Intelligent cleaning product reduction method for crude oil storage tank

Abstract

The application discloses a crude oil storage tank intelligent cleaning product reduction method, and belongs to the technical field of crude oil storage tank cleaning.The method comprises the following steps: step 1001, designing a reduction system; step 1002, observing the thickness of sludge; step 1003, planning a robot path; steps 1004 and 1005, observing the speed of the robot; step 1006, judging whether the thickness of the sludge meets the standard; step 1007, adjusting jet parameters; step 1008, providing cleaning pump pressure; step 1009, ending sludge discharge in the crude oil storage tank; steps 1010 and 1011, observing the oil content after treatment; step 1012, adjusting reduction system parameters; and step 1013, ending the sludge reduction system.In the crude oil storage tank intelligent cleaning product reduction method, a pure physical treatment process is combined with a reduction technology, and an industrial computer is used to control in-tank cleaning of a cleaning robot and out-of-tank reduction process of the cleaning robot, so that the recovery cost is reduced, and the recovery rate is improved, as compared with the prior art.

Description

Technical Field

[0001] A method for reducing the volume of cleaning products in crude oil storage tanks, belonging to the field of crude oil storage tank cleaning technology. Background Technology

[0002] Utilizing oil storage tanks is the primary method for strategic petroleum reserves. During long-term storage and transportation of refined oil products in these tanks, small amounts of impurities such as sand, mud, and heavy metal salts, due to their density difference, will settle and accumulate at the bottom of the tank along with water, forming a thick, black, gelatinous substance known as tank sludge. These impurities and moisture reduce the quality of the refined oil, affecting the accuracy of oil price calculations; accelerate tank corrosion, potentially causing perforation of the bottom plate and oil leaks; and generate and accumulate static electricity, leading to electrostatic discharge accidents. Therefore, cleaning of the tanks is essential. Tank cleaning technology has evolved from manual to mechanical cleaning. With increasing demands for safety, efficiency, cost, and environmental protection in tank cleaning operations, robotic tank cleaning technology, as a cleaning technology that eliminates the need for personnel to enter the tank, has gained increasing attention both domestically and internationally. However, the byproducts of robotic cleaning of refined oil tank sludge, as well as the sludge from water tanks, pose a significant environmental pollution hazard if directly discharged. Therefore, the resource recovery outside the tank is particularly important. However, in the existing technology, most methods for treating oil sludge outside the tank have problems such as high cost and low recovery rate, and the actual application effect is not significant.

[0003] Chinese invention patent application number 202010538240.2, filed on June 12, 2020, entitled "An Unmanned Rapid Cleaning System for Airport Oil Tanks," discloses a technical solution for designing mechanized cleaning, oil and gas recovery, and wastewater discharge compliance technologies for oil tanks. This technical solution provides an unmanned rapid cleaning system for airport oil tanks, including a cleaning module, a gas detection module, a nitrogen generation module, a vacuum suction module, an oil-water separation and reuse module, a wastewater discharge compliance treatment module, an oil and gas recovery treatment module, and a non-destructive testing module. Taking into account the wide distribution of airport oil tanks, the diverse types and capacities of tanks, and the limited operating time, it designs various cleaning devices such as small three-dimensional rotating nozzles, vertical spray guns, side-wall spray guns, and unmanned inspection robots. Furthermore, for large oil tanks, the cleaning robot is equipped with high-frequency guided wave flaw detection equipment to perform defect detection on the tank bottom plate, enabling the detection of the oil tank's operating status and ensuring its long-term continuous operation.

[0004] However, in this technical solution, the cleaning robot is completely controlled by humans inside the crude oil storage tank. In large oil tanks, where visualization is difficult and the bottom area is large, the accuracy of manual operation is low, and the long cleaning time due to the lack of path planning will be more prominent. Furthermore, when the physical properties of the sludge in the crude oil storage tank vary greatly and the sludge discharge volume is unstable, the "filtration + compact cyclone flotation" and "electrocoagulation + compact cyclone flotation" treatment systems, which cannot be adjusted in real time, will result in the inability to guarantee the quality of the discharged products and the waste of a large amount of electricity and chemicals. Summary of the Invention

[0005] The technical problem to be solved by this invention is to overcome the shortcomings of the prior art and provide a method for reducing the amount of crude oil storage tank cleaning products by combining a purely physical processing technology with a reduction technology, and controlling the cleaning robot's in-tank cleaning and external reduction processes through an industrial control computer. Compared with the prior art, this method reduces the recovery cost and improves the recovery rate.

[0006] The technical solution adopted by this invention to solve its technical problem is: a method for reducing the volume of cleaning products in crude oil storage tanks, characterized by including the following steps:

[0007] Step 1001: Design a volume reduction system based on the size of the crude oil storage tank and the ratio of water, oil and sludge in the crude oil storage tank. The volume reduction system includes an industrial control computer and a cleaning robot located inside the crude oil storage tank.

[0008] Step 1002: Measure the distribution of sludge in the crude oil storage tank using the sonar detection device inside the crude oil storage tank and the sonar detection device of the cleaning robot, and establish a sludge distribution database for the industrial control computer to call.

