Regulation type blast furnace slag flushing water waste heat recovery system based on image recognition deep learning

Abstract

The invention discloses a deep learning regulation type blast furnace slag flushing water waste heat recovery system based on image recognition, and relates to the technical field of image recognition and waste heat utilization, in particular to a slag flushing water filtering system which is used for monitoring the concentration of a slag flushing water pipeline, if the concentration of slag flushing water does not meet the preset concentration requirement, sedimentation is conducted again till the preset concentration requirement is met, and then the slag flushing water is recycled. Heat exchange is conducted through the heat pipe, image recognition is conducted on the thickness of dirt on the inner wall of the heat pipe, and when the thickness of the dirt on the inner wall of the heat pipe does not meet the preset thickness requirement, descaling treatment is conducted till the thickness of the dirt on the inner wall of the heat pipe meets the preset thickness requirement; the ORC organic Rankine system is connected with the slag flushing water filtering system and is used for recovering the waste heat of the slag flushing water; and the cooling system is connected with the ORC organic Rankine system and is used for cooling the organic working medium. According to the invention, the deep learning regulation type blast furnace slag flushing water waste heat recovery system based on image recognition can be realized.

Description

Technical Field

[0001] The present invention relates to the technical fields of image recognition and waste heat utilization, and particularly relates to a regulation-type blast furnace slag water waste heat recovery system based on deep learning of image recognition. Background Art

[0002] In the process of steel production, as a core device, the operation efficiency and stability of the blast furnace are directly related to the efficiency of the entire production line. However, during the operation of the blast furnace, the slag water, as an inevitable by-product, not only contains rich waste heat resources but also brings a series of challenges. A large amount of waste heat contained in the slag water, if not effectively recovered and utilized, is not only a great waste of energy but also increases carbon emissions and has an adverse impact on the environment.

[0003] In traditional processes, for the waste heat recovery of blast furnace slag water, it mostly relies on traditional heat exchange equipment. Although these equipment have achieved the recovery and utilization of heat energy to a certain extent, due to the high particle content in the slag water, it is easy to deposit on the surface of traditional heat exchangers during the heat exchange process, forming scale. This not only reduces the heat exchange efficiency of the equipment but also affects the safe and stable operation of the system, increasing the complexity and cost of equipment maintenance. Therefore, it is of great significance to adopt a suitable heat exchange equipment for the heat of blast furnace slag water. The heat pipe heat exchanger has the advantages that there will be no cross leakage between the waste medium and the supply medium, with high heat exchange efficiency, compact structure, no moving parts, light weight, relatively economical, and the complete separation of the hot fluid and the cold fluid. However, the problem of heat exchanger fouling is a major problem in its application.

[0004] With the increasing attention to energy and environmental issues at home and abroad and the proposal of the national "dual carbon" goal, the Organic Rankine Cycle (ORC) technology has attracted much attention. This technology can effectively utilize resources such as industrial waste heat, geothermal heat, and waste heat, and has certain advantages compared with other medium and low temperature heat energy utilization technologies. However, in the face of slag water with high particle content, the ORC system heat pipe heat exchange system cannot avoid the dilemmas of low heat exchange efficiency and equipment fouling.

[0005] In this context, the rise of image recognition technology provides a new idea for the waste heat recovery system. In recent years, the application of image recognition technology in industrial process monitoring has become increasingly widespread. Its ability to monitor and analyze the operation status of the system in real time provides a reliable guarantee for the efficient and stable operation of industrial equipment. Applying image recognition technology to the waste heat recovery system, by monitoring the status of the heat exchanger surface in real time, the fouling phenomenon can be discovered and prevented in time.

[0006] Furthermore, by combining image recognition technology with the organic Rankine cycle technology, a more intelligent and efficient solution for deep waste heat recovery is developed. This solution will use image recognition technology to monitor the heat exchanger status in real time, analyze data through intelligent algorithms, predict and adjust system operation parameters to reduce fouling and improve heat transfer efficiency. At the same time, the high-efficiency waste heat recovery technology of the ORC system will ensure that the waste heat in the slag flushing water is fully utilized, thus achieving the maximum utilization of energy and environmental protection.

