Energy pile green low-carbon building heat exchange pipe maintenance and intelligent operation and maintenance control method

Abstract

The present application belongs to the field of geothermal energy development and utilization, pile foundation engineering and building energy-saving technology, and proposes a heat exchange pipe maintenance and intelligent operation and maintenance control method for energy pile green low-carbon buildings. Preliminary judgment and preliminary alarm positioning are made, and the heat exchange pipe maintenance device is used for fine repair; the learning and calculation of the energy pile buried pipe end control module, the unit equipment control module and the indoor user end control module are realized on the client side; the client calculates the desensitization parameters and summarizes them to the central server for calculation, and then updates the local model to each client until the global shared model is stable. The heat exchange pipe maintenance device based on the principle of electromagnetic induction includes a double-layer intelligent heat exchange pipe structure, a miniature magnetoelectric sensor containing a dredging module, a cable storage tray, a data automatic acquisition instrument and a control end. The present application can realize advanced control management of energy pile green low-carbon buildings, accurate positioning and repair of damaged points; and can reduce energy waste during operation and maintenance stage, and achieve the effect of precise control of green low-carbon buildings.

Description

Technical Field

[0001] This invention relates to the fields of geothermal energy development and utilization, pile foundation engineering, and building energy conservation technology, and in particular to a method for the maintenance and intelligent operation and maintenance control of heat exchange tubes in energy pile green and low-carbon buildings. Background Technology

[0002] Ground source heat pump technology extracts recyclable shallow geothermal energy to regulate indoor building temperatures, essentially functioning as a seasonal energy storage mechanism. Currently, ground source heat pumps are considered the most promising energy-saving and environmentally friendly HVAC technology for achieving green and low-carbon buildings. However, in practice, traditional ground source heat pump systems typically require a large enough land area to install buried pipe heat exchangers, resulting in relatively high initial investment costs. Compared to traditional cooling and heating methods, their advantages are significantly reduced, thus becoming one of the main obstacles restricting the development of ground source heat pump systems.

[0003] Energy piles are an economical, efficient, energy-saving, and emission-reducing new technology that combines ground source heat pump technology with traditional pile foundations. By embedding heat exchange pipes of various shapes within the pile foundation, shallow geothermal energy is converted, satisfying the conventional mechanical functions of pile foundations while simultaneously achieving heat exchange with shallow geothermal energy through the pile body. This serves the dual purpose of simultaneous construction of the pile foundation and the buried pipe heat exchanger. Compared to traditional ground source heat pump technology, it offers significant advantages in investment cost, heat exchange efficiency, and land utilization. The management and maintenance of energy piles during the green and low-carbon building operation phase are crucial for reducing system carbon emissions and extending system lifespan. On the other hand, during the construction or operation of the energy pile system, damage, blockage, or bending of the heat exchange pipes within the pile can hinder the normal circulation of the heat exchange medium, leading to unstable water pressure, corrosion and damage to the main load-bearing components of the pile foundation, and reduced heat exchange efficiency, ultimately causing the entire system to malfunction. However, current technologies for the inspection and maintenance of energy pile systems after construction are still immature.

[0004] Chinese invention patent application number CN202111526641.7, entitled "Fuzzy Control Method and Device for Solar-Ground Source Heat Pump Heating System", establishes a mathematical model of the solar-ground source heat pump heating system and determines input variables and output variables; fuzzifies the input variables and output variables respectively to generate corresponding fuzzy sets; and generates fuzzy control rules based on the fuzzy sets to realize partial variable frequency fuzzy control of the solar-ground source heat pump heating system.

[0005] Chinese invention patent application number CN201610200872.1, entitled "A Ground Source Heat Pump Control Method and System", discloses a method that uses the predicted value of room temperature instead of the real-time temperature value as part of the input of the heat pump unit controller, thereby realizing pre-control.

[0006] Chinese invention patent application number CN202010244228.0, entitled "A Method and Apparatus for Performance Testing and Optimization Control of a Ground Source Heat Pump System," discloses a method for calculating the real-time unit energy efficiency ratio of the system based on collected real-time operating data, thereby further determining the real-time operating status of the system. The system adjustment mode is determined by comparing historical operating data.

[0007] Chinese invention patent application number CN201810818755.0, entitled "Energy Pile Test Monitoring System and Test Method", discloses a test monitoring system including an energy pile body, a loading device, a pile top displacement testing device, a circulating temperature control device, a data acquisition device, and a geotechnical thermal response testing device, which can test and analyze the thermodynamic parameters of the energy pile.

[0008] The aforementioned control methods for ground source heat pump systems primarily focus on controlling the heat pump unit, with few addressing the operation and maintenance control of energy pile-based green and low-carbon buildings. In practical engineering, the operation of energy pile-based green and low-carbon buildings involves dynamic coordination between the indoor user end, the heat pump unit end, and the energy pile buried pipe end. Controlling and managing only the heat pump unit cannot achieve the overall robustness and low-carbon energy saving of the building. Therefore, an easily scalable and simple collaborative control and management method is needed for precise control in practical engineering projects.

[0009] Machine learning is a scientific technique that uses appropriate algorithms to enable computers to automatically analyze and obtain patterns from a type of data, and then use these patterns to predict unknown data.

[0010] Chinese invention patent application number CN201910043376.3, entitled "A Performance Prediction Method for Ground Source Heat Pump Systems", discloses a ground source heat pump performance prediction method based on a data structure including borehole distribution pattern, borehole radius, buried pipe depth, number of buried pipes, lateral spacing of buried pipes, longitudinal spacing of buried pipes, thermal conductivity of the filling material inside the pipe, nominal outer diameter of the U-shaped pipe, and temperature of the soil and rock at the far end.

[0011] Chinese invention patent application number CN202111450170.6, entitled "Control Method and System for Solar-Ground Source Heat Pump System Combining Resource Prediction", discloses a machine learning prediction model for ground source heat pump air conditioning load, and performs system control based on the predicted load.

[0012] The aforementioned machine learning-based ground source heat pump performance prediction and control methods only predict the heat exchange performance of borehole buried pipes and the load of the water collector, and still cannot obtain advanced decision-making for control and management methods. They are not applicable to the intelligent control of green and low-carbon buildings with energy piles, and cannot formulate advanced control schemes that match the three parts of the unit equipment, indoor users, and energy piles, nor can they issue fault alarms.