[0009] Step 1003: The industrial control computer calculates the shortest travel route for the cleaning robot and trains it. The cleaning robot travels in the crude oil storage tank and cleans according to the shortest travel route planned by the industrial control computer. The cleaning robot uses its onboard sonar to monitor the condition of the sludge after rinsing in real time and feeds it back to the industrial control computer and updates the sludge distribution database established in step 1002 in real time.

[0010] Step 1004: The industrial control computer uses the cleaning robot's tank positioning system to observe the moving speed of the cleaning robot inside the crude oil storage tank and establishes a database of robot moving speed.

[0011] Step 1005: Is the cleaning robot unable to move?

[0012] The industrial control computer determines whether the cleaning robot is unable to move. If it is unable to move, it executes step 1007. If the cleaning robot moves normally, it executes step 1006.

[0013] Step 1006: The industrial control computer calls the sludge distribution database to estimate the sludge thickness at the location of the cleaning robot and determines whether the sludge thickness meets the standard. If it meets the standard, proceed to step 1009; if it does not meet the standard, proceed to step 1008.

[0014] Step 1007: The industrial control computer adjusts the jet pressure and jet angle of the jet nozzle, and then returns to step 1004;

[0015] Step 1008: The industrial control computer adjusts the jet pressure of the cleaning pump according to the thickness of the sludge;

[0016] Step 1009: End the sludge removal process from the crude oil storage tank;

[0017] Step 1010: The industrial control computer uses soft measurement technology to measure the oil content of the products of the volume reduction system and establishes a database of oil content of the products of the volume reduction system.

[0018] Step 1011: The industrial control computer calls the oil content database of the product of the volume reduction system to determine whether the oil content meets the standard. If it meets the standard, proceed to step 1013; if it does not meet the standard, proceed to step 1012.

[0019] Step 1012: The industrial control computer adjusts the operating parameters of the reduction system;

[0020] Step 1013: End the sludge reduction system.

[0021] Preferably, in step 1003, the training includes the following steps:

[0022] Step 1003-1: Initialize all parameters, draw an electronic map based on the tank bottom sludge distribution database obtained by sonar detection, and perform grayscale processing;

[0023] Step 1003-2: Plan an optimal path based on the distribution of sludge and perform multi-point segmentation on the path;

[0024] Step 1003-3: Initialize the training model data and record the current Q value. Find the action corresponding to the maximum Q value through the ε-greedy strategy.

[0025] Step 1003-4, obtain the next location information and utilize it. Update the corresponding Q value;

[0026] In the formula: Q(s,a) is the estimated q value of the corresponding action quantity and state quantity, where s represents the state quantity of the Agent and a represents the action quantity of the Agent; Q(s',a) is the true q value of the corresponding action quantity and state quantity, where s represents the state quantity of the Agent and a represents the action quantity of the Agent; A is the action set of the Agent; α is the learning rate; γ is the discount factor; r is the reward value in the reward mechanism;

[0027] Step 1003-5: Determine if the current location is a sludge accumulation point. If it is the target point, proceed to step 1003-6; otherwise, return to step 1003-4.

[0028] Step 1003-6: Determine whether the training has reached the required number of iterations. If the set value has been reached, proceed to step 1003-7; otherwise, return to step 1003-3.

[0029] Step 1003-7, End.

[0030] Preferably, in step 1003-3, a reward mechanism r is set, specifically including the following steps:

[0031] Step 1003-3-1: When the cleaning robot is outside the tank or collides with an obstacle inside the crude oil storage tank, it will receive a negative reward.

[0032] Step 1003-3-2: When the cleaning robot is at the target point, it receives a positive reward.

[0033] Step 1003-3-3: Excluding the above two cases, the cleaning robot will obtain a reward according to the inverse proportional function of the distance between the cleaning robot's location and the target point.

[0034] The reward mechanism uses an inverse proportional function to calculate the distance between two points to obtain the reward value. This allows the cleaning robot to obtain a larger reward value as it gets closer to the target point, thus enabling the training of an optimal autonomous obstacle avoidance path. The reward function is shown in the following formula:

[0035]

[0036] In the formula: a0 represents a positive value, selected through simulation debugging; k is a proportional coefficient, selected through simulation debugging; (x0, y0) are the current coordinates of the cleaning robot; (x goal y goal () represents the coordinates of the expected target point.

[0037] Preferably, in the aforementioned volume reduction system, a cleaning pump is installed outside the crude oil storage tank. The cleaning pump is connected to a cleaning robot via a pipeline. A sludge pump is also installed outside the crude oil storage tank. The sludge pump extracts the sludge from the crude oil storage tank and sends it to the pre-processor. The outlet of the pre-processor is connected to a centrifuge. The outlet of the centrifuge is connected to an oil-water separator and a cracking furnace, respectively. The outlet of the oil-water separator is connected to a wastewater buffer tank and a crude oil storage tank, respectively.

