Hydraulic fracturing vehicle power matching control method based on dynamic programming and genetic algorithm

Abstract

The present application relates to a kind of hydraulic fracturing truck power matching control method based on dynamic programming and genetic algorithm, it includes the following steps, step one: the whole vehicle oil consumption prediction model of full hydraulic fracturing truck based on BP neural network is established;Step two: using dynamic programming algorithm determines the number of variable motor and engine, determines solving interval;Step three: using adaptive genetic algorithm to the displacement of variable motor and the displacement of variable pump are optimized, determine optimal efficiency point;Step four: the control parameter of fracturing truck is obtained to realize the global power matching control of fracturing truck.The present application is based on power transmission route, considers power confluence process, uses backward power matching mode, under the premise of guaranteeing dynamic performance and load requirement, accurate oil consumption prediction model is established using BP neural network, the control parameter of fracturing truck is optimized using dynamic programming algorithm and adaptive genetic algorithm, finally realizes the global power matching control of fracturing truck.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of energy consumption reduction of engineering machinery, in particular to a hydraulic fracturing vehicle power matching control method and control system based on dynamic programming and a genetic algorithm. BACKGROUND

[0002] The fracturing vehicle is high-end fracturing equipment developed according to the geological conditions of oil and gas exploitation and the requirements of fracturing construction technology. The function is to inject high-pressure and large-displacement fracturing fluid into the well, press the stratum, extrude the proppant into the fracture, and improve the permeability of the oil layer to increase the water injection or oil production. The fracturing vehicle has the characteristics of large power, large displacement, high oil consumption and long-time operation like other engineering machinery. In recent years, with the increasing demand for energy, the global unconventional oil and gas exploration and development has entered an active period, and the Ministry of Land and Resources of China considered that the unconventional oil and gas resources such as shale oil and gas, oil sands and oil shale are rich in reserves in the 2009 national unconventional oil and gas resource evaluation work. With the further deepening of exploration and development, the demand for unconventional fracturing equipment in China is increasing, and the research on the fracturing vehicle has also attracted the attention of manufacturers.

[0003] Currently, existing patents on engineering machinery research mainly focus on "pump-engine" matching. Yang Bo, in his work "Power Matching of a Fully Hydraulic Fracturing Truck Based on the MFO Algorithm," addresses the high fuel consumption and cost of mechanical fracturing trucks by proposing the concept of a fully hydraulic fracturing truck. Considering system power loss, he proposes using "operational specific fuel consumption" to measure the actual fuel consumption of the fracturing truck and performs global power matching for the fully hydraulic fracturing truck. Mathematical models are established for the engine's universal characteristics, variable plunger pump efficiency, and overall auxiliary power. An adaptive penalty function method is used to construct a penalty function, and the objective function is established with the lowest operational specific fuel consumption as the optimization goal. Based on the MFO algorithm, the required output pressure and flow rate of the fracturing pump are used as optimization input parameters to optimize the optimal combination of 11 adjustment parameters, including the number of engines to be started, the engine speed, and the plunger pump displacement. This method mainly optimizes the local power matching of the "pump-engine" relationship and does not perform global power matching optimization, leaving room for further fuel consumption reduction. Sany Petroleum Intelligent Equipment Co., Ltd. proposed a fracturing truck displacement control method in its paper "A Fracturing Truck Displacement Control Method and Fracturing Truck". This method sets the displacement of the variable pump to a first displacement value and the displacement of the motor to a second displacement value based on the target displacement value of the fracturing truck. The second displacement value is greater than or equal to the first displacement value. This prevents shocks caused by displacement mismatch when the variable pump switches, thus improving the stability of the hydraulic system during displacement control or adjustment. The actual displacement value of the fracturing truck is obtained based on the first and second displacement values ​​and adjusted to the target displacement value. The main function of this control method is to maintain the stability of the hydraulic system; it does not consider the excessive fuel consumption and emissions caused by power mismatch under different operating conditions. Shu Fuhua proposed a genetic neural network control method for a variable pump in a hydraulic bulldozer in his paper "A Bulldozer Variable Pump Control System Based on Genetic Neural Network". Under a set throttle position, the control system adjusts the displacement of the variable pump in real time according to changes in the external load of the engine, thereby achieving a reasonable match between the engine and the variable pump. Although an optimization algorithm was used, it only optimized the local power matching of the "pump-engine" and did not perform global power matching optimization.

[0004] In summary, existing global power matching control methods for fully hydraulic fracturing trucks are rare, and their control effectiveness and technical shortcomings need to be improved. Therefore, this paper proposes a power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention is based on the power transmission path, considers the power merging process, and adopts a backward power matching method. Under the premise of ensuring power and load requirements, an accurate fuel consumption prediction model is established using a BP neural network. Dynamic programming algorithm and adaptive genetic algorithm are used to optimize the power matching control parameters of the fracturing truck and determine the control parameters of the fracturing truck. Finally, the power matching control of the fracturing truck is realized.

[0006] To achieve the above objectives, the solution adopted by the present invention is as follows:

[0007] A power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms includes the following steps:

[0008] Step 1: Establish a fuel consumption prediction model for the fully hydraulic fracturing truck based on a BP neural network;

[0009] A fuel consumption prediction model for a fully hydraulic fracturing truck was established. The overall structure of the fully hydraulic fracturing truck, with its multi-pump and multi-motor hydraulic transmission system, was determined. Speed ​​and power transmission models for the hydraulic system of the fully hydraulic fracturing truck were established. Influence factor models for variable motor and variable pump efficiency and engine fuel consumption rate were also established. The mathematical model for predicting the overall fuel consumption of the fully hydraulic fracturing truck, determined based on the engine speed and torque, is shown below:

[0010]

[0011] In the formula: G t p1 represents the fuel consumption of the entire vehicle per unit time; q1 represents the output pressure of the fracturing pump; b represents the output flow rate of the fracturing pump; e Indicates the engine's fuel consumption rate; η Y Indicates the fracturing pump efficiency; η m Indicates the efficiency of a variable motor; η p Indicates the efficiency of a variable pump; n e This indicates the engine speed; n4 indicates the engine speed of the fracturing truck; T4 indicates the engine torque of the fracturing truck.