[0013] Chinese utility model patent application number CN201721077937.4, entitled "A Permanently Detectable Polyethylene Composite Pipe", discloses a permanently detectable polyethylene composite pipe, comprising a double-layer structure pipe composed of a polyethylene layer and a modified PP coating layer. The pipe body is damaged by detecting the current of the tracer metal strip between the inner and outer layers using a multimeter.

[0014] Chinese invention patent application number CN202111660338.6, entitled "Anti-blockage and dredging system for municipal road engineering and its working method", discloses an anti-blockage and dredging system for municipal road engineering composed of a dredging trolley, a trolley drive mechanism, a visual inspection mechanism, an industrial and mining inspection mechanism and a display mechanism.

[0015] In the aforementioned composite pipeline technologies, the metal strips require assembly and connection, which can easily lead to short circuits or open circuits during the construction phase. Furthermore, municipal pipeline dredging systems are limited to large-diameter municipal pipe networks. Existing technologies cannot be applied to the inspection and maintenance of deeply buried, small-diameter energy pile heat exchanger pipes. Summary of the Invention

[0016] To overcome the following shortcomings and problems of existing technologies: (1) Traditional control and management methods are based on real-time control using electrical signals, and it is difficult to achieve advance warning functions; (2) The heat exchange tubes inside the energy pile are permanent or semi-permanent designs, and cannot be repaired after a failure. Therefore, this invention provides a method for the inspection and intelligent operation and maintenance control of heat exchange tubes in green and low-carbon buildings for energy piles.

[0017] The technical solution of the present invention is: a method for the inspection and intelligent operation and maintenance control of heat exchange tubes in a green and low-carbon building with energy piles, including a control operation method and a heat exchange tube inspection device based on the principle of electromagnetic induction; a preliminary judgment is made and a preliminary alarm location is made according to the control operation method, and on this basis, the heat exchange tube inspection device based on the principle of electromagnetic induction is arranged to carry out refined repair work.

[0018] The control operation method establishes a federated learning algorithm framework for the operation control of green low-carbon buildings with energy piles based on a client-server architecture; the learning and calculation of all sub-modules are implemented on the client, including three major modules: the energy pile buried pipe end control module, the unit equipment control module, and the indoor user end control module; the desensitized parameters calculated by the client are aggregated to the central server for calculation, and then distributed to each client to update its local model until the globally shared model is robust.

[0019] The specific steps are as follows:

[0020] D1: Acquire indoor user terminal operating status parameters, indoor user terminal control parameters, unit operating status parameters, unit control parameters, energy pile buried pipe terminal operating status parameters, and energy pile buried pipe terminal control parameters; preprocess the data of each parameter, analyze the data characteristics, and allocate them to the corresponding clients;

[0021] D2: Each client updates its local model with the following parameters: the local support vector regression model for energy pile buried pipe control, the local decision tree model for unit equipment control, and the local random forest model for indoor user terminal control.

[0022] D3: The three clients—the energy pile buried pipe end control module, the unit equipment control module, and the indoor user control module—upload encrypted and de-identified parameters to the central server using public keys; the relevant data for client i is represented as follows: Can be encrypted as

[0023] D4: The central server uses its private key to decrypt the encrypted and anonymized parameters uploaded by the three clients, performs secure aggregation, and then updates the global shared model; the central server then... Decryption of encrypted and anonymized parameters is performed. Federated learning on the central server obtains a globally shared model using one or more synchronization methods, including gradient averaging, federated averaging, and knowledge distillation. The global model update process in the nth round is as follows:

[0024] Preferably, all client models are aggregated through federated learning based on knowledge distillation, according to the formula... Perform global model weight updates on three clients, where W... n+1 These are the global model parameters for the nth round. The client sub-model weights are uploaded to the server by client i in the nth round. After each round of model weight updates, the central server calculates the global model error and accuracy. The central server can also control the transmission speed and the shutdown of model training.

[0025] D5: The central server generates a public key for encrypting data transmission from the global shared model and distributes it to each client; it updates the iteration results of other clients as new sample data attribute parameters based on the global shared model and the local models of each client; for example, the energy pile buried pipe end control module support vector regression model client updates the unit equipment control status parameters and indoor user end equipment status parameters based on the global shared model.

[0026] D6: Repeat steps D2 to D5 iteratively until the global shared model is robust. Finally, the client calculates the corresponding results based on the global shared model. Calculate the matching control states in the green and low-carbon building of the energy piles, including the control parameters of the energy pile buried pipe end, the unit equipment control parameters, and the indoor user end control parameters. Specifically, this includes the set temperature of the fan coil unit, the opening degree of the indoor return water control valve, the opening degree of the indoor supply water control valve, the opening degree of the indoor user end water collector manifold control valve, the opening degree of the water collector branch pipe control valve, the opening degree of the indoor user end water distributor manifold control valve, the opening degree of the water distributor branch pipe control valve, the opening degree of the makeup water pump control valve, the opening degree of the cooling circulation pump control valve, the opening degree of the chilled water circulation pump control valve, the opening degree of the expansion valve, the opening degree of the buried pipe end water collector manifold control valve, the opening degree of the water collector branch pipe control valve, the opening degree of the buried pipe end water distributor manifold control valve, the opening degree of the water distributor branch pipe control valve, the opening degree of the energy pile return water control valve, and the opening degree of the energy pile supply water control valve.

[0027] The energy pile buried pipe end control module uses indoor user terminal operating status parameters, unit operating status parameters, and energy pile buried pipe end operating status parameters as sample attribute data to establish a support vector regression local model for energy pile buried pipe end control, predict and calculate the control scheme, and encrypt and upload it to the central processing unit; the specific steps are as follows:

[0028] D2.1.1: Decrypt the public key of the encrypted data using the private key, update the iteration results of the energy pile buried pipe end control module, unit equipment control module, and indoor user control module client according to the global shared model of the central server as the new sample data attribute parameters, and use the water level, pressure difference, manifold flow, manifold pressure, manifold temperature, buried pipe water supply temperature, buried pipe return water temperature, buried pipe water supply flow, buried pipe return water flow, buried pipe water supply pressure, and buried pipe return water pressure as the sample attribute space set, and use the energy pile buried pipe end control parameters as the learning target;

[0029] D2.1.2: To resolve the relationship between the operating state parameters and control parameters of the energy pile buried pipe end, a relaxation variable ξ is introduced. i{ ξ i * A nonlinear segmentation support vector classifier considering soft margins is constructed using the penalty coefficient C; the prediction accuracy and stability of the local support vector regression model for the control parameters of the energy pile buried pipe end are represented by the loss function; the loss function is converted into a conditional extremum function.