[0007] The system and method for combined cooling and heating supply of blast furnace slag flushing water and low-grade steam with the publication number CN116875747A connect the slag flushing water pretreatment system through a filter tank via a water intake pump, and the outlet end is connected to the slag flushing water inlet of the waste heat utilization device. The slag flushing water outlet of the waste heat utilization device is connected to the inside of the filter tank near the outlet end through a pipeline; the waste heat utilization device is also provided with a low-grade steam inlet; the pretreated blast furnace slag flushing water is sent into the waste heat utilization device, and the slag flushing water after utilization is returned to the cooling tower and the cold water tank for circulating slag flushing. However, the connection between the filter tank and the slag flushing water pretreatment system through the water intake pump cannot completely remove the impurities in the slag flushing water, so the problem of blockage of the heat exchange equipment cannot be effectively solved.

[0008] The blast furnace slag flushing water waste heat recovery system and its working method with the publication number CN115369195A solve the problem of difficult recovery and utilization of the waste heat of blast furnace slag flushing water in summer; it promotes the efficient circulation of the slag flushing water in the mode of large flow and small temperature difference. During the circulation process of the slag flushing water, the vertical parallel negative pressure flash evaporation technology is adopted to produce sufficient flash negative pressure steam, and the waste steam is used to drive the waste steam bromine refrigerator to produce sufficient chilled water, realizing the efficient recovery and utilization of the waste heat of blast furnace slag flushing water in summer. However, by using the waste steam to drive the waste steam bromine refrigerator to produce sufficient chilled water, the utilization effect of the slag flushing water waste heat is too single, reducing the utilization effect of the slag flushing water waste heat.

[0009] The blast furnace slag flushing water waste heat utilization system with the publication number CN218372360U includes a slag flushing water circulation loop, an organic working fluid circulation loop and a water vapor circulation loop. Among them, there is a slag flushing water filter and an evaporator on the slag flushing water circulation loop, and an evaporator, a compressor, a condenser, an economizer and an expansion valve on the organic working fluid circulation loop; there is a flash tank, a steam accumulator, a power generation steam turbine, a generator, a condenser, a deaerator and a feed water pump on the steam circulation loop. Among them, a pressurized water pump is provided between the flash tank and the condenser, and the condenser is located between the outlet of the heating water pump and the inlet of the flash tank. This system can utilize low-grade slag flushing hot water and generate steam above 100°C, and generate electric energy in combination with the saturated steam power generation system, which can make full use of the waste heat. By coupling the slag flushing water waste heat system with the ORC system, it can make full use of the slag flushing water waste heat. However, it lacks a slag flushing water descaling system and is prone to the condition of evaporator blockage.

[0010] Therefore, how to implement a regulated blast furnace slag flushing water waste heat recovery system based on image recognition deep learning is a technical problem that urgently needs to be solved at present. Summary of the Invention

[0011] In view of the existing problems such as waste heat utilization of slag flushing water and scaling of the slag flushing water heat exchange system, an embodiment of the present invention provides a regulated blast furnace slag flushing water waste heat recovery system based on image recognition deep learning. Based on AI image recognition and particle concentration monitoring, the possibility of scaling in the slag flushing water system is reduced, and the scaling problem in the heat exchange process of the slag flushing water waste heat system is effectively solved. At the same time, the state of the slag flushing water is monitored and regulated in real time. Through coupling with the ORC organic Rankine cycle system, deep recovery of the waste heat of the slag flushing water waste heat system is realized, and the energy utilization efficiency is improved.

[0012] In a first aspect, the present invention provides a regulated blast furnace slag flushing water waste heat recovery system based on image recognition deep learning, including: a slag flushing water filtration system, an ORC organic Rankine cycle system, and a cooling system;

[0013] The slag flushing water filtration system is used to monitor the concentration of the slag flushing water pipeline. If the concentration of the slag flushing water does not meet the preset concentration requirement, it will be sedimented again until it meets the preset concentration requirement, and then heat exchange will be carried out by a heat pipe. At the same time, image recognition is performed on the thickness of the dirt on the inner wall of the heat pipe. When the thickness of the dirt on the inner wall of the heat pipe does not meet the preset thickness requirement, descaling treatment is carried out until the thickness of the dirt on the inner wall of the heat pipe meets the preset thickness requirement;

[0014] The ORC organic Rankine cycle system is connected to the slag flushing water filtration system and is used to complete the waste heat recovery of the slag flushing water;

[0015] The cooling system is connected to the ORC organic Rankine cycle system and is used to complete the cooling of the organic working fluid.