[0012] The method for obtaining the engine speed and torque of the fully hydraulic fracturing truck is as follows:

[0013]

[0014] Where: n m Indicates the number of motors in operation; V m V represents the displacement of a variable motor. Y η represents the displacement of the fracturing pump; i1 represents the ratio of the variable displacement motor to the fracturing pump drive; Vp represents the displacement of the variable displacement pump; i2 represents the ratio of the engine drive to the variable displacement pump drive; η represents the displacement of the variable displacement pump; i1 represents the ratio of the variable displacement motor to the variable displacement pump drive; η represents the displacement of the variable displacement pump; i2 represents the displacement of the variable displacement pump; i2 represents the displacement of the engine drive; i3 represents the displacement of the variable displacement pump; i4 represents the displacement of the variable displacement pump; i5 represents the displacement of the variable displacement pump; i6 represents the displacement of the variable displacement pump; i7 represents the displacement of the variable displacement motor; i8 represents the displacement of the variable displacement pump; i9 represents the displacement of the variable displacement pump; i1 represents the ratio of the engine drive to the variable displacement pump drive; i2 represents the displacement of the variable displacement pump; i2 represents the displacement ... Y1Indicates the volumetric efficiency of the fracturing pump; η m1 Indicates the motor's volumetric efficiency; η p1 Indicates the efficiency of the variable pump;

[0015] The efficiency-affecting parameters are input, and the current efficiency of the variable motor, variable pump, and engine fuel consumption are predicted by a trained BP neural network.

[0016] Step 2: Use dynamic programming algorithm to determine the number of motors and engines, and determine the solution interval;

[0017] Step 21: Determine the objective function for the optimization process; In the optimization process of fracturing truck operation, the control mode with the lowest hourly fuel consumption is the optimal mode under the current working condition. The objective function for the minimum fuel consumption of fracturing truck operation is as follows:

[0018]

[0019] In the formula: F represents the total fuel consumption of the engine when the fracturing truck is in operation; T represents the operating time of the fracturing truck;

[0020] Step 22: Determine the dynamic programming model; the number of engines engaged ranges from 1 to all engines operating, with each engine engaged representing a stage. Calculate the maximum flow rate that can be output under the current operating conditions when each engine carries two pumps; the state transition equation of the dynamic programming model is as follows:

[0021]

[0022] In the formula: f k+1 (Q k+1 ) represents the minimum fuel consumption generated when the flow is distributed to the pumps on the 1st to (k+1th)th engines; f k (Q k ) represents the minimum fuel consumption generated when the flow is distributed to the pumps on engines 1 through k; g k+1 (q k+1 ) represents q k+1 The minimum fuel consumption obtained by distributing the flow rate to the pump on the (k+1)th engine; Q k This represents the flow rate output of the pumps driven by the 1st to the kth engines, where 0 < Q. k ≤k·q;q k This represents the flow rate allocated to the pump on the k-th engine, 0 ≤ q k ≤q;

[0023] Step 23: Construct the algorithm optimization process, obtain the optimal number of motors to be put into operation for variable motors and the optimal number of motors to be started for engines, and provide the recommended displacement of motors and pumps for controlling variable motors and variable pumps to provide a reference range for the initial population setting of the genetic algorithm;

[0024] Step 3: Use an adaptive genetic algorithm to optimize the displacement of the variable motor and the variable pump to determine the optimal efficiency point;

[0025] First, calculate the crossover and mutation probabilities in the adaptive genetic algorithm; then, based on the criterion for determining the degree of population dispersion, arcsin(f ave / f max The distribution of the population is determined, and the crossover and mutation probabilities are iteratively updated. Finally, the most efficient variable motor displacement and variable pump displacement under the current operating conditions are determined.

[0026] Step 4: Obtain the control parameters of the fracturing truck to achieve global power matching control of the fracturing truck;

[0027] The control parameters of the fracturing truck include: the optimal engine speed determined in step 1, the number of variable motors engaged and the number of engines started determined in step 2, and the displacement of variable motors and the displacement of variable pumps determined in step 3; the global power matching control of the fracturing truck is achieved through the control parameters of the fracturing truck.

[0028] Preferably, the speed transmission model in step 1 refers to the speed transmission model of the hydraulic system of the fully hydraulic fracturing truck, as shown below:

[0029]

[0030] In the formula: n1 represents the fracturing pump speed; n2 represents the variable displacement motor speed; q 2m n3 represents the flow rate input to the variable motor; n3 represents the speed of the variable pump; q 2q This represents the flow rate output by the variable pump; n p This indicates the number of variable pumps in operation.

[0031] Preferably, the power transfer model in step 1 refers to the power transfer model of the hydraulic system of the fully hydraulic fracturing truck, as shown below:

[0032]

[0033] In the formula: P Y T1 represents the fracturing pump power; T2 represents the fracturing pump input torque; T2 represents the output torque of the single variable motor; P m p1 represents the total output power of the variable displacement motor; p2 represents the hydraulic system pressure; P H q represents the total power of the hydraulic system. 2pT3 represents the output flow rate of the variable pump; T3 represents the input torque of the variable pump; η p P represents the efficiency of a variable pump. p P represents the total input power of the variable pump. e This indicates the total output power of the engine.

[0034] Preferably, the efficiency of the variable motor and variable pump, and the engine fuel consumption rate in step 1 are as follows:

[0035] The variable motor efficiency η m The effects of hydraulic system pressure, motor speed, and motor displacement are as shown in the following formula:

[0036] η m =f1(p2,n2,V) m );

[0037] In the formula: f1 represents the functional relationship between the variable motor efficiency and the hydraulic system pressure, motor speed, and motor displacement; p2 represents the hydraulic system pressure;

[0038] The variable pump efficiency η p The effects of hydraulic system pressure, pump speed, and pump displacement are shown in the following formula:

[0039] η p =f2(p2,n3,V) p );

[0040] In the formula: f2 represents the functional relationship between the variable pump efficiency and the hydraulic system pressure, pump speed and pump displacement;

[0041] The engine's fuel consumption rate b e The influence of engine speed and engine torque is as shown in the following formula:

[0042] b e =f3(n4,T4);

[0043] In the formula: f3 represents the functional relationship between the engine's fuel consumption rate and the engine speed and engine torque.