[0038] Preferably, the preprocessor includes a preprocessor base, a plurality of supports are provided on the surface of the preprocessor base, a preprocessing chamber is provided between the supports, a dosing port and a cleaning water inlet are provided at the preprocessor cover at the upper port of the preprocessing chamber, an oily sludge inlet is provided on one side of the top of the preprocessing chamber, and an oily sludge outlet is provided at the bottom of the preprocessing chamber; a stirring mechanism is fixed on the preprocessor cover.

[0039] Preferably, the stirring mechanism includes a driving mechanism, the output shaft of which is coaxially fixed with a stirring shaft, the stirring shaft enters the pretreatment chamber, and a blade is provided at the lower part of the stirring shaft.

[0040] Preferably, the stirring power of the stirring mechanism and its calculation formula are as follows:

[0041]

[0042] Where: Pj represents stirring power; Np represents power number; The impeller's average density is represented by N; the impeller's rotational speed is represented by d; and the impeller's diameter is represented by d.

[0043] Preferably, the formula for calculating the diameter of the stirring shaft is:

[0044]

[0045] Where: d3 represents the shaft diameter; P j [τ] represents the stirring power; [τ] represents the allowable shear stress; N represents the impeller speed.

[0046] Preferably, the oil-water separator includes a separator base, a cylinder mounted on the surface of the separator base, an inlet component at the top of one end of the cylinder, and a sludge discharge funnel connected to the cylinder via a liquid inlet pipe after the inlet component enters the cylinder. A flow stabilizing component and a sludge-collecting component are sequentially arranged on the side of the sludge discharge funnel. An oil collection device is provided at the other end of the cylinder, with an oil collection port leading out from the oil collection device. A bidirectional liquid collection component is provided at the bottom of the oil collection device, and a drain port is provided at the bottom of the bidirectional liquid collection component.

[0047] Preferably, the sludge discharge funnel includes a funnel body, a sludge discharge pipe is provided at the bottom of the funnel body, the sludge discharge pipe passes through the cylinder and leads out, a funnel top cover is provided at the upper port of the funnel body, the liquid inlet pipe enters the funnel body vertically from the center of the funnel top cover, and a number of top cover openings are also provided on the surface of the funnel top cover.

[0048] Compared with the prior art, the beneficial effects of this invention are:

[0049] 1. In this intelligent cleaning product reduction method for crude oil storage tanks, a purely physical treatment process is combined with reduction technology. The cleaning robot's in-tank cleaning and external reduction processes are controlled by an industrial control computer. Compared with existing technologies, this reduces recycling costs and improves the recovery rate.

[0050] 2. In this intelligent cleaning product reduction method for crude oil storage tanks, a machine learning algorithm for path planning was designed. The distribution of oil sludge at the bottom of the tank was mapped electronically, and the cleaning path of the cleaning robot was planned using the Q-Learning algorithm and the ε-greedy strategy. This enabled the industrial control computer to autonomously control the cleaning robot to perform cleaning operations. At the same time, the distribution of oil sludge at the bottom of the tank before and after cleaning could be compared in the industrial control computer database, and secondary training of the cleaning path could be considered, which greatly ensured the cleaning effect.

[0051] 3. In this intelligent cleaning product reduction method for crude oil storage tanks, a reward mechanism is set when planning the path of the cleaning robot. This setting excludes the cleaning robot from boundary and collision conditions. The inverse proportional function of the distance between the real-time position of the cleaning robot and the target point position is calculated to obtain a reward for training. The design of this inverse proportional function ensures the accuracy of the training path.

[0052] 4. In this intelligent cleaning product reduction method for crude oil storage tanks, a skid-mounted sludge reduction treatment device was designed based on the distribution characteristics of oilfield tanks and the physical properties of sludge. The device consists of a pre-processor, a horizontal screw press, and a horizontal oil-water separator to complete three-phase separation and recover crude oil. The pyrolysis furnace then performs harmless treatment on the separated sludge. A matching intelligent monitoring algorithm was designed to monitor in real time whether the products of the reduction treatment system meet the standards and adjust the working conditions of the post-treatment system to truly achieve the goal of rational resource utilization while ensuring the effectiveness of the reduction treatment.

[0053] 5. This intelligent cleaning method for reducing the volume of waste products in crude oil storage tanks establishes a comprehensive intelligent control system capable of collecting complete industrial data. This data supports subsequent industrial analysis and system upgrades, enabling the entire system to achieve resource recovery and energy conservation. Its engineering implementation will yield significant social and environmental benefits. Attached Figure Description

[0054] Figure 1 Flowchart of a method for reducing the volume of cleaning products in intelligent cleaning of crude oil storage tanks.

[0055] Figure 2 This is a schematic diagram of an intelligent cleaning product reduction system for crude oil storage tanks.

[0056] Figure 3 A schematic diagram of the preprocessor structure for an intelligent cleaning product reduction system for crude oil storage tanks.

[0057] Figure 4 This is a schematic diagram of the oil-water separator in an intelligent cleaning product reduction system for crude oil storage tanks.