[0030]

[0031] Where w is the normal vector of the hyperplane, C is the penalty coefficient, and ξ, ξ * y is the relaxation factor, ∈ is the hyperparameter that determines the boundary width; i The actual test results for the training samples; f(X)i )=w·Φ(X)+b is the classification hyperplane of the support vector regression local model for the control of the buried pipe end of the energy pile; i is the training sample number; N is the number of training set samples; Φ(X) is the nonlinear mapping function;

[0032] D2.1.3: The above conditional extremum function is transformed into a multivariate function by using the Lagrange function. The partial derivatives of the Lagrange function with respect to the optimization objectives w, b, ξ are set to 0 to obtain the Lagrange multipliers. The conditional extremum function is then transformed into a dual function, thereby finding the minimum value of the prediction boundary.

[0033] D2.1.4: The nonlinear mapping function Φ(X) contained in the classification hyperplane of the local support vector regression model of the energy pile buried pipe control end, and its inner product φ(X) i ) T φ(X j The processing is carried out using a combination of kernel functions: the combination kernel function is one or more of the following: Gaussian kernel, linear kernel, polynomial kernel, and Sigmoid kernel;

[0034] D2.1.5: Optimize the model parameters in the local support vector regression model for the control of the buried pipe end of the energy pile: insensitive loss function ∈, penalty coefficient C, hyperparameters γ, λ, α, c, d in the kernel function; select one or more optimization methods from simulated annealing, grid search, particle swarm optimization, PSO algorithm, and genetic algorithm;

[0035] D2.1.6: Input the operating status parameters of the target energy pile buried pipe end, and use the trained support vector regression model to predict and calculate the control parameters of the target building's energy pile buried pipe end;

[0036] D2.1.7 Compare the calculation result of step D2.1.6 with the device control threshold. If the calculation result meets the control threshold, the predicted calculation result is used as advance prediction control information and converted into encrypted parameters using the public key and uploaded to the central server for iteration of the globally shared model of federated learning. If the calculation result exceeds the control threshold, then step D2.1.8 is executed.

[0037] D2.1.8: Output the calculated control parameters of the buried pipe end of the energy pile and issue an alarm. Use the equipment control threshold as the advanced prediction control information and convert it into encrypted parameters using the public key to upload to the central server for the global shared model iteration of federated learning. At the same time, arrange the heat exchange tube maintenance device based on the principle of electromagnetic induction to carry out fine detection and positioning.

[0038] Preferably, the model parameters are optimized using a segmented simulated annealing method, and the steps are as follows:

[0039] D2.1.5.1: Randomly generate an initial parameter set for interactive verification. The error value EEP is denoted as the current annealing system state E0, the initial temperature T0, the temperature of the first annealing stage T1, and the annealing end temperature T2.

[0040] D2.1.5.2: According to the perturbation algorithm m′ i =m i +s·(μ-0.5)(B i -A i The model parameters are perturbed to form a new parameter set, and the current annealing system state E is obtained through interactive verification. n Calculate ΔE = E n -E n-1 .

[0041] D2.1.5.3: If ΔE < 0, accept the new parameter set and jump to step D2.1.5.5; otherwise, accept the corresponding parameter set according to the Metropolis criterion exp(ΔE / kT)-μ > 0 and jump to step D2.1.5.5; if none of the above conditions are met, reject the critical state and return to step D2.1.5.2, re-perturb to generate a new parameter set, and perform interactive verification until the parameter set acceptance condition in step D2.1.5.3 is met.

[0042] D2.1.5.4: Upon obtaining a new state, proceed according to the cooling plan. Cool down the temperature. If the set temperature T1 is not reached, return to step D2.1.5.2. When the set temperature T1 is reached, start a new annealing plan.

[0043] D2.1.5.5: Following the new perturbation method and annealing plan, continue to perturb the parameter set of the first-stage annealing and calculate the corresponding state parameters E. n .

[0044] D2.1.5.6: If ΔE < 0, accept the new parameter set and jump to step D2.1.5.7; otherwise, accept the corresponding parameter set according to the Metropolis criterion and jump to step D2.1.5.7; if none of the above conditions are met, reject the parameter set, return to step D2.1.5.6, re-perturb to generate a new parameter set, and perform interactive verification until the parameter set acceptance condition in step D2.1.5.9 is met.

[0045] D2.1.5.7 sets the end temperature to T2 as the algorithm exit point and the global maximum number of EEP calculations to N. Annealing stops when T2 or N is reached, and the cross-validation error E at this critical state is accepted. n It should be the lowest E nThe corresponding parameters should be the optimal prediction parameters; otherwise, return to step D2.1.5.5.

[0046] The unit equipment control module uses indoor user terminal operating status parameters, unit operating status parameters, and energy pile buried pipe terminal operating status parameters as sample attribute data to establish a decision tree local model of unit equipment operating control parameters. This model determines the control parameters and valve openings of the heat pump and circulating pump, and then encrypts and uploads them to the central server. The decision tree model is generated based on one or more of the ID3, C4.5, and CART algorithms. The specific steps are as follows:

[0047] D2.2.1: Decrypt the public key of the encrypted data using the private key, update the iteration results of the energy pile buried pipe terminal control module, unit equipment control module, and indoor user control module client according to the global shared model of the central server as the new sample data attribute parameters; calculate the information gain of the unit operation status parameters of the training samples according to the information entropy, and sort all unit operation status parameters according to the information gain.

[0048] The method for calculating information entropy is as follows;

[0049] Total information entropy of a given sample:

[0050] Sample subset {s 1j ,s 2j ,…,s mj Total information entropy: P ij For sample subsets s j Category C i The sample probability.