[0044] Preferably, the optimization process of the construction algorithm in step 23 is as follows:

[0045] Step 31: Based on the current hydraulic system pressure, calculate the maximum flow rate that a single engine can output using the engine fuel consumption neural network and the variable pump efficiency neural network;

[0046] Step 32: Design the step size to decompose the flow rate that a single pump can output, solve for the optimal fuel consumption and pump displacement at each flow rate, and divide the calculation into 3 stages, taking the process of calculating one engine as one stage.

[0047] Step 33: Decompose the input flow rate Q with the designed step size, and use the flow range of the second engine as a variable to solve for the optimal fuel consumption and the displacement of each pump within the flow range that the two engines can output.

[0048] Step 34: Go through all stages to determine the optimal combination of pump displacements for the engine output flow rate Q.

[0049] Preferably, the method for obtaining the crossover probability and mutation probability in step 3 is as follows:

[0050]

[0051]

[0052] In the formula: P c P represents the crossover probability; m f represents the probability of mutation; ave f represents the average fitness of the population. max k1 represents the maximum fitness of the population; k2 represents the crossover probability weight; k3 represents the mutation probability weight.

[0053] Preferably, in step 3, the determination condition for the degree of population dispersion is arcsin(f ave / f max To determine the distribution of the population and iteratively update the crossover and mutation probabilities, the following steps are taken:

[0054] arcsin(f ave / f max As a criterion for determining the degree of population dispersion, its value is determined by the average fitness value f of the population. ave And the maximum fitness of the population f max Decide;

[0055] When arcsin(f) ave / f max When π / 6 < π / 6, it indicates that the population is dispersed. The smaller the value of π / 6, the easier it is to determine that the population is dispersed. At this time, the crossover probability is adaptively increased to allow chromosome genes to be fully exchanged and to evolve superior individuals. At the same time, the mutation probability is adaptively decreased to reduce the probability of destroying superior individuals and speed up the convergence speed.

[0056] When arcsin(f) ave / f maxWhen π / 6 ≥ π / 6, it indicates that the population distribution is concentrated. The larger π / 6 is, the easier it is to judge that the population distribution is concentrated. At this time, the population difference is not large and the population is homogeneous. In this case, the crossover probability value is adaptively reduced to prevent the destruction of existing high-quality genes. At the same time, the mutation probability value is adaptively increased to improve the global search ability of the population and prevent it from getting trapped in local optima.

[0057] The second aspect of this invention proposes a control system based on dynamic programming and genetic algorithm for power matching control of hydraulic fracturing trucks. This system can achieve the lowest fuel consumption under the current operating conditions by collecting the fracturing pump pressure output by the pressure sensor and the fracturing pump speed output by the speed sensor. The control system includes: a fracturing pump part, a hydraulic transmission part, an engine part, and a control system part.

[0058] The fracturing pump section includes a first pressure sensor, a fracturing pump, a first connecting shaft, a first speed sensor, and a fracturing pump gear. The first pressure sensor is installed at the outlet of the fracturing pump, the first speed sensor is installed near the first connecting shaft, and the fracturing pump gear and the fracturing pump are coaxially and rigidly connected through the first connecting shaft.

[0059] The hydraulic transmission system includes a variable motor gear set that meshes with the fracturing pump gear, a variable motor, a flow sensor, a high-pressure pipeline, a second pressure sensor, an accumulator, a variable pump, a variable pump gear set, a low-pressure pipeline, a third pressure sensor, a back pressure valve, and an oil tank. The fracturing pump and the variable motor of the hydraulic transmission system, as well as the engine and the variable pump of the hydraulic system, all employ gear transmission.

[0060] The engine section includes an engine gear reduction gearbox, a second speed sensor, a torque sensor, an engine, and a fuel consumption metering device;

[0061] The control system includes a signal input section and a control output section. The signal input section includes a first pressure sensor, a first speed sensor, a second pressure sensor, a third pressure sensor, a second speed sensor, a torque sensor, and a fuel consumption metering device. The control output section includes a vehicle controller, which contains a variable displacement motor control system, a variable motor activation quantity control system, a variable pump displacement control system, an engine activation quantity control system, and an engine speed control system.

[0062] Preferably, the control system of the fully hydraulic fracturing truck uses a hydraulic system for power confluence, employing 3 engines, each engine driving 2 parallel variable pumps, which together output high-pressure hydraulic oil. After being combined through high-pressure pipelines, the oil is output to 12 parallel variable motors to drive fracturing pumps to output high-pressure, high-flow fracturing fluid.

[0063] The variable displacement motor's oil inlet draws oil from the low-pressure pipeline and transmits it to the high-pressure pipeline through the variable displacement motor's oil outlet to output high-pressure oil. The variable displacement motor's oil inlet is connected to the high-pressure pipeline, and the variable displacement motor's oil outlet is connected to the low-pressure pipeline. A flow sensor is installed at the variable displacement motor's oil inlet. A second pressure sensor and an accumulator are installed in the middle of the high-pressure pipeline. A third pressure sensor is installed on the low-pressure pipeline. The first end of the back pressure valve is connected to the low-pressure pipeline, and the second end is connected to the oil tank.

[0064] The variable pump gear set is connected to the engine gear reducer via gear meshing;

[0065] Each engine is equipped with a second speed sensor and a torque sensor on its drive shaft, and a fuel consumption metering device to record the amount of fuel consumed per unit time when the engine is working.

[0066] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0067] (1) This invention uses the speed and power relationship of the fully hydraulic fracturing truck and the fuel consumption prediction model of the fully hydraulic fracturing truck constructed by the BP neural network toolbox to predict the actual fuel consumption of the fracturing truck under various parameters, thereby reducing the frequency of on-site testing, improving optimization efficiency, and shortening the optimization cycle.

[0068] (2) This invention uses dynamic programming algorithm and adaptive genetic algorithm to optimize the control of the five control parameters of the fracturing truck. The combination of power matching control variables is solved by optimization algorithm, and the coordinated control of each variable finally maximizes the efficiency of the whole machine.

