point of the axis of the straight pipe section (1) and the inner wall of the outer arc side of the bend section (2) is located in the middle of the intersection area of ​​the bypass pipe (4) and the bend section (2). The airflow from the outlet end (45) of the pipe (4) into the bend section (2) is tangent to the inner wall of the outer arc side of the bend section (2); the inner diameter of the inlet straight pipe section (1) and the outlet straight pipe section (3) are the same, the inner diameter of the bend section (2) is smaller than the inner diameter of the inlet straight pipe section (1) and the outlet straight pipe section (3), and the connection between the inner wall of the bend section (2) and the inner wall of the inlet straight pipe section (1) and the inner wall of the outlet straight pipe section (3) is connected by a bevel transition.

2. The bypass self-supplying air bend structure for pneumatic conveying as described in claim 1, characterized in that, The bypass pipe (4) consists of a variable diameter section (42), an arc section (43), and a straight section (44) that are smoothly connected from the air inlet end (41) to the air outlet end (45). The inner diameter of the variable diameter section (42) gradually decreases from the air inlet end (41) to the arc section (43). The inner diameters of the arc section (43) and the straight section (44) are the same as the inner diameter of the small end of the variable diameter section (42).

3. The bypass self-supplying air bend structure for pneumatic conveying as described in claim 2, characterized in that, The section of the bend (2) connected to the outlet (45) of the bypass pipe (4) has a reduced diameter section (21), the inner diameter of which is smaller than the inner diameter of the rest of the bend (2).

4. The bypass self-supplying air bend structure for pneumatic conveying as described in claim 3, characterized in that, The inner diameter of the reduced diameter section (21) in the bend section (2) is The inner diameter of the main pipe in the remaining parts of the bend section (2) is... The inner diameter of the straight section (44) of the bypass pipe (4) is The distance between the center of the air inlet (41) and air outlet (45) of the bypass pipe (4) projected onto the axis of the straight inlet pipe section (1) is: It satisfies the following relationship: ， in, The value is 0.75~0.

85. It is 2.5~3.

5. The value is 2.5~5.

5.

5. The bypass self-supplying air bend structure for pneumatic conveying as described in claim 4, characterized in that, It is 0.

85. It is 3.

22. It is 2.

98.

6. The bypass self-supplying air bend structure for pneumatic conveying as described in any one of claims 1-5, characterized in that, The bypass pipeline (4) is equipped with a regulating valve (5).

7. The bypass self-supplying air bend structure for pneumatic conveying as described in claim 6, characterized in that, A pressure sensor (6) is provided at the air inlet end (41) of the straight inlet pipe section (1) facing the bypass pipe (4). Multiple vibration sensors (7) are provided on the outer arc side and inner arc side of the bend pipe section (2) from the part connected to the bypass pipe (4) to the part connected to the straight outlet pipe section (3). The regulating valve (5) is an electric ball valve. The pressure sensor (6) and all vibration sensors (7) are electrically connected to the controller of the electric ball valve.

8. The bypass self-supplying air bend structure for pneumatic conveying as described in claim 7, characterized in that, The controller of the electric ball valve is configured to include a pressure feedback control channel based on a pressure sensor (6) and an operating point optimization control channel based on a vibration sensor (7), as well as an instruction synthesis and execution channel; the pressure feedback control channel uses a sampling period of During operation, based on the deviation between the pressure setpoint and the pressure feedback value of the pressure sensor (6), a dynamic valve position compensation value is generated; the operating point optimization control channel optimizes the update cycle. run, The collision state evaluation index and closed-loop performance index are calculated based on the pressure feedback value and the vibration signal detected by the vibration sensor (7), and the nominal valve position is optimized and updated. The command synthesis and execution channel synthesizes the nominal valve position and the valve position dynamic compensation amount into a synthesized valve position command to drive the actuator of the electric ball valve.

9. A method for controlling the air supply of a bypass self-supplying bend structure used in pneumatic conveying, characterized in that, The bypass self-supplying bend structure for pneumatic conveying as described in claim 8 includes the following steps: S1: Set the pressure setpoint, sampling period, optimization update period, allowable valve position set and pressure control deviation upper limit, select the initial nominal valve position from the known safety valve positions, and establish the initial Gaussian process proxy model of the closed-loop performance index and pressure constraint function; S2: After each sampling cycle, read the pressure feedback value of the pressure sensor and calculate the pressure control deviation. The filter super-torsion controller generates the valve position dynamic compensation amount, which is combined with the current nominal valve position to form a composite valve position command. S3: Implements safety supervision, mode switching, valve position command amplitude limiting, and valve position command rate limitation on the synthesized valve position command, generates a restricted valve position command and drives the electric ball valve to reach the actual valve position, changes the flow rate of the bypass pipeline, and corrects the pressure feedback value. S4: Record the vibration signals of each vibration sensor. After each optimization update cycle, calculate the collision event count index, vibration energy index, closed-loop performance index and pressure constraint function, and add the observation results of this cycle to the Gaussian process sample set. S5: Update the objective function proxy model, pressure constraint proxy model, safety set, and reachable safety set, and select the nominal valve position for the next optimization update cycle within the reachable safety set; S6: Repeat steps S3 to S5. When the pressure control deviation, nominal valve position change, and closed-loop performance index change continuously meet the convergence conditions, it is determined that the system has entered steady-state operation. When the operating conditions change, the nominal valve position is updated again and converged to a new safe operating point.