[0016] In some examples, the slag flushing water filtration system includes a water pump, a primary sedimentation tank, a filter, a three-way control valve, a heat pipe, a vibration actuator, a monitoring terminal, and a descaling tank;

[0017] The slag flushing water is sent into the primary sedimentation tank by a water pump for sedimentation. After being filtered by a filter, the concentration of the slag flushing water pipeline is monitored. If the concentration of the slag flushing water meets the preset concentration requirement, it is sent into the heat pipe for heat exchange. If the concentration of the slag flushing water does not meet the preset concentration requirement, the three-way control valve is adjusted through the monitoring terminal, and the slag flushing water is sent into the primary sedimentation tank for re-sedimentation. After being filtered by the filter, the concentration of the slag flushing water pipeline is monitored until it meets the preset concentration requirement, and then it is sent into the heat pipe for heat exchange. Then, image recognition is performed on the thickness of the dirt on the inner wall of the heat pipe. When the thickness of the dirt on the inner wall of the heat pipe does not meet the preset thickness requirement, the vibration actuator is controlled through the monitoring terminal to perform descaling treatment on the heat pipe wall surface. The waste water after descaling is sent into the descaling tank through the monitoring terminal to control the three-way control valve.

[0018] In some instances, the ORC (Organic Rankine Cycle) system includes a steam turbine, a condenser, a first circulation pump, a cooling pond, and a booster pump connected to the heat pipe of the slag water filtration system;

[0019] The organic working fluid exchanges heat with the slag water in the heat pipe. After the heat exchange is completed, a saturated organic working fluid vapor is formed, which is sent to the steam turbine for power generation. After being condensed by the condenser, it is sent to the cooling pond through the first circulation pump, and after being pressurized by the booster pump, it is sent back into the heat pipe for heat exchange again, and the cycle is repeated to complete the waste heat recovery of the slag water.

[0020] In some instances, the cooling system includes a cooling tower, a second circulation pump, a valve, and a liquid storage tank connected to the condenser of the ORC system;

[0021] After the cooling water exchanges heat with the organic working fluid in the condenser, it is sent to the cooling tower for cooling, then sent to the liquid storage tank through the second circulation pump and the valve for storage, and then sent back into the condenser for heat exchange, and the cycle is repeated to complete the cooling of the organic working fluid.

[0022] In some instances, the monitoring of the concentration of the slag water pipeline includes:

[0023] Monitoring the concentration of the slag water pipeline, and comparing the concentration of the slag water pipeline with the data in the database to determine whether it meets the preset concentration requirements.

[0024] In some instances, the comparison of the concentration of the slag water pipeline with the data in the database includes:

[0025] Using a neural network model to compare the concentration of the slag water pipeline with the data in the database.

[0026] In some instances, the image recognition of the thickness of the dirt on the inner wall of the heat pipe includes:

[0027] Using a neural network model of the dirt on the inner wall of the heat pipe to perform image recognition on the thickness of the dirt on the inner wall of the heat pipe to determine whether the thickness of the dirt on the inner wall of the heat pipe meets the preset thickness requirements.

[0028] In some instances, the neural network model of the dirt on the inner wall of the heat pipe is pre-trained based on the real-time obtained image information of the inner wall of the heat pipe by using human-computer interactive monitoring and analysis. During the training process, the stochastic gradient descent algorithm is used to optimize the parameters, and the optimal network model is obtained by evaluating the neural network model of the dirt on the inner wall of the heat pipe through human-computer interactive monitoring and analysis.

[0029] In some instances, a neural network model for fouling on the inner wall of a heat pipe is trained using historical data. In the human-machine interaction mode, in the early stage, the images of the fouling on the inner wall of the heat pipe collected in real time are imported into the neural network model for fouling on the inner wall of the heat pipe with the optimal parameters trained, through comparing the difference between the real-time data of the fouling on the heat pipe wall surface and the prediction result of the neural network model for fouling on the inner wall of the heat pipe, to determine whether there is an abnormality in the heat transfer state of the heat pipe; if the difference between the real-time data of the heat pipe wall surface or the heat transfer effect and the prediction result of the neural network model for fouling on the inner wall of the heat pipe exceeds a preset threshold, the diagnosis result is marked as abnormal, then for the diagnosed abnormality, the neural network model for fouling on the inner wall of the heat pipe further analyzes the cause, and adjusts the vibration execution vibration frequency and number of times automatically or provides an alarm and operation suggestions to the operator.