[0058] Figure 5 A schematic diagram of the sludge discharge funnel structure of the oil-water separator in the intelligent cleaning product reduction system for crude oil storage tanks.

[0059] Figure 6 for Figure 5 Top view.

[0060] The components are as follows: 1. Drive mechanism; 2. Motor base; 3. Cleaning water inlet; 4. Pre-processor top cover; 5. Paddle; 6. Pre-treatment chamber; 7. Oily sludge outlet; 8. Support; 9. Pre-processor base; 10. Oily sludge inlet; 11. Dosing port; 12. Inlet component; 13. Liquid inlet pipe; 14. Vent port; 15. Flow stabilizing component; 16. Cylinder; 17. Polymer structure; 18. Oil collection device; 19. Saddle support; 20. Oil collection port; 21. Two-way liquid collection component; 22. Drain port; 23. Separator base; 24. Sludge discharge funnel; 25. Sludge discharge port; 26. End cap; 27. Funnel top cover; 28. Funnel body; 29. ​​Sludge discharge pipe; 30. Top cover opening. Detailed Implementation

[0061] Figures 1-6 This is the preferred embodiment of the present invention, which is described below in conjunction with the accompanying drawings. Figures 1-6 The present invention will be further described below.

[0062] like Figure 1 As shown, a method for reducing the volume of cleaning products in crude oil storage tanks includes the following steps:

[0063] Step 1001: Begin by designing an intelligent cleaning product reduction system for crude oil storage tanks (hereinafter referred to as the reduction system) based on the size of the crude oil storage tank, the ratio of water, oil, and sludge in the crude oil storage tank, and the indicator requirements (processing capacity, sludge moisture content).

[0064] like Figure 2 As shown, the volume reduction system includes a crude oil storage tank, a cleaning robot installed inside the tank, and a cleaning pump installed outside the tank. The cleaning pump is connected to the cleaning robot via pipeline, and the cleaning robot cleans the inside of the crude oil storage tank. When hoisting the cleaning robot, firstly, pulleys are installed on the top of the crude oil storage tank to lift the cleaning robot to the top; then, a mounting plate is laid to guide the cleaning robot from the edge of the tank top to near the manhole at the top of the tower; finally, a tripod with pulleys is erected above the manhole, and the cleaning robot is then lowered into the crude oil storage tank.

[0065] A sludge pump is installed outside the crude oil storage tank. The sludge pump pumps out the oil sludge from the crude oil storage tank and sends it to the pre-processor. The outlet of the pre-processor is connected to a centrifuge. The outlet of the centrifuge is connected to an oil-water separator and a cracking furnace. The oil-water mixture obtained by the centrifuge is sent to the oil-water separator for oil-water separation. The separated wastewater and oil are sent to the wastewater buffer tank and the crude oil storage tank, respectively. The sludge separated by the centrifuge is sent to the cracking furnace for treatment.

[0066] like Figure 3 As shown, the preprocessor includes a preprocessor base 9, on which multiple supports 8 are skid-mounted. A preprocessing chamber 6 is located between the supports 8. A preprocessor cover 4 is located at the upper port of the preprocessing chamber 6. A dosing port 11 and a cleaning water inlet 3, communicating with the preprocessing chamber 6, are located on both sides of the preprocessor cover 4. An oily sludge inlet 10 for feeding oily sludge into the preprocessing chamber 6 is located on one side of the top of the preprocessing chamber 6, and an oily sludge outlet 7 is located at the bottom of the preprocessing chamber 6. A drive mechanism 1 is fixed to the preprocessor cover 4 via a motor mount 2. The drive mechanism 1 includes a drive motor and a reducer connected to the drive motor. A stirring shaft is coaxially fixed to the output shaft of the reducer. The stirring shaft passes downward through the preprocessor cover 4 and enters the preprocessing chamber 6. A paddle 5 for stirring the oily sludge in the preprocessing chamber 6 is located at the lower part of the stirring shaft.

[0067] A preprocessor is used to mix oily sludge, water, and cleaning agents, ensuring thorough agitation. The preprocessor employs a two-stage, tandem agitator to enhance mixing efficiency and features a leg-type support for skid-mounted operation, facilitating mobile production. Simultaneously, calculations are performed on the agitator power, cylinder structure design, and shaft diameter based on design conditions. The material volumetric flow rate and Reynolds number are calculated using the following formula:

[0068]

[0069] Where: Pj represents stirring power; Np represents power number; The average density of the material is represented by ; N represents the impeller speed; and d represents the impeller diameter. Combining the above calculations of stirring power, allowable stress, and impeller speed data, the shaft diameter of the stirring shaft is calculated using the following formula:

[0070]

[0071] Where: d3 represents the shaft diameter; P j [τ] represents the stirring power; [τ] represents the allowable shear stress; N represents the impeller speed.