[0069] (3) The control method of the present invention improves the accuracy of prediction and the speed of calculation by using multiple sets of constraints, such as engine external characteristic curve constraints, variable pump and variable motor speed range constraints, and hydraulic system maximum pressure constraints. Attached Figure Description

[0070] Figure 1 This is a flowchart of a power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms, according to an embodiment of the present invention.

[0071] Figure 2 This is a schematic diagram illustrating the overall intelligent collaborative optimization algorithm design principle of this invention.

[0072] Figure 3 This is a schematic diagram of the dynamic programming optimization algorithm in an embodiment of the present invention.

[0073] Figure 4 This is a schematic diagram of the genetic algorithm in an embodiment of the present invention;

[0074] Figure 5 This is a diagram illustrating the hydraulic principle and hardware configuration system of an embodiment of the present invention.

[0075] 1. First pressure sensor; 2. Fracturing pump; 3. First connecting shaft; 4. First speed sensor; 5. Fracturing pump gear; 6. Variable displacement motor gear set; 7-12, 17, 18. Variable displacement motor; 19. Flow sensor; 20. High-pressure pipeline; 21. Second pressure sensor; 22. Accumulator; 23-28. Variable displacement pump; 29-31. Variable displacement pump gear set; 32. Low-pressure pipeline; 33. Third pressure sensor; 34. Back pressure valve; 35-37. Engine gear reduction gearbox; 38-40. Second speed sensor; 41-43. Torque sensor; 44-46. Engine; 47-49. Fuel consumption metering device; 50. Fuel tank. Detailed Implementation

[0076] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

[0077] This invention, through its power transmission path and considering the power confluence process, employs a backward power matching method. While ensuring power and load requirements, it uses a BP neural network to establish an accurate fuel consumption prediction model and determine the optimal engine speed. Dynamic programming and adaptive genetic algorithms are used to optimize the control parameters of the fracturing truck, determining the optimal number of variable displacement motors and the optimal number of engine starts, as well as the optimal motor displacement and pump displacement. Finally, power matching control of the fracturing truck is achieved through these control parameters. Figure 1 The diagram shows a flowchart of a power matching control method for a hydraulic fracturing truck based on dynamic programming and genetic algorithms, according to an embodiment of the present invention. Figure 2 The diagram shown is a schematic diagram illustrating the overall intelligent collaborative optimization algorithm design principle of an embodiment of the present invention.

[0078] This invention provides a power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms. To demonstrate the applicability of this invention, it is applied to an example, specifically including the following steps:

[0079] S1: Establish a fuel consumption prediction model for the fully hydraulic fracturing truck based on BP neural network;

[0080] A fuel consumption prediction model for the fully hydraulic fracturing truck was established. The overall structure of the fully hydraulic fracturing truck, with its multi-pump and multi-motor hydraulic transmission system, was determined. Speed ​​transmission and power transmission models for the hydraulic system of the fully hydraulic fracturing truck were established. The speed transmission model refers to the speed transmission model of the hydraulic system of the fully hydraulic fracturing truck, as shown below:

[0081]

[0082] In the formula: n1 represents the fracturing pump speed; n2 represents the variable displacement motor speed; q 2m n3 represents the flow rate input to the variable motor; n3 represents the speed of the variable pump; q 2q This represents the flow rate output by the variable pump; n p This indicates the number of variable pumps in operation.

[0083] Table 1 shows an example of the initial parameters of a fracturing pump. The table details the names, symbols, and specific values ​​of the initial parameters.

[0084] Table 1. Example of initial parameters for fracturing pump variables.

[0085]

[0086] The power transfer model refers to the power transfer model of the hydraulic system of a fully hydraulic fracturing truck, as shown below:

[0087]

[0088] In the formula: P Y T1 represents the fracturing pump power; T2 represents the fracturing pump input torque; T2 represents the output torque of the single variable motor; P m p1 represents the total output power of the variable displacement motor; p2 represents the hydraulic system pressure; P H q represents the total power of the hydraulic system. 2p T3 represents the output flow rate of the variable pump; T3 represents the input torque of the variable pump; η p P represents the efficiency of a variable pump. p P represents the total input power of the variable pump. e This indicates the total output power of the engine.

[0089] Establish a model of the influencing factors of variable motor, variable pump efficiency, and engine fuel consumption rate; variable motor efficiency η m The effects of hydraulic system pressure, motor speed, and motor displacement are as shown in the following formula:

[0090] η m =f1(p2,n2,V) m );

[0091] In the formula: f1 represents the functional relationship between the variable motor efficiency and the hydraulic system pressure, motor speed and motor displacement; p2 represents the hydraulic system pressure.

[0092] Table 2 shows an example of the initial parameters for a variable displacement motor. The table details the parameter names, symbols, and initial parameters included in the variable displacement motor.

[0093] Table 2 Example of initial parameters for variable displacement motor

[0094]

[0095] Variable pump efficiency η p The effects of hydraulic system pressure, pump speed, and pump displacement are shown in the following formula:

[0096] η p =f2(p2,n3,V) p );

[0097] In the formula: f2 represents the functional relationship between the variable pump efficiency and the hydraulic system pressure, pump speed and pump displacement.

[0098] Engine fuel consumption rate b e The influence of engine speed and engine torque is as shown in the following formula:

[0099] b e =f3(n4,T4);

[0100] In the formula: f3 represents the functional relationship between the engine's fuel consumption rate and the engine speed and engine torque.

[0101] Finally, the mathematical model for predicting the vehicle's fuel consumption, determined based on the engine speed and torque of the fully hydraulic fracturing truck, is shown below:

[0102]

[0103] In the formula: G t p1 represents the fuel consumption of the entire vehicle per unit time; q1 represents the output pressure of the fracturing pump; b represents the output flow rate of the fracturing pump; e Indicates the engine's fuel consumption rate; η Y Indicates the fracturing pump efficiency; η m Indicates the efficiency of a variable motor; η p Indicates the efficiency of a variable pump; n e n4 represents the engine speed; n4 represents the engine speed of the fracturing truck; T4 represents the engine torque of the fracturing truck.