[0030] In some instances, the ORC organic Rankine system adopts a screw expander power generation device, and the organic working fluid can be selected from one or several optimized combinations of R245fa, R142b, R141b, R123, n-butane, and isopropane according to seasons and the temperature of the slag flushing water.

[0031] Generally speaking, compared with the prior art through the above technical solutions conceived by the present invention, the following beneficial effects can be achieved:

[0032] (1) Through the system coupling of the blast furnace slag flushing and filtering system and the ORC Rankine cycle, the present invention can deeply utilize the waste heat of the blast furnace slag flushing water, solves the problems of easy fouling and blockage of the blast furnace slag flushing water heat exchanger, the decline of the heat transfer efficiency, the influence on the heat source stability and the power generation efficiency, effectively recovers the waste heat in the blast furnace slag flushing water, and realizes the stable operation and long life of the waste heat recovery of the blast furnace slag flushing water.

[0033] (2) By introducing the concentration monitoring and image recognition technology, the waste heat recovery system of the blast furnace slag flushing water can monitor the water quality of the slag flushing water and the state of the heat exchanger in real time, and adjust the operation parameters in time; through the pre-training using the human-machine interactive monitoring and analysis, evaluating the image recognition neural network model for fouling on the inner wall of the heat pipe, obtaining the optimal network model, forming the optimal vibration descaling frequency and process, ensuring the heat transfer efficiency of the heat exchanger while reducing the number of descaling operations, and ensuring the system to operate in a high-efficiency and stable state; through the setting of the vibration actuator, when the fouling of the heat exchanger is identified at the monitoring end to reach the critical value, vibration descaling is carried out, which can not only prevent the fouling problem of the heat exchanger, extend the service life of the equipment, but also solve the problem that the heat pipe heat exchanger is difficult to be applied in the waste heat recovery of the blast furnace slag flushing water; in addition, through the reasonable adjustment of the three-way valve and the descaling mechanism by the deep learning mechanism, the thermal resistance of the heat transfer equipment in the blast furnace slag flushing water heat transfer process can be reduced, and the overall efficiency of the organic Rankine cycle system can be effectively improved. Description of the Drawings

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those skilled in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0035] Figure 1 It is a schematic diagram of a blast furnace slag water waste heat recovery system based on image recognition deep learning regulation provided by an embodiment of the present invention;

[0036] Figure 2 It is a schematic diagram of a control method based on image recognition deep learning regulation provided by an embodiment of the present invention. Detailed implementation manners

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the protection scope of the present invention.

[0038] In the following description, specific embodiments of the present invention will be described with reference to steps and symbols executed by one or more computers, unless otherwise specified. Therefore, these steps and operations will be mentioned several times as being executed by a computer. The computer execution referred to herein includes the operation of a computer processing unit representing electronic signals in a structured form of data. This operation transforms the data or maintains it at a position in the computer's memory system, which can be reconfigured or otherwise changed in a manner well known to those skilled in the art to change the operation of the computer. The data structure maintained by the data is the physical location of the memory, which has specific characteristics defined by the data format. However, the principles of the present invention are described in the above text, which does not represent a limitation. Those skilled in the art will understand that the following various steps and operations can also be implemented in hardware.

[0039] The term "module" or "unit" used herein can be regarded as a software object executed on the computing system. Different components, modules, engines, and services herein can be regarded as implementation objects on the computing system. The devices and methods herein are preferably implemented in software, and of course, can also be implemented in hardware, all within the protection scope of the present invention.

[0040] Those skilled in the art can understand that, unless specifically stated, the singular forms "a", "an", and "the" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of the present invention means the presence of features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or their groups. It should be understood that when we say an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more related listed items.