[0051] Information entropy value of the samples based on attribute A: The corresponding information gain is: Gain(A) = I(s1,s2,…,s) m )-E(A).

[0052] D2.2.2: Select the unit operating state parameter with the largest information gain as the optimal attribute, divide the samples based on the optimal attribute, and form the root node by taking the samples with the same optimal attribute value as the same sample set.

[0053] D2.2.3: Treat each root node as a complete dataset, divide the samples according to the suboptimal attribute, and form leaf nodes by taking samples with the same suboptimal attribute value as the same sample set. Use the REP method, PEP method or MEP method to prune the decision tree, and iterate to form a decision tree.

[0054] D2.2.4: Compare the determined unit equipment control parameters with the equipment control thresholds. If the calculation result meets the equipment control threshold, the decision calculation result is used as advanced predictive control information, which is converted into encrypted parameters using a public key and uploaded to the central server for iteration of the globally shared model of federated learning. If the calculation result exceeds the equipment control threshold, proceed to step D2.2.5.

[0055] D2.2.5: Output the unit equipment control results of the decision and issue an alarm. Use the equipment control threshold as the advanced predictive control information, convert it into encrypted parameters using the public key, and upload it to the central server for the global shared model iteration of federated learning.

[0056] The indoor user terminal control module, based on the indoor user terminal operating status parameters and the unit operating status parameters, establishes an indoor user terminal control random forest local model on the client side, using the indoor user terminal operating status parameters as sample attributes. This model calculates the indoor equipment set temperature and chilled water control parameters as the optimal control scheme and uploads it to the central server in encrypted form. The specific steps are as follows:

[0057] D2.3.1: Decrypt the public key of the encrypted data using the private key, and update the iteration results of the energy pile buried pipe terminal control module, unit equipment control module, and indoor user control module client according to the global shared model of the central server as the new sample data attribute parameters; use the Bootstrap sampling method to randomly generate a subset of samples from the indoor user terminal control training samples as the training samples of one of the decision tree models, and repeat the sampling k times to form k decision tree training samples.

[0058] D2.3.2: Train decision trees based on a subset of attributes from k decision tree training samples to form k independent random decision trees;

[0059] D2.3.3: The random forest randomly selects a classifier to vote on the indoor user terminal control schemes predicted by k decision trees, and the voting result is taken as the optimal control scheme for the indoor user terminal.

[0060] D2.3.4: Compare the indoor user terminal control parameters determined in the optimal control scheme with the equipment control threshold. When the calculation result meets the equipment control threshold, the decision calculation result is used as advanced predictive control information, which is converted into encrypted parameters using a public key and uploaded to the central server for iteration of the globally shared model of federated learning. When the calculation result exceeds the equipment control threshold, proceed to step D2.3.5.

[0061] D2.3.5: Output the indoor user terminal control results and issue an alarm. Use the equipment control threshold as the advanced predictive control information, convert it into encrypted parameters using the public key, and upload it to the central server for global shared model iteration in federated learning. At the same time, arrange a heat exchange tube maintenance device based on the principle of electromagnetic induction to perform fine detection and positioning.

[0062] The data cleaning targets include erroneous data, abnormal data, incomplete data, and duplicate data. Data cleaning involves identifying invalid, outlier, and missing values; and processing outliers, invalid, and missing values.

[0063] Optionally, the identification of invalid, outlier, and missing values ​​can be based on one of the following: 1. Box plots calculate the maximum, minimum, median, and upper and lower quartiles of the dataset. Outliers or invalid values ​​can be identified based on the upper and lower quartiles of the box plot; 2. The 3σ rule is often used when the data follows a normal distribution. In this case, outliers are defined as values ​​in a set of measurements that deviate from the mean by more than three times the standard deviation; 3. Scatter plots visually identify outliers, i.e., invalid values, by showing the positional relationship between two sets of data.

[0064] Optionally, the handling of invalid, outlier, and missing values ​​can be based on one of the following: 1. Direct deletion; 2. Filling in invalid, outlier, and missing values ​​according to the statistical data characteristics of mean, median, and mode; 3. Fitting invalid, outlier, and missing values ​​according to regression model or maximum likelihood estimation.

[0065] The indoor user terminal operating status parameters include: indoor ambient temperature, outdoor ambient temperature, personnel behavior status, equipment operating status, indoor user terminal manifold water level, differential pressure, manifold flow rate, manifold pressure, manifold temperature, indoor supply water temperature, indoor return water temperature, indoor supply water flow rate, indoor return water flow rate, indoor supply water pressure, and indoor return water pressure; the unit operating status parameters include: condenser working liquid level, condenser inlet temperature, evaporator working liquid level, evaporator inlet temperature, and compressor pressure; the energy pile buried pipe terminal operating status parameters include: buried pipe terminal manifold water level, differential pressure, manifold flow rate, manifold pressure, manifold temperature, buried pipe supply water temperature, buried pipe return water temperature, buried pipe supply water flow rate, buried pipe return water flow rate, buried pipe supply water pressure, and buried pipe return water pressure.

[0066] The indoor user-end control parameters include: the set temperature of the fan coil unit, the opening degree of the indoor return water control valve, the opening degree of the indoor supply water control valve, the opening degree of the indoor user-end water collector manifold control valve, the opening degree of the water collector branch pipe control valve, the opening degree of the indoor user-end water distributor manifold control valve, and the opening degree of the water distributor branch pipe control valve; the unit control parameters include: the opening degree of the makeup water pump control valve, the opening degree of the cooling circulation pump control valve, the opening degree of the chilled water circulation pump control valve, and the opening degree of the expansion valve; the energy pile buried pipe end control parameters include: the opening degree of the buried pipe end water collector manifold control valve, the opening degree of the water collector branch pipe control valve, the opening degree of the buried pipe end water distributor manifold control valve, the opening degree of the water distributor branch pipe control valve, the opening degree of the energy pile return water control valve, and the opening degree of the energy pile supply water control valve.