[0104] Table 3 shows an example of engine initial parameters, which details the names, symbols, and initial parameters of the engine.

[0105] Table 3 Example of Engine Initial Parameters

[0106]

[0107] The methods for obtaining the engine speed and torque of a fully hydraulic fracturing truck are as follows:

[0108]

[0109] Where: n mIndicates the number of motors in operation; V m V represents the displacement of a variable motor. Y η represents the displacement of the fracturing pump; i1 represents the ratio of the variable displacement motor to the fracturing pump drive; Vp represents the displacement of the variable displacement pump; i2 represents the ratio of the engine drive to the variable displacement pump drive; η represents the displacement of the variable displacement pump; i1 represents the ratio of the variable displacement motor to the variable displacement pump drive; η represents the displacement of the variable displacement pump; i2 represents the displacement of the variable displacement pump; i2 represents the displacement of the engine drive; i3 represents the displacement of the variable displacement pump; i4 represents the displacement of the variable displacement pump; i5 represents the displacement of the variable displacement pump; i6 represents the displacement of the variable displacement pump; i7 represents the displacement of the variable displacement motor; i8 represents the displacement of the variable displacement pump; i9 represents the displacement of the variable displacement pump; i1 represents the ratio of the engine drive to the variable displacement pump drive; i2 represents the displacement of the variable displacement pump; i2 represents the displacement ... Y1 Indicates the volumetric efficiency of the fracturing pump; η m1 Indicates the motor's volumetric efficiency; η p1 This indicates the efficiency of the variable pump.

[0110] The efficiency-affecting parameters are input, and the current efficiency of the variable motor, variable pump, and engine fuel consumption are predicted by a trained BP neural network.

[0111] S2: Use dynamic programming to determine the number of variables, motors and engines, and to determine the solution interval;

[0112] S21: Determine the objective function for the optimization process; In the operation optimization process of the fracturing truck, the control method with the lowest hourly fuel consumption is the optimal method for the current working condition, which corresponds to the minimum fuel consumption; The objective function for the minimum fuel consumption of the fracturing truck is as follows:

[0113]

[0114] In the formula: F represents the total fuel consumption of the engine when the fracturing truck is in operation; T represents the operating time of the fracturing truck.

[0115] S22: Determine the dynamic programming model; the number of engines in operation ranges from 1 to all engines, with each engine activation constituting a stage. Calculate the upper limit q of the flow rate that the two pumps on each engine can output under the current operating conditions; the state transition equation of the dynamic programming model is as follows:

[0116]

[0117] In the formula: f k+1 (Q k+1 ) represents the minimum fuel consumption generated when the flow is distributed to the pumps on the 1st to (k+1th)th engines; f k (Q k ) represents the minimum fuel consumption generated when the flow is distributed to the pumps on engines 1 through k; g k+1 (q k+1 ) represents q k+1 The minimum fuel consumption obtained by distributing the flow rate to the pump on the (k+1)th engine; Q k This represents the flow rate output of the pumps driven by the 1st to the kth engines, where 0 < Q. k ≤k·q;q k This represents the flow rate allocated to the pump on the k-th engine, 0 ≤ qk ≤q.

[0118] S23: Construct the algorithm optimization process to obtain the optimal number of motors to be put into operation and the optimal number of motors to be started, and provide recommended motor and pump displacements for controlling the variable motors and pumps to provide a reference range for the initial population settings of the genetic algorithm; such as Figure 3 The diagram shown is a schematic of the dynamic programming optimization algorithm according to an embodiment of the present invention.

[0119] The algorithm optimization process is constructed as follows:

[0120] S231: Based on the current hydraulic system pressure, the maximum flow rate that a single engine can output is calculated using the engine fuel consumption neural network and the variable pump efficiency neural network;

[0121] S232: Design step size decomposition to determine the flow rate that a single pump can output, solve for the optimal fuel consumption and pump displacement at each flow rate, and divide the calculation into 3 stages, with the calculation of one engine as one stage.

[0122] S233: Decompose the input flow rate Q with the designed step size, and use the flow range of the second engine as a variable to solve for the optimal fuel consumption and pump displacement within the flow range that the two engines can output.

[0123] S234: Similarly, after going through all stages, the optimal combination of pump displacements for the engine output flow rate Q is finally determined.

[0124] S3: Use an adaptive genetic algorithm to optimize the displacement of the variable motor and the variable pump to determine the optimal efficiency point;

[0125] First, calculate the crossover and mutation probabilities in the adaptive genetic algorithm; then, based on the criterion for determining the degree of population dispersion, arcsin(f ave / f max The distribution of the population is determined, and the crossover and mutation probabilities are iteratively updated. Finally, the optimal motor displacement of the optimal variable motor and the optimal pump displacement of the optimal variable pump are determined. The methods for obtaining the crossover and mutation probabilities are as follows:

[0126]

[0127]

[0128] In the formula: P c P represents the crossover probability; m f represents the probability of mutation; ave f represents the average fitness of the population. maxk1 represents the maximum fitness of the population; k2 represents the crossover probability weight; k3 represents the mutation probability weight.

[0129] According to the criteria for determining the degree of population dispersion, arcsin(f ave / f max To determine the distribution of the population and iteratively update the crossover and mutation probabilities, the following steps are taken:

[0130] arcsin(f ave / f max As a criterion for determining the degree of population dispersion, its value is determined by the average fitness value f of the population. ave And the maximum fitness of the population f max Decide.

[0131] When arcsin(f) ave / f max When π / 6 < π / 6, it indicates that the population is dispersed. The smaller the value of π / 6, the easier it is to determine that the population is dispersed. At this time, the population has a large degree of difference, is rich in population, and has good diversity. In this case, the crossover probability value is adaptively increased to allow chromosome genes to be fully exchanged and evolve into new, superior individuals. At the same time, the mutation probability value is adaptively decreased to reduce the probability of destroying superior individuals and accelerate the convergence speed.