[0041] In the first embodiment of the present invention, there is provided a regulation-type blast furnace slag water waste heat recovery system based on image recognition deep learning. As Figure 1 shown, the system is divided into three parts: a slag water filtration system, an ORC organic Rankine system, and a cooling system;

[0042] Among them, the slag water filtration system includes a water pump, a primary sedimentation tank, a filter, a three-way regulating valve, a heat pipe, a vibration actuator, a monitoring end, and a descaling tank;

[0043] The ORC organic Rankine cycle includes a steam turbine, a condenser, a first circulation pump, a cooling pond, and a booster pump;

[0044] The cooling system includes a cooling tower, a second circulation pump, a valve, and a liquid storage tank;

[0045] In the embodiment of the present invention, an image recognition deep learning regulation-type descaling method includes the following steps:

[0046] S1: The concentration of the slag water pipeline is monitored in real time at node A. According to the comparison with the database data, if the concentration requirement is met, it enters the next node. If not, the slag water is sent to the primary sedimentation tank node through the three-way regulating valve. The comparison data results are constructed into a training database and imported into the neural network model. A generation model of the fouling wall thickness under the long-term operation of the slag water flow rate and concentration is established. According to the previous human-computer interaction training process, the parameters are optimized through the genetic algorithm;

[0047] Among them, the input parameters during the training process of the neural network model are: the flow rate and concentration of the slag water, and the output parameters are: the generation situation of the fouling wall thickness through image recognition within the same time. The stored training data is: the influence model of the concentration and flow rate on the wall thickness within a single time range, and the fouling wall thickness cleaning time node in the human-computer interaction mode.

[0048] S2: At node B, parameters are derived through the neural network model for image recognition of fouling on the inner wall of the heat pipe. Based on the real-time acquired image information, preliminary training is carried out using human-machine interactive monitoring and analysis. During the training process, the stochastic gradient descent algorithm is used for parameter optimization, and the optimal network model is obtained by evaluating the neural network model for image recognition of fouling on the inner wall of the heat pipe through human-machine interactive monitoring and analysis. When the judged data meets the requirements, heat exchange is carried out through the heat pipe. When the requirements are not met, the monitoring end controls the vibration actuator to vibrate the heat pipe wall surface for wall surface descaling, and through the three-way regulating valve, the descaling water is sent into the descaling pool.

[0049] In the embodiment of the present invention, in the slag flushing water filtration system, the path of the slag flushing water is in turn: water pump - primary sedimentation tank - filter - concentration monitoring node - three-way regulating valve - heat pipe - image recognition node - water pump - three-way regulating valve.

[0050] In the embodiment of the present invention, in the slag flushing water filtration system, a concentration real-time monitoring end is set at node A, and an image recognition monitoring end for fouling on the inner wall of the heat pipe is set at node B and fed back to the monitoring end.

[0051] In the embodiment of the present invention, if the particle concentration at node A does not meet the heat exchange requirements, the three-way regulating valve is regulated through the monitoring end, and the slag flushing water is sent into the primary sedimentation tank - filter in sequence.

[0052] In the embodiment of the present invention, if the fouling degree at node B does not meet the heat exchange requirements at the picture recognition monitoring end, the vibration actuator is regulated through the monitoring end to vibrate the wall surface to clean the wall surface scale layer, and through the three-way regulating valve, the slag flushing water is sent into the descaling pool in sequence.

[0053] In the embodiment of the present invention, historical data is used to train the neural network model for image recognition of fouling on the inner wall of the heat pipe. In the human-machine interaction mode, the real-time collected data is input into the neural network model for image recognition of fouling on the inner wall of the heat pipe with the optimal parameters trained with the assistance of humans in the early stage. By comparing the difference between the real-time data of fouling on the heat pipe wall surface and the model prediction result, it is determined whether there is an abnormality in the heat exchange state of the heat pipe; if the difference between the real-time data of the heat pipe wall surface or the heat exchange effect and the model prediction result exceeds the preset threshold, the diagnosis result is marked as abnormal. For the diagnosed abnormality, the neural network model for image recognition of fouling on the inner wall of the heat pipe further analyzes the cause, and solves the abnormal situation by automatically adjusting the vibration frequency and number of vibrations of the actuator or providing an alarm and operation suggestions to the operator.

[0054] In the second embodiment of the present invention, to facilitate better implementation of the system provided by the embodiment of the present invention, the embodiment of the present invention also provides a method based on the above system. The meanings of the nouns are the same as those in the above system, and the specific implementation details can refer to the description in the method embodiment.

[0055] Please refer toFigure 1 , such as Figure 1 shown in a diagram of a regulated blast furnace slag water waste heat recovery system based on image recognition deep learning. The system is divided into three parts, namely: I: slag water filtration system; II: ORC organic Rankine system; III: cooling system.