[0067] The heat exchanger tube maintenance device based on the electromagnetic induction principle includes a double-layer intelligent heat exchanger tube structure 1, a miniature magnetoelectric sensor 3 with a dredging module, a cable tray 4, an automatic data acquisition device 5, and a control terminal 6. The double-layer intelligent heat exchanger tube structure 1 is U-shaped, with its two ends connected to the automatic data acquisition device 5 via the cable tray 4, and the automatic data acquisition device 5 is connected to the control terminal 6. The double-layer intelligent heat exchanger tube structure 1 consists of an inner polyethylene tube 1-A, an outer polyethylene tube 1-B, and a magnetic induction ring 2. The inner polyethylene tube 1-A is located inside the inner wall of the outer polyethylene tube 1-B, and the two are filled with lubricating oil and threaded together. The magnetic induction ring 2 is embedded in the wall of the outer polyethylene tube 1-B and is arranged at equal intervals along the axial direction of the double-layer intelligent heat exchanger tube structure 1. The miniature magnetoelectric sensor 3 with a dredging module moves inside the inner polyethylene tube 1-A and includes a dredging component, a series motor 3-5, and a drive wheel 3. -7 and cable 3-8; one end of series motor 3-5 is connected to cable 3-8, and the other end is fixed to the unblocking component through mounting gear 3-3 and connecting component 3-2; multiple drive wheels 3-7 are connected to the side of series motor 3-5 and slide on the inner wall of inner polyethylene pipe 1-A; magnetoelectric sensor 3-6, surround camera 3-4 and lighting lamp 3-9 are all installed on the surface of series motor 3-5; when an electrical signal is passed into magnetoelectric sensor 3-6, the electromagnetic frequency changes when it passes through magnetic induction ring 2, and the electrical signal is transmitted to data automatic acquisition instrument 5 through the connected cable 3-8; data automatic acquisition instrument 5 automatically records the time when magnetoelectric sensor 3-6 passes through each magnetic induction ring 2 in sequence, and transmits this information to control terminal 6. When no electrical signal is transmitted to the next magnetic induction ring 2 after passing through a certain magnetic induction ring 2 for a set time, it is determined that there is a blockage between the two magnetic induction rings 2. The detection principle is as follows: When an electrical signal is applied to the magnetoelectric sensor, the electromagnetic frequency will change when the sensor passes through the magnetic induction ring. The electrical signal is transmitted to the automatic data acquisition instrument 5 through the connected cable. The automatic data acquisition instrument 5 automatically records the time when the sensor passes through each magnetic ring in sequence and transmits this information to the control terminal. If there is no electrical signal transmission to the next measurement point for a long time after passing through a certain measurement point, it is determined that there may be a blockage between the two measurement points.

[0068] The magnetic induction ring 2 has a width of 5-8 mm and is integrally formed with the outer polyethylene pipe 1-B; the spacing of the magnetic induction rings 2 is 0.3-0.5 m. The magnetic induction rings are arranged at equal intervals along the axial direction of the heat exchange pipe, and the spacing is determined according to the design requirements of the energy pile. The inner wall of the outer polyethylene pipe and the outer wall of the inner polyethylene pipe are threaded, lubricating oil is applied between the two layers of pipe, and the two layers of pipe are tightened using the threads.

[0069] The unblocking component is a spiral metal wire 3-1 and / or a hot melt rod.

[0070] The heat exchanger tube inspection device based on the principle of electromagnetic induction is used to detect blockages and damage in heat exchanger tubes, and includes the following steps:

[0071] S1: After the initial fault alarm is issued by the Energy Pile Green Building Federation Learning Operation Control, before the test begins, turn on the switch of the automatic data acquisition instrument 5, and use a magnetic induction ring to move around the magnetic electric sensor 3-6. When the magnetic induction ring encounters the sensing point of the magnetic electric sensor 3-6, an audible and visual alarm will be issued, and the automatic data acquisition instrument 5 will show an indication. If the automatic data acquisition instrument 5 is working normally, proceed with the subsequent steps.

[0072] S2: During testing, a vertically arranged pipeline is placed inside the energy pile. The miniature magnetoelectric sensor 3, which includes a dredging module, is inserted along the pipeline. The magnetoelectric sensor 3-6 is inserted by its own gravity and the cable tray 4 rotates. When the center of the magnetoelectric sensor 3-6 intersects with the magnetic induction ring 2, the automatic data acquisition instrument 5 emits a beep and is accompanied by a light indicator. The surround camera 3-4 is activated to monitor the pipeline in real time, including images and sound waveforms. The automatic data acquisition instrument 5 automatically records the number of each measurement point on the magnetic induction ring 2, the time interval between two adjacent measurement points on the magnetic induction ring 2, and real-time images and sound waveform information.

[0073] S3: The automatic data acquisition instrument 5 transmits the automatically recorded data to the control terminal 6, and uses one or more of the following methods to analyze the acquired data and accurately locate the damage point: isolated forest algorithm, X-ray digital imaging technology, ultrasonic detection, ultrasonic guided wave, ultrasonic C-scan, ultrasonic phased array, and high temperature thickness measurement method; the control terminal 6 accurately locates the blockage point based on the abnormal time and location of the analyzed data.

[0074] In step S3, when analyzing the collected data using the isolated forest algorithm, an isolated tree is built in the attribute space of the sample set to isolate attributes. Within the same attribute space, outliers are separated through model training, while normal values ​​are located in deeper child nodes of the isolated tree. The specific steps are as follows:

[0075] S3.1: The current signal, acoustic signal, and electromagnetic signal transmitted from the magnetoelectric sensor 3-6 to the control terminal 6 are used as sample attributes to form a sample set, represented as X=(x1,x2,x3,…,x…). n), where x n =(x n1 ,x n2 ,x n3 ,…,x nm ), x nm The nth type attribute signal at time m;

[0076] S3.2: Use the attribute dimension q as the attribute cut point to segment attribute types, forming different isolated trees. For each isolated tree, use the attribute value cut point p as the outlier cut point to segment the attribute values ​​within the attribute space; Q is a random value belonging to the range (0, n), and p is a random value belonging to the range min(x). n ) < p < max(x) n Random values ​​within a given range;

[0077] S3.3: Based on the split of attribute cut point q and outlier cut point p, a hyperplane is formed in the sample space. Samples smaller than the outlier cut point p are classified into one class and form the left node of the isolated tree, and vice versa.