[0132] When arcsin(f) ave / f max When π / 6 ≥ π / 6, it indicates that the population distribution is concentrated. The larger π / 6 is, the easier it is to determine that the population distribution is concentrated. At this point, the population diversity is small, and the population is homogeneous. Therefore, the crossover probability is adaptively reduced to prevent the destruction of existing high-quality genes; simultaneously, the mutation probability is adaptively increased to improve the population's global search ability and prevent it from getting trapped in local optima. Figure 4 Therefore, this is a schematic diagram of the genetic algorithm in an embodiment of the present invention.

[0133] S4: Obtain the control parameters of the fracturing truck to achieve global power matching control of the fracturing truck;

[0134] The control parameters of the fracturing truck include: the optimal engine speed determined by S1, the optimal number of variable displacement motors engaged and the optimal number of engine starts determined by S2, and the optimal displacement of the variable displacement motors and the optimal displacement of the variable displacement pump determined by S3.

[0135] The second aspect of this invention proposes a control system based on dynamic programming and genetic algorithm for power matching control of hydraulic fracturing trucks. This system can achieve the lowest fuel consumption under the current operating conditions by collecting the fracturing pump pressure output by the pressure sensor and the fracturing pump speed output by the speed sensor. The control system includes: a fracturing pump section, a hydraulic transmission section, an engine section, and a control system section.

[0136] The fracturing pump section includes a first pressure sensor 1, a fracturing pump 2, a first connecting shaft 3, a first speed sensor 4, and a fracturing pump gear 5. The first pressure sensor 1 is installed at the outlet of the fracturing pump 2, the first speed sensor 4 is installed near the first connecting shaft 3, and the fracturing pump gear 5 is coaxially rigidly connected to the fracturing pump 2 through the first connecting shaft 3.

[0137] The hydraulic transmission system includes a variable motor gear set 6, a variable motor 7-18, a flow sensor 19, a high-pressure pipeline 20, a second pressure sensor 21, an accumulator 22, a variable pump 23-28, a variable pump gear set 29-31, a low-pressure pipeline 32, a third pressure sensor 33, a back pressure valve 34, and an oil tank 50. The variable motor 7-18 and engine 44-46 of the hydraulic transmission system, as well as the variable pump 23-28 of the hydraulic system, all use gear transmission.

[0138] The engine section includes an engine gear reduction gearbox 35-37, a second speed sensor 38-40, a torque sensor 41-43, an engine 44-46, and a fuel consumption metering device 47-49.

[0139] The control system includes a signal input section and a control output section. The signal input section includes a first pressure sensor 1, a first speed sensor 4, a second pressure sensor 21, a third pressure sensor 33, second speed sensors 38-40, torque sensors 41-43, and fuel consumption metering devices 47-49. The control output section includes a vehicle controller, which contains a displacement control system for variable motors 7-18, a control system for the number of variable motors 7-18 in operation, a displacement control system for variable pumps 23-28, a control system for the number of engines 44-46 in operation, and an engine speed control system.

[0140] The fully hydraulic fracturing truck uses a hydraulic system for power confluence, employing three engines 44-46, each driving two parallel variable pumps 23-28, which together output high-pressure hydraulic oil. After confluence via high-pressure pipeline 20, the oil is output to twelve parallel variable motors 7-18, which drive fracturing pump 2 to output high-pressure, high-flow fracturing fluid.

[0141] The variable displacement motor 7-18 draws oil from the low-pressure line 32 through its inlet and transmits it to the high-pressure line 20 through its outlet to output high-pressure oil. The inlet of the variable displacement motor 7-18 is connected to the high-pressure line 20, and its outlet is connected to the low-pressure line 32. A flow sensor 19 is installed at the inlet of the variable displacement motor 7-18. A second pressure sensor 21 and an accumulator 22 are installed in the middle of the high-pressure line 20. A third pressure sensor 33 is installed on the low-pressure line 32. The first end of the back pressure valve 34 is connected to the low-pressure line 32, and the second end is connected to the oil tank 50. The variable displacement pump gear set 29-31 is connected to the engine gear reduction gearbox 35-37 via gear meshing. Each engine 44-46 has a second speed sensor 38-40 and a torque sensor 41-43 installed on its drive shaft, and is equipped with a fuel consumption metering device 47-49 to record the amount of fuel consumed per unit time during engine 44-46 operation. Figure 5 This is a diagram illustrating the hydraulic principle and hardware configuration system of an embodiment of the present invention.

[0142] Using a fracturing pump operating at 40 MPa and 1000 L / min as the optimal working condition, both genetic algorithm and intelligent collaborative optimization algorithm were used for optimization. The results are shown in Table 4. As can be seen from Table 4, the intelligent collaborative optimization algorithm has fewer optimization iterations than the genetic algorithm, achieves lower optimal fuel consumption, and expands the optimization variables from two continuous variables to two discrete variables and two continuous variables. This indicates that the efficiency and accuracy of the intelligent collaborative optimization algorithm are superior to those of the genetic algorithm.

[0143] Table 4 Optimization results of genetic algorithm and intelligent cooperative optimization algorithm

[0144]

[0145] Table 5 lists the efficiency comparison of the fracturing truck before and after the improvement. Experiments were conducted to compare the efficiency using the control methods before and after the improvement, verifying the energy-saving effect under four operating conditions: low pressure and low flow, high pressure and low flow, high pressure and high flow, and low pressure and high flow. The verification results show that the improved energy-saving control method has a better energy-saving effect than the previous method.

[0146] Table 5 Comparison of fracturing truck efficiency before and after improvement

[0147]

[0148] In summary, the results of the power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms presented in this case demonstrate its excellent effectiveness.

[0149] (1) The embodiments of the present invention accurately predict the actual fuel consumption of the fracturing truck under various parameters by using the speed and power relationship of the fully hydraulic fracturing truck and the fuel consumption prediction model of the fully hydraulic fracturing truck constructed by the BP neural network toolbox, thereby reducing the frequency of on-site testing, improving optimization efficiency, and shortening the optimization cycle.

[0150] (2) In this embodiment of the invention, dynamic programming algorithm and adaptive genetic algorithm are used to optimize the control of the five control parameters of the fracturing truck. The power matching control variables are solved by combining the optimization algorithm. The coordinated control of each variable can maximize the efficiency of the whole machine. The superiority of this method can be clearly seen by comparing the number of optimization iterations, the efficiency and accuracy of the algorithm through the data in the table.