[0056] The slag water filtration system is sequentially connected by pipelines to a water pump 1, a primary sedimentation tank 2, a filter 3, a three-way regulating valve 501, a heat pipe 4, a vibration actuator 6, a monitoring end 7, a three-way regulating valve 502, and a descaling tank 8.

[0057] The ORC organic Rankine cycle is provided with a steam turbine 9, a condenser 10, a first circulation pump 1101, a cooling pond 12, and a booster pump 1102 connected to the heat pipe 4 of the slag water filtration system.

[0058] The cooling system is provided with a cooling tower 13, a second circulation pump, valves, and a liquid storage tank 14 connected to the condenser 10 of the ORC organic Rankine cycle system.

[0059] such as Figure 2 shown, its specific working process is as follows: The slag water is sent into the primary sedimentation tank 2 by the water pump 1 for sedimentation. After being filtered by the filter 3, it is monitored by the particle concentration monitoring system at node A. If the concentration of the slag water meets the heat exchange requirements, it is sent into the heat pipe 4 for heat exchange. If the particle concentration of the slag water does not meet the heat exchange requirements, the three-way regulating valve 501 is adjusted through the monitoring end 7, and the slag water is sent back into the primary sedimentation tank 2 for re-sedimentation. After being filtered by the filter 3 and monitored by the concentration monitoring system at node A, and meeting the heat exchange requirements, it is sent into the heat pipe 4 for heat exchange. Node B is an image recognition system for monitoring the fouling on the inner wall of the heat pipe. If it is recognized that the thickness of the inner wall of the heat pipe 4 cannot meet the heat exchange requirements, the vibration actuator 6 is regulated through the monitoring end 7 to remove the scale on the heat pipe wall surface. The waste water after descaling is sent into the descaling tank 8 through the regulation of the three-way regulating valve 502 by the monitoring end 7.

[0060] The organic working fluid exchanges heat with the slag water in the heat pipe 4. After the heat exchange is completed, a saturated organic working fluid steam is formed and sent into the steam turbine 9 for power generation. After being condensed by the condenser 10, it is sent into the cooling pond by the first circulation pump 1101, and after being pressurized by the booster pump 1102, it is sent back into the heat pipe 4 for heat exchange again, circulating in turn to complete the waste heat recovery of the slag water.

[0061] After the cooling water exchanges heat with the organic working fluid in the condenser 10, it is sent into the cooling tower 13 for cooling, sent into the liquid storage tank 14 for storage through the second circulation pump, and then sent into the condenser 10 for heat exchange, circulating in turn to complete the cooling of the organic working fluid.

[0062] In an embodiment of the present invention, the ORC organic Rankine system adopts a screw expander power generation device, and the organic working fluid can be selected from one or several optimized combinations of R245fa, R142b, R141b, R123, n-butane, and isopropane according to seasons and the temperature of slag flushing water.

[0063] The above has introduced in detail an image recognition deep learning-regulated blast furnace slag flushing water waste heat recovery system provided by an embodiment of the present invention. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A blast furnace slag flushing water waste heat recovery system based on image recognition deep learning control, characterized in that: include: Slag flushing water filtration system, ORC organic Rankine system and cooling system; The slag flushing water filtration system is used to monitor the concentration of the slag flushing water pipeline. If the slag flushing water concentration does not meet the preset concentration requirement, it is precipitated again until it meets the preset concentration requirement, and then the heat pipe is used for heat exchange. At the same time, the thickness of the dirt on the inner wall of the heat pipe is image-recognized. When the thickness of the dirt on the inner wall of the heat pipe does not meet the preset thickness requirement, descaling is performed until the thickness of the dirt on the inner wall of the heat pipe meets the preset thickness requirement. The ORC organic Rankine system is connected to the slag flushing water filtration system to complete the waste heat recovery of the slag flushing water; The cooling system is connected to the ORC organic Rankine system and is used to cool the organic working fluid.