[0078] S3.4: Repeat steps S3.2 to S3.3 until each isolated tree leaf node contains only one sample, thus completing the training of the isolated tree model;

[0079] S3.5: Take the sample data x at each time step i By traversing and calculating the data in an isolated tree model, the average height h(x) of each sample data can be obtained. i The outlier score of the sample data is calculated using the following formula:

[0080]

[0081] Where h(x) is the height of x in each isolated tree, c(n) is the average path length for a given number of samples n, and E(h(x)) is the expected path length of x in multiple isolated trees;

[0082] When the score S(x,ψ) is greater than 0.75, the signal corresponding to the sample at that moment is considered an abnormal signal;

[0083] When the score S(x,ψ) is between 0.4 and 0.75, the signal corresponding to the sample at that time is considered a suspected abnormal signal;

[0084] When the score S(x,ψ) is less than 0.4, the sample at that moment is considered a normal signal.

[0085] The heat exchanger tube maintenance device based on the principle of electromagnetic induction is used for unblocking and repairing heat exchanger pipelines, and includes the following steps:

[0086] T1: Turn on the surround camera 3-4 and the lighting 3-9, and transmit the data to the control terminal 6 in real time, so that the surround camera 3-4 is aimed at the blockage;

[0087] T2: Begin unblocking work. When the blockage is found to be caused by pipe heat fusion, turn on the series motor 3-5 controlling the spiral metal wire 3-1. After the spiral metal wire 3-1 rotates to the working speed, continue to use the gravity of the magnetoelectric sensor 3-6 to slowly lower the sensor cable 3-8. The spiral metal wire 3-1 stirs the blockage until the blockage is cleared. When the blockage is caused by pipe heat fusion, replace the spiral metal wire 3-1 and connecting component 3-2 of the unblocking module with a heat fusion rod and place the magnetoelectric sensor 3-6 at the blockage again. Turn on the heating rod to the heat fusion temperature of the polyethylene material and use the gravity of the magnetoelectric sensor 3-6 to slowly lower the sensor cable 3-8 until the blockage is cleared.

[0088] T3: After removing the miniature magnetoelectric sensor 3 containing the unblocking module from the pipe, perform pipe pressure testing and flushing to clean the foreign objects remaining in the pipe, and the unblocking work is completed.

[0089] T4: After accurately locating the damage point, repair the heat exchange pipeline. The damaged section of the pipe is unscrewed along the threads to replace it, or an automatic repair agent is used to repair the damaged section. This heat exchange pipeline repair method is applied to double-layer intelligent heat exchange tube structures. Because there are pre-made threads and friction-reducing lubricating oil between the double-layer intelligent heat exchange tubes, the damaged section of the pipe can be unscrewed along the threads to replace it, or an automatic repair agent can be used to repair the damaged section.

[0090] The automatic repair agent is one or more of polyethylene powder, polyethylene granules, nylon fiber filaments, ethylene glycol, and adhesive; when the automatic repair agent is nylon fiber filaments, the fineness is 200-300D, and the length of the nylon fiber filaments is 3-5mm. This ensures that the mixture does not clump inside the pipe, thus reducing the effectiveness of the repair agent. The nylon fiber filaments are intertwined with the polyethylene powder, rubber powder, and granules, making it difficult for the polyethylene powder to fall off, and thus participating in filling the holes in the heat exchange pipeline.

[0091] Because the pipe wall has a double-layer structure, with fiber filaments running through both sides of the pores, the polyethylene microparticles are firmly fixed. The repair process and method of the automatic repair agent are as follows:

[0092] P1: After locating the damage point, seal either end of the pipe inlet / outlet according to the pressure test standard, and inject an automatic repair agent into the pipe through the other pipe opening;

[0093] P2: After injecting the repair agent, pressurize the pipe, but the pressure should not exceed the rated pressure that the polyethylene pipe used can withstand.

[0094] P3: When the pressure gauge reading is stable, the initial repair of the damaged point is completed; after the automatic repair agent has fully reached the strength requirements, the pipeline is pressure tested. Under the test pressure, the pressure should be stabilized for 15 to 20 minutes. After stabilization, the pressure drop should be 1% to 3%, and there should be no leakage to be considered as the repair work is completed.

[0095] The ultrasonic phased array method described in step S3 above involves controlling the time delay of the excitation and reception pulses of each array element in the array transducer to change the phase relationship when a certain array element inside the object receives sound waves, thereby changing the focal point and the direction of the sound beam to synthesize a phased array beam and thus scan for information.

[0096] The ultrasonic guided wave testing described in step S3 above involves placing several sensor probes at multiple locations within the pipeline to form a sensor array that detects the condition of the pipeline wall. As the information sensed by the sensors passes through the guided wave and is reflected, the damage to the pipeline wall is detected. Subsequent analysis and calculation of the relevant data then pinpoint the problem within the pipeline wall. To generate an appropriate waveform for typical pipe wall thicknesses, a much lower frequency than conventional ultrasonic testing is required. Guided waves typically operate at frequencies of 60–100 kHz. Therefore, the detection sensitivity of guided waves for individual defects is relatively low compared to ultrasonic testing, which typically uses frequencies in the MHz range. However, guided wave testing has the advantage of propagating over long distances of 20–30 meters with minimal attenuation. Thus, a large-scale inspection can be performed by fixing a pulse-echo array at a single location, making it particularly suitable for detecting internal and external corrosion of in-service pipelines and dangerous defects in welds. Low-frequency guided wave distance ultrasonic testing is used for rapid inspection of pipelines in service; internal and external corrosion can be detected in a single operation, and planar defects in the pipeline cross-section can also be detected.

[0097] The ultrasonic C-scan detection described in step S3 above is a deep scan of the pipeline. Its basic working principle is that ultrasonic waves are generated inside the component being inspected through a reflective probe. The probe then transmits the received information about the defect location back as ultrasonic waves, and the software displays various indicators of the reflected waves, thereby determining the specific information about the location and size of the pipeline defect.

[0098] The beneficial effects of this invention: Compared with existing control technologies, this invention has the following technical advantages:

[0099] (1) There are many design and construction methods for energy pile buried pipes, but there is currently no method for the management and operation of energy pile buried pipes. This invention proposes a support vector regression advance control method suitable for energy pile buried pipes, which has high computational efficiency and can improve service life.