[0072] After preprocessing, further liquid-solid separation is required. In this volume reduction system, an LW series horizontal screw discharge sedimentation centrifuge is selected to achieve liquid-solid separation. A multi-stage centrifugation process is designed. After the first-stage centrifuge finishes, the centrifuged product is transferred to the next stage centrifuge for multiple centrifugations. Particulate matter concentration sensors are installed at the outlet of each centrifuge to detect the content of solid particles in the sludge mixture. The data is transmitted to the industrial control computer, which determines whether the sludge mixture meets the separation standard and decides the number of repetitions until the concentration of the sludge mixture is below the standard.

[0073] like Figure 4 As shown, the oil-water separator includes a separator base 23, on which a cylinder 16 is mounted via a saddle-type support 19. End caps 26 are provided at both ends of the cylinder 16. An inlet component 12 is provided at the top of one end of the cylinder 16. The inlet component 12 is a perforated box-type inlet component, whose main function is to absorb the kinetic energy of the high-speed liquid flow entering the oil-water separator, reducing the impact and disturbance of the inlet jet on the flow field. A vent 14 is also provided at the top of the cylinder 16.

[0074] After the inlet component 12 enters the interior of the cylinder 16, it is connected to the sludge discharge funnel 24 through the liquid inlet pipe 13. The discharge pipe of the sludge discharge funnel 24 passes downward through the cylinder 16 and leads out as the sludge discharge port 25. A flow stabilizing component 15 is provided on one side of the sludge discharge funnel 24. The flow stabilizing component 15 adopts a bulk packing box type flow stabilizing component. Its function is to stabilize and improve the flow conditions in the main separation zone, making it more conducive to the separation process and avoiding the secondary eddies and backflow phenomena that still exist in the main separation zone.

[0075] A polymer structure 17 is provided on one side of the flow stabilizing component 15. The main function of the polymer structure 17 in the gravity oil-water separator is to enhance the separation performance of the equipment. An oil collection device 18 is provided at the other end of the cylinder 16 relative to the inlet component 12. The oil collection device 18 is located on the inner wall of the corresponding side end cap 26. An oil collection port 20 is also led out from the oil collection device 18 and extends from the end cap 26. A bidirectional liquid collection component 21 is provided at the bottom of the oil collection device 18. The bidirectional liquid collection component 21 allows as much equipment space as possible to be used for separation, thereby maximizing the volume utilization rate of the equipment. A drain port 22 is provided at the bottom of the bidirectional liquid collection component 21.

[0076] like Figures 5-6 As shown, the sludge discharge funnel 24 includes a funnel body 28, the lower part of which is conical. A sludge discharge pipe 29 is provided at the tip of the cone, and the sludge discharge pipe 29 passes through the aforementioned cylinder 16 and extends out. A funnel top cover 27 is provided at the upper end of the funnel body 28. The aforementioned liquid inlet pipe 13 enters the funnel body 28 vertically from the center of the funnel top cover 27. Several top cover openings 30 are also provided on the surface of the funnel top cover 27.

[0077] Since centrifuges cannot completely remove sediment, a small amount of sediment remains in the oil-water mixture entering the oil-water separator. If left untreated, this sediment can clog the bulk packing of the flow stabilizing component 15 over time, preventing separation. Therefore, it is essential to remove the sediment. To address this issue, a sludge discharge funnel 24 is used. Utilizing the principle of gravity settling, the oil-water mixture from the inlet, after being dissipated by the inlet component 12, enters the sludge discharge funnel 24 along the liquid inlet pipe 13. The oily wastewater, after settling, overflows from the top cover opening 30, preventing excessively high flow velocity or strong turbulence, thus stabilizing the flow and reducing impact and disturbance to the main separation zone.

[0078] Based on the conventional pyrolysis furnace, the gas nozzles in the preheating and heating sections of the furnace are replaced with resistance wires surrounding the furnace. The resistance wires provide uniform heating, rapid temperature rise, and do not produce nitrogen oxides. By changing the heat source of the pyrolysis furnace, the goal of low-carbon and environmentally friendly pyrolysis process is achieved.

[0079] After being dehydrated by a centrifuge, the oily sludge is conveyed to the electrothermal pyrolysis furnace by a feeding mechanism. The sludge is slowly propelled through the furnace tubes by a screw propeller. The average temperature inside the furnace is 700℃. The oily sludge exchanges heat thoroughly with the high-temperature circulating gas inside the furnace, reaching a temperature of 550℃. This process distills out the water and light components of the crude oil from the sludge and pyrolyzes the heavy components of the crude oil. The distilled components reach a temperature of approximately 400℃. The distilled mixture then enters a heat exchanger to exchange heat with the fuel gas, preheating it. Afterward, it enters a spray cooling device, where the oil and gas are cooled to 50℃, and the cooling water temperature increases from 30℃ to 50℃. The gasoline and diesel components cool down, and the oil-water mixture is then pumped to an oil-water separator for oil-water separation.