[0151] (3) The control method of this invention improves the accuracy of prediction and the speed of calculation by using multiple sets of constraints, such as engine external characteristic curve constraints, variable pump and variable motor speed range constraints, and hydraulic system maximum pressure constraints. The effect verification on four working conditions of fracturing trucks, namely low pressure and low flow, high pressure and low flow, high pressure and high flow, and low pressure and high flow, all prove that the improved control method has a good effect compared with the original method.

[0152] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A power matching control method for a hydraulic fracturing truck based on dynamic programming and genetic algorithm, characterized in that, It includes the following steps: Step 1: Establish a fuel consumption prediction model for the fully hydraulic fracturing truck based on a BP neural network; A fuel consumption prediction model for a fully hydraulic fracturing truck was established. The overall structure of the fully hydraulic fracturing truck, with its multi-pump and multi-motor hydraulic transmission system, was determined. Speed ​​and power transmission models for the hydraulic system of the fully hydraulic fracturing truck were established. Influence factor models for variable motor and variable pump efficiency and engine fuel consumption rate were also established. The mathematical model for predicting the overall fuel consumption of the fully hydraulic fracturing truck, determined based on the engine speed and torque, is shown below: In the formula: G t p1 represents the fuel consumption of the entire vehicle per unit time; q1 represents the output pressure of the fracturing pump; b represents the output flow rate of the fracturing pump; e Indicates the engine's fuel consumption rate; η Y Indicates the fracturing pump efficiency; η m Indicates the variable motor efficiency; η p Indicates the efficiency of a variable pump; n e This indicates the engine speed; n4 indicates the engine speed of the fracturing truck; T4 indicates the engine torque of the fracturing truck. The method for obtaining the engine speed and torque of the fully hydraulic fracturing truck is as follows: Where: n m Indicates the number of motors in operation; V m V represents the displacement of a variable motor. Y i1 represents the displacement of the fracturing pump; i2 represents the ratio of the variable motor to the fracturing pump drive; Vp represents the displacement of the variable pump; i2 represents the ratio of the engine to the variable pump drive. η Y1 This indicates the volumetric efficiency of the fracturing pump; η m1 Indicates the motor's volumetric efficiency; η p1 Indicates the efficiency of the variable pump; The efficiency-affecting parameters are input, and the current efficiency of the variable motor, variable pump, and engine fuel consumption are predicted by a trained BP neural network. Step 2: Use dynamic programming algorithm to determine the number of motors and engines, and determine the solution interval; Step 21: Determine the objective function for the optimization process; In the optimization process of fracturing truck operation, the control mode with the lowest hourly fuel consumption is the optimal mode under the current working condition. The objective function for the minimum fuel consumption of fracturing truck operation is as follows: In the formula: F represents the total fuel consumption of the engine when the fracturing truck is in operation; T represents the operating time of the fracturing truck; Step 22: Determine the dynamic programming model; the number of engines engaged ranges from 1 to all engines operating, with each engine engaged representing a stage. Calculate the maximum flow rate that can be output under the current operating conditions when each engine carries two pumps; the state transition equation of the dynamic programming model is as follows: In the formula: f k+1 (Q k+1 ) represents the minimum fuel consumption generated when the flow is distributed to the pumps on the 1st to (k+1th)th engines; f k (Q k ) represents the minimum fuel consumption generated when the flow is distributed to the pumps on engines 1 through k; g k+1 (q k+1 ) represents q k+1 The minimum fuel consumption obtained by distributing the flow rate to the pump on the (k+1)th engine; Q k This represents the flow rate output of the pumps driven by the 1st to the kth engines, where 0 < Q. k ≤k·q;q k This represents the flow rate allocated to the pump on the k-th engine, 0 ≤ q k ≤q; Step 23: Construct the algorithm optimization process, obtain the optimal number of motors to be put into operation for variable motors and the optimal number of motors to be started for engines, and provide the recommended displacement of motors and pumps for controlling variable motors and variable pumps to provide a reference range for the initial population setting of the genetic algorithm; Step 3: Use an adaptive genetic algorithm to optimize the displacement of the variable motor and the variable pump to determine the optimal efficiency point; First, calculate the crossover and mutation probabilities in the adaptive genetic algorithm; then, based on the criterion for determining the degree of population dispersion, arcsin(f ave / f max The distribution of the population is determined, and the crossover and mutation probabilities are iteratively updated. Finally, the most efficient variable motor displacement and variable pump displacement under the current operating conditions are determined. Step 4: Obtain the control parameters of the fracturing truck to achieve global power matching control of the fracturing truck; The control parameters of the fracturing truck include: the optimal engine speed determined in step 1, the number of variable motors engaged and the number of engines started determined in step 2, and the displacement of variable motors and the displacement of variable pumps determined in step 3; the global power matching control of the fracturing truck is achieved through the control parameters of the fracturing truck.

2. The power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms according to claim 1, characterized in that, The speed transmission model in step 1 refers to the speed transmission model of the hydraulic system of the fully hydraulic fracturing truck, as shown below: In the formula: n1 represents the fracturing pump speed; n2 represents the variable displacement motor speed; q 2m n3 represents the flow rate input to the variable motor; n3 represents the speed of the variable pump; q 2q This represents the flow rate output by the variable pump; n p This indicates the number of variable pumps in operation.

3. The power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms according to claim 1, characterized in that, The power transfer model in step 1 refers to the power transfer model of the hydraulic system of the fully hydraulic fracturing truck, as shown below: In the formula: P Y T1 represents the fracturing pump power; T2 represents the fracturing pump input torque; T2 represents the output torque of the single variable motor; P m p1 represents the total output power of the variable displacement motor; p2 represents the hydraulic system pressure; P H q represents the total power of the hydraulic system. 2p T3 represents the output flow rate of the variable pump; T3 represents the input torque of the variable pump. η p P represents the efficiency of a variable pump. p P represents the total input power of the variable pump. e This indicates the total output power of the engine.