[0100] (2) Traditional ground source heat pump control methods mostly rely on communication signals to control the heat pump unit in real time, which cannot meet the advanced dynamic control requirements of energy pile green and low-carbon buildings. The intelligent operation and maintenance control method proposed in this invention can learn the control model from the collected historical data and predict the control decision scheme for the next moment.

[0101] (3) The intelligent operation and maintenance control method for green and low-carbon buildings based on federated learning control management proposed in this invention realizes the integrated control of three parts: green and low-carbon building user end, unit selection, and energy pile buried pipe end, which improves the scientific rationality of the control scheme and reduces the operation and management cost.

[0102] (4) The pipeline maintenance device proposed in this invention, which combines a double-layer intelligent heat exchange tube structure with a micro magnetoelectric sensor, has simple operation steps, strong operability, and is easy to control and implement. The double-layer intelligent heat exchange tube structure with embedded metal ring structure is simple and easy to prefabricate; the micro magnetoelectric sensor has the functions of monitoring and unblocking pipelines, and has multiple functions.

[0103] (5) The present invention proposes a complete set of methods for detecting, clearing and repairing the blockage and damage of the embedded pipes in the energy pile structure during the operation and maintenance process. During the detection process, if the pipe is found to be blocked, the monitoring module and the clearing module are immediately activated to clear the pipe. If the pipe is damaged, the inner heat exchange pipe can be unscrewed and replaced immediately, thereby improving the efficiency of finding and solving problems.

[0104] (6) The automatic repair agent is readily available and economical. The repair agent is easy to use and has a good repair effect, which strengthens the structure of the heat exchange tube. Attached Figure Description

[0105] Figure 1 This is a schematic diagram of an optional federated learning method for the operation control of green and low-carbon buildings with energy piles in an embodiment of the present invention;

[0106] Figure 2 This is a schematic diagram of an optional federated learning algorithm framework for the operation control of green and low-carbon buildings with energy piles in an embodiment of the present invention.

[0107] Figure 3 This is a schematic diagram of an optional energy pile buried pipe end control support vector regression local model structure in an embodiment of the present invention;

[0108] Figure 4 This is a schematic diagram of an optional energy pile buried pipe end control support vector regression local model flow in an embodiment of the present invention;

[0109] Figure 5 This is a schematic diagram of an optional simulated annealing improved support vector regression algorithm in an embodiment of the present invention;

[0110] Figure 6 This is a schematic diagram of a local model structure of a decision tree for optional unit equipment operation control parameters in an embodiment of the present invention;

[0111] Figure 7 This is a schematic diagram of a partial decision tree model for optional unit equipment operation control parameters in an embodiment of the present invention;

[0112] Figure 8 This is a schematic diagram of an optional indoor user-end control random forest local model in an embodiment of the present invention;

[0113] Figure 9 This is a schematic diagram of an optional indoor user-side control random forest local model in an embodiment of the present invention.

[0114] Figure 10 This is a schematic diagram of an optional heat exchanger tube maintenance device based on the principle of electromagnetic induction in an embodiment of the present invention.

[0115] Figure 11 This is a schematic diagram of an optional double-layer intelligent heat exchange tube structure in an embodiment of the present invention;

[0116] Figure 12 This is a schematic diagram of the cross-sectional structure of an optional double-layer intelligent heat exchange tube in an embodiment of the present invention;

[0117] Figure 13 This is a schematic diagram of an optional micro magnetoelectric sensor structure containing a dredging module in an embodiment of the present invention;

[0118] Figure 14 This is a schematic diagram of an optional micro magnetoelectric sensor structure containing a dredging module in an embodiment of the present invention. (AA is a cross-sectional schematic diagram.)

[0119] Figure 15 This is a schematic diagram of an optional isolated forest algorithm for anomaly signal detection in an embodiment of the present invention.

[0120] In the diagram: 1-Double-layer intelligent heat exchange tube structure, 1-A-Inner polyethylene tube, 1-B-Outer polyethylene tube, 2-Magnetic induction ring, 3-Miniature magnetoelectric sensor with unblocking module, 3-1-Helical metal wire, 3-2-Connecting component, 3-3-Gear, 3-4-Surround camera, 3-5-Series motor, 3-6-Magnetic sensor, 3-7-Drive wheel, 3-8-Cable, 3-9-Lighting lamp, 4-Cable storage tray, 5-Automatic data acquisition device, 6-Control terminal. Detailed Implementation

[0121] The specific embodiments of this invention are described in detail below with reference to the accompanying drawings. However, the scope of protection of this invention is not limited to the description of these embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0122] Figure 1 This invention provides an optional federated learning algorithm structure diagram for the operation control of green and low-carbon buildings using energy piles. It mainly consists of four parts: a field equipment terminal, a data acquisition terminal, a server terminal, and a control signal terminal. The algorithm is based on a client-server architecture and establishes a framework for this federated learning algorithm. Figure 2 This is a framework diagram of the federated learning algorithm for the operation and control of energy piles in green and low-carbon buildings. The learning and computation of all sub-modules are implemented on the client side, including three main client modules: the energy pile buried pipe control module, the unit equipment control module, and the indoor user terminal control module. Finally, the anonymized parameters calculated by each client are aggregated to the central server for calculation, and then distributed to each data holder (client) to update their local model until the global model is robust. The specific steps are as follows:

[0123] Step D1: Collect and organize the indoor user terminal operation status parameters and control parameters, unit operation status parameters and control parameters, and energy pile buried pipe terminal operation status parameters and control parameters of a green low-carbon building in Dalian City using on-site survey technology, questionnaire survey technology, and image intelligent recognition technology. Perform simple data preprocessing, analyze the characteristics of sample data, and allocate them to the corresponding computing module client. The indoor user terminal operating status parameters include: indoor ambient temperature, outdoor ambient temperature, personnel behavior status, equipment operating status, indoor user terminal manifold water level, differential pressure, manifold flow rate, manifold pressure, manifold temperature, indoor supply water temperature, indoor return water temperature, indoor supply water flow rate, indoor return water flow rate, indoor supply water pressure, and indoor return water pressure; the unit operating status parameters include: condenser working liquid level, condenser inlet temperature, evaporator working liquid level, evaporator inlet temperature, and compressor pressure; the energy pile buried pipe terminal operating status parameters include: buried pipe terminal manifold water level, differential pressure, manifold flow rate, manifold pressure, manifold temperature, buried pipe supply water temperature, buried pipe return water temperature, buried pipe supply water flow rate, buried pipe return water flow rate, buried pipe supply water pressure, and buried pipe return water pressure, as shown in Table 1.