2. The system according to claim 1, characterized in that The slag flushing water filtration system includes a water pump, a primary sedimentation tank, a filter, a three-way regulating valve, a heat pipe, a vibration actuator, a monitoring terminal and a descaling tank; The slag flushing water is pumped into the primary sedimentation tank for sedimentation. After being filtered through the filter, the concentration of the slag flushing water pipeline is monitored. If the concentration of the slag flushing water meets the preset concentration requirement, it is sent to the heat pipe for heat exchange. If the concentration of the slag flushing water does not meet the preset concentration requirement, the three-way regulating valve is adjusted by the monitoring end to send the slag flushing water to the primary sedimentation tank for sedimentation again. After being filtered through the filter, the concentration of the slag flushing water pipeline is monitored until it meets the preset concentration requirement, and then it is sent to the heat pipe for heat exchange. Then, the thickness of the dirt on the inner wall of the heat pipe is image recognized. When the thickness of the dirt on the inner wall of the heat pipe does not meet the preset thickness requirement, the vibration actuator is adjusted by the monitoring end to descale the wall of the heat pipe. The descaling wastewater is sent to the descaling tank through the three-way regulating valve regulated by the monitoring end.

3. The system according to claim 2, characterized in that The ORC organic Rankine system includes a steam turbine connected to the heat pipe of the slag flushing water filtration system, a condenser, a first circulation pump, a cooling tank and a booster pump; The organic working fluid exchanges heat with the slag flushing water in the heat pipe. After the heat exchange is completed, saturated organic working fluid steam is formed and sent to the steam turbine for power generation. After condensation by the condenser, it is sent to the cooling pool through the first circulation pump. After being pressurized by the booster pump, it is sent to the heat pipe again for heat exchange, and circulated in sequence to complete the waste heat recovery of the slag flushing water.

4. The system according to claim 3, characterized in that The cooling system includes a cooling tower connected to the condenser of the ORC organic Rankine system, a second circulation pump, a valve and a liquid storage tank; After the cooling water exchanges heat with the organic working fluid in the condenser, it is sent to the cooling tower for cooling, and is sent to the liquid storage tank for storage through the second circulation pump and valve. Finally, it is sent to the condenser for heat exchange and circulated in sequence to complete the cooling of the organic working fluid.

5. The system according to claim 4, characterized in that The monitoring of the slag flushing water pipeline includes: Monitor the flow rate of the slag flushing water pipeline and store the concentration and flow rate information in the basic database. Monitor the concentration of the slag flushing water pipeline and compare it with the data in the database to determine whether it meets the preset concentration requirements.

6. The system according to claim 5, characterized in that The comparison of the concentration of the slag flushing water pipeline with the data in the database includes: A neural network model is used to compare the concentration of slag flushing water pipeline with the data in the database.

7. The system according to claim 6, characterized in that The image recognition of the thickness of the dirt on the inner wall of the heat pipe includes: The heat pipe inner wall dirt neural network model performs image recognition on the thickness of the heat pipe inner wall dirt to determine whether the thickness of the heat pipe inner wall dirt meets the preset thickness requirement.

8. The system according to claim 7, characterized in that The heat pipe inner wall dirt neural network model is based on real-time acquired heat pipe inner wall image information, and uses human-computer interactive monitoring and analysis for preliminary training. During the training process, a stochastic gradient descent algorithm is used to optimize parameters. The heat pipe inner wall dirt neural network model is evaluated through human-computer interactive monitoring and analysis to obtain the optimal network model.

9. The system according to claim 8, characterized in that The method of training the heat pipe inner wall fouling neural network model with historical data includes: Based on the wall thickness generation model obtained by image recognition under different flow rates and concentrations of slag flushing water pipelines and long-term operating conditions, a generation model of flow rate, concentration and dirt wall thickness under time span is established and stored in the basic database.

10. The system according to claim 9, characterized in that The heat pipe inner wall fouling neural network model is trained using historical data. In the human-computer interaction mode, the heat pipe inner wall fouling image collected in real time is imported into the heat pipe inner wall fouling neural network model with the best trained parameters with human assistance in the early stage. By comparing the difference between the real-time data of heat pipe wall fouling and the predicted results of the heat pipe inner wall fouling neural network model, it is determined whether there is an abnormality in the heat exchange state of the heat pipe; if the difference between the real-time data of the heat pipe wall or heat exchange effect and the predicted results of the heat pipe inner wall fouling neural network model exceeds the preset threshold, the diagnosis result is marked as abnormal. For the diagnosed abnormality, the heat pipe inner wall fouling neural network model further analyzes the cause, and automatically adjusts the vibration frequency and number of vibration executions or provides alarms and operation suggestions to the operator.

Images