4. The power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms according to claim 1, characterized in that, The efficiency of the variable motor and variable pump, and the engine fuel consumption rate in step 1 are specifically as follows: The variable motor efficiency η m The effects of hydraulic system pressure, motor speed, and motor displacement are as shown in the following formula: η m =f1(p2,n2,V m ); In the formula: f1 represents the functional relationship between the variable motor efficiency and the hydraulic system pressure, motor speed, and motor displacement; p2 represents the hydraulic system pressure; The variable pump efficiency η p The effects of hydraulic system pressure, pump speed, and pump displacement are shown in the following formula: η p =f2(p2,n3,V p ); In the formula: f2 represents the functional relationship between the variable pump efficiency and the hydraulic system pressure, pump speed and pump displacement; The engine's fuel consumption rate b e The influence of engine speed and engine torque is as shown in the following formula: b e =f3(n4,T4); In the formula: f3 represents the functional relationship between the engine's fuel consumption rate and the engine speed and engine torque.

5. The power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms according to claim 1, characterized in that, The optimization process for the construction algorithm in step 23 is as follows: Step 31: Based on the current hydraulic system pressure, calculate the maximum flow rate that a single engine can output using the engine fuel consumption neural network and the variable pump efficiency neural network; Step 32: Design the step size to decompose the flow rate that a single pump can output, solve for the optimal fuel consumption and pump displacement at each flow rate, and divide the calculation into 3 stages, taking the process of calculating one engine as one stage. Step 33: Decompose the input flow rate Q with the designed step size, and use the flow range of the second engine as a variable to solve for the optimal fuel consumption and the displacement of each pump within the flow range that the two engines can output. Step 34: Go through all stages to determine the optimal combination of pump displacements for the engine output flow rate Q.

6. The power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms according to claim 1, characterized in that, The methods for obtaining the crossover probability and mutation probability in step 3 are as follows: In the formula: P c P represents the crossover probability; m f represents the probability of mutation; ave f represents the average fitness of the population. max k1 represents the maximum fitness of the population; k2 represents the crossover probability weight; k3 represents the mutation probability weight.

7. The power matching control method for hydraulic fracturing trucks based on dynamic programming and genetic algorithms according to claim 1, characterized in that, In step 3, the determination condition for the degree of population dispersion is arcsin(f ave / f max To determine the distribution of the population and iteratively update the crossover and mutation probabilities, the following steps are taken: arcsin(f ave / f max As a criterion for determining the degree of population dispersion, its value is determined by the average fitness value f of the population. ave And the maximum fitness of the population f max Decide; When arcsin(f) ave / f max When the crossover probability is less than π / 6, it indicates that the population is dispersed. The smaller the crossover probability is, the easier it is to determine that the population is dispersed. In this case, increasing the crossover probability allows chromosome genes to be fully exchanged, resulting in the evolution of superior individuals. At the same time, decreasing the mutation probability reduces the probability of destroying superior individuals and speeds up the convergence. When arcsin(f) ave / f max When π / 6 ≥ π / 6, it indicates that the population distribution is concentrated. The larger π / 6 is, the easier it is to determine that the population distribution is concentrated. At this time, the crossover probability should be reduced to prevent the destruction of existing high-quality genes. At the same time, the mutation probability should be increased to improve the global search capability of the population and prevent it from getting trapped in local optima.

8. A control system for implementing the power matching control method for a hydraulic fracturing truck based on dynamic programming and genetic algorithm according to any one of claims 1 to 7, characterized in that, The system can achieve the lowest fuel consumption under the current operating conditions by collecting the fracturing pump pressure output by the pressure sensor and the fracturing pump speed output by the speed sensor. The control system includes: fracturing pump part, hydraulic transmission part, engine part and control system part. The fracturing pump section includes a first pressure sensor, a fracturing pump, a first connecting shaft, a first speed sensor, and a fracturing pump gear. The first pressure sensor is installed at the outlet of the fracturing pump, the first speed sensor is installed near the first connecting shaft, and the fracturing pump gear and the fracturing pump are coaxially and rigidly connected through the first connecting shaft. The hydraulic transmission system includes a variable motor gear set that meshes with the fracturing pump gear, a variable motor, a flow sensor, a high-pressure pipeline, a second pressure sensor, an accumulator, a variable pump, a variable pump gear set, a low-pressure pipeline, a third pressure sensor, a back pressure valve, and an oil tank. The fracturing pump and the variable motor of the hydraulic transmission system, as well as the engine and the variable pump of the hydraulic system, all employ gear transmission. The engine section includes an engine gear reduction gearbox, a second speed sensor, a torque sensor, an engine, and a fuel consumption metering device; The control system includes a signal input section and a control output section. The signal input section includes a first pressure sensor, a first speed sensor, a second pressure sensor, a third pressure sensor, a second speed sensor, a torque sensor, and a fuel consumption metering device. The control output section includes a vehicle controller, which contains a variable displacement motor control system, a variable motor activation quantity control system, a variable pump displacement control system, an engine activation quantity control system, and an engine speed control system.

9. The control system of the hydraulic fracturing truck power matching control method based on dynamic programming and genetic algorithm according to claim 8, characterized in that, The fully hydraulic fracturing truck uses a hydraulic system for power confluence, employing 3 engines, each engine driving 2 parallel variable pumps, which together output high-pressure hydraulic oil. After being combined through high-pressure pipelines, the oil is output to 12 parallel variable motors to drive fracturing pumps to output high-pressure, high-flow fracturing fluid. The variable displacement motor's oil inlet draws oil from the low-pressure pipeline and transmits it to the high-pressure pipeline through the variable displacement motor's oil outlet to output high-pressure oil. The variable displacement motor's oil inlet is connected to the high-pressure pipeline, and the variable displacement motor's oil outlet is connected to the low-pressure pipeline. A flow sensor is installed at the variable displacement motor's oil inlet. A second pressure sensor and an accumulator are installed in the middle of the high-pressure pipeline. A third pressure sensor is installed on the low-pressure pipeline. The first end of the back pressure valve is connected to the low-pressure pipeline, and the second end is connected to the oil tank. The variable pump gear set is connected to the engine gear reducer via gear meshing; Each engine is equipped with a second speed sensor and a torque sensor on its drive shaft, and a fuel consumption metering device to record the amount of fuel consumed per unit time when the engine is working.

Images