[0080] The oil content of the oily sludge discharged from the pyrolysis furnace is measured using soft sensing and uploaded to an industrial control computer to establish an oil content database. This database serves as data for subsequent soft sensing technology machine learning and is compared with national standards. If the oil content does not meet the standards, parameters such as the pretreatment unit speed, chemical dosage, centrifuge speed, and pyrolysis unit temperature are adjusted in real time to control the oil content. This allows for complete control of the oil content in the oily sludge and dynamic adjustment of chemical dosage and centrifuge speed based on the oil content, achieving energy conservation and emission reduction. A crude oil storage tank intelligent cleaning product reduction system, centered on an industrial control computer and integrating machine learning data acquisition and real-time intelligent control, has been established.

[0081] Non-condensable gases that cannot be cooled by the spray cooling device enter the liquid seal and gas buffer tank. This part of the gas is mixed with external natural gas. After being pressure-stabilized by the gas buffer tank, it is passed into the heat exchange device to exchange heat with the pyrolysis gas from the heating furnace. After the temperature rises, it is used as fuel for the burner of the heating furnace, realizing the reuse of pyrolysis gas. The oily sludge after pyrolysis is loose and meets the requirements for discharge.

[0082] Step 1002: Observe the thickness of the sludge;

[0083] The distribution of sludge inside crude oil storage tanks is measured using sonar detection devices inside crude oil storage tanks and sonar detection devices on cleaning robots, and a sludge distribution database is established for use by industrial control computers.

[0084] Sonar is installed in front of and directly below the cleaning robot to monitor the sludge condition in real time and transmit the information to the industrial control computer to establish a sludge distribution database. The industrial control computer performs real-time analysis and issues instructions. If the sludge in front is thick or thin, it intelligently feeds back to the frequency converter to adjust the water pressure at the outlet of the cleaning pump.

[0085] Step 1003: Plan and train the robot's cleaning path;

[0086] The cleaning robot uses its onboard sonar to monitor the condition of the washed sludge in real time, feeding this data back to the industrial control computer and updating the sludge distribution database established in step 1002. The industrial control computer pre-sets sludge cleaning targets. After the cleaning robot reaches the sludge area at the bottom of the oil tank, the sonar activates, collecting sludge distribution information. The industrial control computer then calculates the shortest path for the cleaning robot using the Q-Learning algorithm and the ε-greedy strategy, and trains the robot accordingly. The training process is as follows:

[0087] Step 1003-1: Initialize all parameters, draw an electronic map based on the tank bottom sludge distribution database obtained by sonar detection, and perform grayscale processing;

[0088] Step 1003-2: Plan an optimal path based on the distribution of sludge and perform multi-point segmentation on the path;

[0089] Step 1003-3: Initialize the training model data and record the current Q value. Find the action corresponding to the maximum Q value through the ε-greedy strategy.

[0090] This step includes a reward mechanism r, specifically the following steps:

[0091] Step 1003-3-1: When the cleaning robot is outside the tank or collides with an obstacle inside the crude oil storage tank, it will receive a large negative reward.

[0092] Step 1003-3-2: When the cleaning robot is at the target point, it receives a large positive reward.

[0093] Step 1003-3-3: Excluding the above two cases, the cleaning robot will obtain a reward according to the inverse proportional function of the distance between the cleaning robot's location and the target point.

[0094] The reward mechanism uses an inverse proportional function to calculate the distance between two points to obtain the reward value. This allows the cleaning robot to obtain a larger reward value as it gets closer to the target point, thus enabling the training of an optimal autonomous obstacle avoidance path. The reward function is shown in the following formula:

[0095]

[0096] In the formula: a0 represents a positive value, selected through simulation debugging; k is a proportional coefficient, selected through simulation debugging; (x0, y0) are the current coordinates of the cleaning robot; (x goal y goal () represents the coordinates of the expected target point.

[0097] After training, the cleaning robot begins its cleaning operation. After the first cleaning cycle, the sonar will re-collect the sludge distribution information after the sludge settles and compare it with the previous database. This will create a new electronic map of the areas with poor cleaning results. The industrial control computer will then train the robot to plan its path on the new electronic map and issue cleaning commands. The robot will perform a second cleaning on the areas with poor cleaning results. After each sludge settling cycle, the industrial control computer will continuously update the sludge distribution database in the crude oil storage tank, assess the sludge content, and determine the number of cleaning cycles to repeat until the sludge content is below the target level, at which point the cleaning process ends.

[0098] Step 1003-4, obtain the next location information and utilize it. Update the corresponding Q value;

[0099] Step 1003-5: Determine if the current location is a sludge accumulation point. If it is the target point, proceed to step (5); otherwise, return to step 1003-2.

[0100] Step 1003-6: Determine whether the training has reached the required number of iterations. If the set value has been reached, proceed to step 1003-7; otherwise, return to step 1003-3.

[0101] Step 1003-7, End.

[0102] Step 1004: Observe the robot's speed;

[0103] The moving speed of the cleaning robot inside the crude oil storage tank was observed using the tank positioning system of the cleaning robot, and a database of robot moving speed was established.

[0104] Step 1005: Is the cleaning robot unable to move?

[0105] The industrial control computer determines whether the cleaning robot is unable to move. If it is unable to move, it executes step 1007. If the cleaning robot moves normally, it executes step 1006.