10. The air supply control method for the bypass self-supplying bend structure for pneumatic conveying as described in claim 9, characterized in that, In step S2, the pressure control deviation is defined as: ， In the formula: The pressure control deviation at the k-th sampling time is expressed in units of... ; Pressure setpoint, unit: ; The pressure feedback value obtained by the pressure sensor at the k-th sampling time is expressed in units of... ; A filtered over-torsion controller with a first-order filtering state is used to generate the valve position dynamic compensation quantity, which is in continuous form according to the sampling period. Discretize it, and write it as: ， In the formula: In the filtering state, The rate at which the filter state changes over time; For filter gain, ; It is a symbolic function; The valve position dynamic compensation is generated by the following formula: ， In the formula: This is the dynamic compensation amount for the valve position; This is a super-distorted integral state. It is an integral state. The rate of change over time; and To control the gain, and , The values ​​of each gain are determined based on the actual valve position's effect on the pressure feedback value and the dynamic response. Limit the dynamic compensation amount of the valve position: ， In the formula: This is the dynamic compensation amount for the valve position after amplitude limiting; Valve position compensation limit; This indicates that the input is restricted to a range. Inside; In the Within each optimization update cycle, the synthesized valve position command is: ， In the formula: This is a synthesized valve position command; For the first The nominal valve position used in each optimization update cycle; In step S3, the synthesized valve position command is sequentially subjected to valve position command amplitude limiting and valve position command rate limiting: ， ， In the formula: For valve position commands with limited amplitude; and These are the minimum and maximum allowable valve positions, respectively. This is a restricted valve position command; The maximum permissible rate of valve position change, in units of ; The sampling period is expressed in units of 10 ... Electric ball valve according to Make the actual valve position follow the restricted valve position command; In step S4, the first The optimization update cycle is: ， In the formula: For the first The time range of each optimization update cycle; Four vibration sensors were installed at selected measuring points on the bend section. The operation point optimization control channel processed the signals from each vibration sensor within this cycle. Filtered vibration signal at each measuring point To perform event identification, the collision event count metric is defined as follows: ， In the formula: For the first Within the first optimization update cycle, the first Collision event count index at the measuring point; The number of candidate vibration events identified; For the first The characteristic amplitude of each candidate event; For the first Event recognition threshold for measurement points; It is an indicator function; It only indicates the number of vibration events that meet the criteria, and is not equivalent to the uncalibrated count of actual collisions; No. The vibration energy index of the measuring point is defined as follows: ， In the formula: It is a dimensionless vibrational energy index; This is a reference value for vibration acceleration, in units of... ; For the pre-processed first Vibration acceleration at the measuring point, in units of , For integration variables, corresponding to the time range ; The four measurement point indices are synthesized after normalization as follows: ， In the formula: and These are the normalized collision event count index and the normalized vibration energy index, respectively. and The first Reference values ​​for measuring points; For the first The weight of the measurement points ; Closed-loop performance index is defined as: ， In the formula: For the first The closed-loop performance index of each optimization and update cycle; the smaller the value, the better the overall performance. , and These are the weights for the collision event count, vibration energy, and pressure tracking term, respectively. This is a normalized reference value for pressure, in units of ; The pressure constraint function is defined as follows: ， In the formula: This is the pressure constraint function; The upper limit of the allowable pressure control deviation, in units of ;when When this occurs, it indicates that the optimization update cycle meets the pressure constraint; The run-point optimization control channel employs one-dimensional safe Bayesian run-point optimization, establishing Gaussian process surrogate models for both closed-loop performance indices and pressure constraint functions; let Let represent the objective function and the constraint function, respectively. Then their posterior mean and posterior variance are: ， ， In the formula: For the first Each optimization update cycle is at the valve position. The posterior mean at the location; This represents the number of existing samples; Valve position The kernel function vector between the kernel function and existing sample points; The kernel matrix of the existing sample points; To observe the noise variance; It is the identity matrix; This is a vector of existing observations; The posterior standard deviation; Valve position The prior kernel function value at that position; the posterior standard deviation is used to characterize the prediction uncertainty of the surrogate model at that valve position; In step S5, the confidence bounds of the objective function and the constraint functions are defined as follows: ， In the formula: and These are the lower confidence bound and the upper confidence bound, respectively. For the first Confidence coefficient for each optimization update cycle; Take in the constraint direction Under safe conditions, the safe set is defined as: ， In the formula: For the first A safety set for each optimization and update cycle; valve position set ; The upper confidence bound is defined for the pressure-constrained surrogate model; the upper confidence bound judgment is used to screen candidate valve positions that satisfy the pressure constraints while considering model uncertainties. The security correction set and the reachable security region are defined as follows: ， ， In the formula: For the first A set of security fixes for each optimization and update cycle; This is the candidate nominal valve position correction amount; Optimize the maximum allowable nominal valve position change during the update cycle; It is the safe reachable range, which is the set of adjustable safety valve positions; , and The meaning is the same as before; Within the reachable safety set, using the lower confidence bound of the objective function as the acquisition function, the next nominal valve position can be determined by the following formula: ， In the formula: The nominal valve position for the next optimization update cycle; This indicates the valve position that causes the data acquisition function within the parentheses to reach its minimum value. ; and These are the posterior mean and posterior standard deviation of the surrogate model for closed-loop performance indices, respectively; this control method balances current performance prediction and the exploration value of unknown valve positions within a safe range; if If no candidate valve position meets the update conditions, then maintain... .

Images