[0106] The cleaning robot adopts a dual-drive system of "pneumatic + hydraulic". The cleaning robot collects position information through sonar, and the industrial control computer analyzes the speed and position and issues commands to start the cleaning robot. At the bottom of the oil tank, the cleaning robot mainly moves by pneumatically driven wheels. When encountering thick sludge or complex working conditions, the robot moves relatively slowly. After the sonar collects position information in real time, the industrial control computer calculates the robot's real-time moving speed and, based on the sludge distribution database established in step 1002, comprehensively judges the difficulty of the robot's movement. The industrial control computer feeds back the commands to the frequency converter. The frequency converter can adjust the water pressure by changing the valve opening and the direction of the cleaning pump outlet, providing auxiliary power to the cleaning robot, thereby achieving the function of hydraulic assisted drive.

[0107] Step 1006: Determine if the sludge thickness meets the standard;

[0108] The industrial control computer calls the sludge distribution database to estimate the sludge thickness at the location of the cleaning robot and determines whether the sludge thickness meets the standard. If it does, proceed to step 1009; otherwise, proceed to step 1008.

[0109] Step 1007: Adjust the jet parameters;

[0110] The industrial control computer uses the sludge distribution database and the cleaning robot speed database to comprehensively estimate the difficulty of the cleaning robot's movement, adjusts the jet nozzle pressure and jet angle, and then returns to step 1004. Because the sludge thickness is directly proportional to the difficulty of movement, increasing the jet pressure can achieve the dual effect of assisting the drive and enhancing the cleaning ability.

[0111] Step 1008: Provide cleaning pump pressure;

[0112] The industrial control computer adjusts the jet pressure of the cleaning pump according to the thickness of the sludge to achieve thorough cleaning.

[0113] Step 1009: End the sludge removal process from the crude oil storage tank.

[0114] Step 1010: Observe the oil content after post-processing;

[0115] The oil content of the products of the weight reduction system was measured using soft measurement technology, and a database of the oil content of the products of the weight reduction system was established.

[0116] Step 1011: Determine whether the emissions meet the standards;

[0117] The industrial control computer calls the oil content database of the product from the reduction system, compares the oil content of the product with the national standard, and determines whether the oil content meets the standard. If it meets the standard, proceed to step 1013; if it does not meet the standard, proceed to step 1012.

[0118] Step 1012: Adjust the parameters of the volume reduction system;

[0119] The industrial control computer adjusts parameters such as the speed of the pre-processor agitator, the dosage of chemicals, the speed of the centrifuge, and the temperature of the pyrolysis unit to achieve energy saving and emission reduction while ensuring the treatment effect.

[0120] Step 1013: End the sludge reduction system.

[0121] Furthermore, an internal database for the oil tank will be established to provide a data foundation for future equipment fault diagnosis, process monitoring, and improvement of the machine learning accuracy of this intelligent system.

[0122] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for intelligent cleaning of crude oil storage tank with reduced product loss, characterized by: The method comprises the following steps: Step 1001, designing a reduction system according to the volume of the crude oil storage tank, the proportion of water, oil and mud in the crude oil storage tank, the reduction system comprising an industrial computer and a cleaning robot located in the crude oil storage tank; Step 1002, measuring the distribution of oil and mud in the crude oil storage tank by using a sonar detection device in the crude oil storage tank and a sonar detection device of the cleaning robot, and establishing an oil and mud distribution database for calling by the industrial computer; Step 1003, the industrial computer calculates the shortest travel route of the cleaning robot and trains it, the cleaning robot travels in the crude oil storage tank according to the shortest travel route planned by the industrial computer and cleans, the cleaning robot carries out real-time monitoring and feedback of the mud condition after flushing to the industrial computer through the sonar, and the oil and mud distribution database established in step 1002 is updated in real time; Step 1004, the industrial computer observes the moving speed of the cleaning robot in the crude oil storage tank by using a positioning system in the cleaning robot tank, and establishes a robot moving speed database; Step 1005, whether the cleaning robot cannot travel; If the cleaning robot cannot travel, step 1007 is executed, and if the cleaning robot travels normally, step 1006 is executed; Step 1006, the industrial computer estimates the oil and mud thickness at the position of the cleaning robot by calling the oil and mud distribution database, and judges whether the oil and mud thickness meets the standard, if it meets the standard, step 1009 is executed, and if it does not meet the standard, step 1008 is executed; Step 1007, the industrial computer adjusts the jet pressure and jet angle of the jet nozzle, and returns to step 1004; Step 1008, the industrial computer adjusts the jet pressure of the cleaning pump according to the oil and mud thickness; Step 1009, ending the mud discharge in the crude oil storage tank; Step 1010, the industrial computer measures the oil content of the reduction system product by using soft measurement technology, and establishes a reduction system product oil content database; Step 1011, the industrial computer calls the reduction system product oil content database, judges whether the oil content meets the standard, if it meets the standard, step 1013 is executed, and if it does not meet the standard, step 1012 is executed; Step 1012, the industrial computer adjusts the operating parameters of the reduction system; Step 1013, ending the oil and mud reduction system.

2. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 1, characterized by: In the step 1003, the training comprises the following steps: Step 1003-1, initializing parameters, drawing an electronic map according to the oil and mud distribution database obtained by sonar detection and carrying out gray scale processing; Step 1003-2, planning an optimal path according to the oil and mud distribution and carrying out multi-point segmentation processing on the path; Step 1003-3, initializing training model data, recording the current Q value, and finding the maximum Q value corresponding action through the ε-greedy strategy; Step 1003-4, acquire next position information and utilize update corresponding Q value; In the formula: Q(s, a) is an estimated q value corresponding to an action amount and a state amount, wherein s represents a state amount of the agent, and a represents an action amount of the agent; Q(s', a) is a true q value corresponding to an action amount and a state amount, wherein s represents a state amount of the agent, and a represents an action amount of the agent; A is an action set of the agent; a is a learning rate; g is a discount factor; and r is a reward value in a reward mechanism; Step 1003-5, judging whether the position is an oil sludge gathering point, if yes, entering step 1003-6, otherwise returning to step 1003-4; Step 1003-6, judging whether the training reaches the iteration number, if yes, entering step 1003-7, otherwise returning to step 1003-3; Step 1003-7, ending.

3. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 2, characterized by: In the step 1003-3, a reward mechanism r is set, which specifically includes the following steps: Step 1003-3-1, when the cleaning robot is in an area outside the tank or collides with an obstacle in the crude oil storage tank, a negative reward is obtained; Step 1003-3-2, when the cleaning robot is at the target point, a positive reward is obtained; Step 1003-3-3, excluding the above two cases, the cleaning robot will obtain a reward according to the inverse function of the distance between the position of the cleaning robot and the target point; The method of applying the inverse function of the distance between two points to obtain the reward value in the reward mechanism can make the cleaning robot obtain a larger reward value when it is closer to the target point, and can train an optimal autonomous obstacle avoidance path. The reward function is shown in the following formula: In the formula: a0 represents a positive value, selected through simulation debugging; k is a proportional coefficient, selected through simulation debugging; (x0, y0) are the current coordinates of the cleaning robot; (x goal y goal () represents the coordinates of the expected target point.

4. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 1, characterized by: In the reduction system, a cleaning pump is arranged outside the crude oil storage tank, the cleaning pump is connected with the cleaning robot through a pipeline, a sewage pump is also arranged outside the crude oil storage tank, the sewage pump sends the oil sludge in the crude oil storage tank to the pretreater after being pumped out, the outlet of the pretreater is connected with the centrifuge, the outlet of the centrifuge is connected with the oil-water separator and the cracking furnace respectively, and the outlet of the oil-water separator is connected with the sewage buffer tank and the crude oil storage tank.

5. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 4, characterized in that: The pretreater includes a pretreater base, a plurality of supports are arranged on the surface of the pretreater base, a pretreatment cavity is arranged between the supports, a dosing port and a cleaning water inlet are arranged at the pretreater upper cover of the pretreatment cavity at the upper end, an oily sludge inlet is arranged at one side of the top of the pretreatment cavity, and an oily sludge outlet is arranged at the bottom of the pretreatment cavity; a stirring mechanism is fixed on the pretreater upper cover.

6. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 5, characterized in that: The stirring mechanism includes a driving mechanism, a stirring shaft is coaxially fixed on the output shaft of the driving mechanism, the stirring shaft enters the pretreatment cavity, and a paddle is arranged at the lower part of the stirring shaft.

7. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 5, characterized by: The stirring power and the calculation formula of the stirring mechanism are as follows: wherein: Pj represents the stirring power; Np represents the power number; wherein: Pj represents the stirring power; Np represents the power number; wherein: Pj represents the stirring power; Np represents the power number; 8. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 6, characterized by: The calculation formula of the stirring shaft diameter is as follows: wherein: d3 represents the shaft diameter; P j represents the stirring power; [τ] represents the allowable shear stress; N represents the impeller rotational speed.

9. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 4, characterized by: The oil-water separator comprises a separator base, a cylinder body is mounted on the surface of the separator base, an inlet member is arranged on the top of one end of the cylinder body, and after entering the cylinder body, the inlet member is connected to a sludge discharge funnel through a liquid inlet pipe; a flow stabilizing member and an aggregation member are sequentially arranged on the side of the sludge discharge funnel; an oil collecting device is arranged on the other end of the cylinder body, an oil collecting port is led out at the oil collecting device, a two-way liquid collecting member is arranged at the bottom of the oil collecting device, and a water discharge port is arranged at the bottom of the two-way liquid collecting member.

10. The intelligent cleaning product reduction method for crude oil storage tanks according to claim 9, characterized in that: The sludge discharge funnel comprises a funnel body, a sludge discharge pipe is arranged at the bottom of the funnel body and led out after penetrating through the cylinder body, a funnel top cover is arranged at the upper end of the funnel body, the liquid inlet pipe vertically enters the funnel body from the center of the funnel top cover, and a plurality of top cover openings are formed in the surface of the funnel top cover.

Images