[0124] Table 1. Training sample space and sample quality of the operation control model for a bridge de-icing system using energy piles in Jiangyin City. Step D2: The three clients—the energy pile buried pipe end control module, the unit equipment control module, and the user terminal control module—update their respective local models. The local models of the three clients are as follows:

[0125] 1. Energy Pile Buried Pipe Control Module: Based on the indoor user terminal operating status parameters, unit operating status parameters, and energy pile buried pipe terminal operating status parameters, a local support vector regression model for energy pile buried pipe terminal control is established. This model predicts and calculates the control scheme, which is then encrypted and uploaded to the central processing unit. Figure 3 This is a local model structure diagram of support vector regression for the control of buried pipe end of an energy pile, provided in an embodiment of the present invention. Figure 4 The flowchart for the local support vector regression model for the control of the buried pipe end of the energy pile is shown below. The specific steps of the model in the s-th round are as follows:

[0126] Step D2.1.1: Decrypt the public key of the encrypted data using the private key, and update the iteration results of the relevant clients of the energy pile buried pipe end control module, unit equipment control module, and indoor user control module according to the global shared model of the central server as the new sample data attribute parameters. Use the opening parameter of the energy pile buried pipe end control valve as the learning target.

[0127] Step D2.1.2: To resolve the relationship between the sample parameters of the operating status of the energy pile buried pipe end and the control parameters of the energy pile buried pipe end, a slack variable ξ is introduced. i , A nonlinear segmentation support vector classifier considering soft margins is constructed using the penalty coefficient C. The model's prediction accuracy and inherent stability can be represented by the loss function (L = max(0, |z|-∈)); the loss function is then converted into a conditional extremum function.

[0128] Step D2.1.3: The conditional extremum function needs to be transformed into a multivariate function for solution using the Lagrange function. By setting the partial derivatives of the Lagrange function with respect to the optimization objectives w, b, ξ to 0, the Lagrange multipliers are obtained, which can transform the original conditional extremum function into the dual function, thereby finding the minimum value of the prediction boundary.

[0129] Step D2.1.4: The nonlinear mapping function Φ(X) contained in the classification hyperplane of the local support vector regression model of the energy pile buried pipe control end, its inner product φ(X) i ) T φ(X j The processing of ) can be done using a Gaussian kernel:

[0130] Step D2.1.5: Optimize the support vector regression model parameters based on the training set data: insensitive loss function ∈, penalty coefficient C, and hyperparameters γ, λ, α, c, and d in the kernel function.

[0131] Preferably, a segmented simulated annealing method is used to optimize the model parameters. Figure 5 The flowchart of a local model of an alternative support vector regression algorithm improved by simulated annealing is as follows:

[0132] Step D2.1.5.1: Randomly generate an initial parameter set for interactive verification. Record the error value EEP as the current annealing system state E0, the initial temperature T0, the temperature of annealing stage one T1, and the annealing end temperature as T2.

[0133] Step D2.1.5.2: According to the perturbation algorithm m′ i =m i +s*(μ-0.5)(B i -A i The parameters are perturbed to form a new parameter set, and the current annealing system state E is obtained through interactive verification. n Calculate ΔE = E n -E n-1 .

[0134] Step D2.1.5.3: If ΔE < 0, accept the new parameter set and jump to step D2.1.5.5; otherwise, accept the corresponding parameter set according to the Metropolis criterion exp(ΔE / kT)-μ > 0 and jump to step D2.1.5.5; if none of the above conditions are met, reject the critical state and return to step D2.1.5.2, re-perturb to generate a new parameter set, and perform interactive verification until the parameter set acceptance condition in step D2.1.5.3 is met.

[0135] Step D2.1.5.4: Upon obtaining the new state, proceed with the cooling plan. Cool down the temperature. If the set temperature T1 is not reached, return to step D2.1.5.2. When the set temperature T1 is reached, start a new annealing plan.

[0136] Step D2.1.5.5: According to the new perturbation method and annealing plan, continue to perturb the parameter set of the first-stage annealing and calculate the corresponding state parameters E. n .

[0137] Step D2.1.5.6: If ΔE < 0, accept the new parameter set and jump to step D2.1.5.7; otherwise, accept the corresponding parameter set according to the Metropolis criterion and jump to step D2.1.5.7; if none of the above conditions are met, reject the parameter set and return to step D2.1.5.5, re-perturb to generate a new parameter set, and perform interactive verification until the parameter set acceptance condition in step D2.1.5.7 is met.

[0138] Step D2.1.5.7 sets the end temperature to T2 as the algorithm exit, and sets the global maximum number of EEP calculations to N. Stop annealing when T2 or N is reached. At this time, the cross - validation error E of the accepted critical state n should be the lowest E n , and the corresponding parameters should be the best prediction parameters. Otherwise, return to step D2.1.5.5.

[0139] Step D2.1.6: Input the operating state parameters of the buried pipe end of the target energy pile, and use the trained support vector regression model to predict and calculate the control parameters of the buried pipe of the target building.

[0140] Step D2.1.7: Compare the calculation result with the device control threshold. If the calculation result meets the control threshold, convert the predicted calculation result into encrypted parameters using the public key and upload them to the central server for the global shared model iteration of federated learning. If the calculation result exceeds the control threshold, execute step D2.1.8.

[0141] Step D2.1.8: Output the calculated control parameters of the buried pipe end of the energy pile and issue an alarm. Use the device control threshold as the advanced prediction control information, convert it into encrypted parameters using the public key, and upload them to the central server for the global shared model iteration of federated learning. At the same time, arrange a heat exchange pipe maintenance device based on the principle of electromagnetic induction for refined detection and positioning.

[0142] 2. Unit equipment end control module: Figure 6 It is a structure diagram of an ID3 decision tree local model for an optional control parameter of unit equipment operation. Based on the operating state parameters of the indoor user end, the operating state parameters of the unit, and the operating state parameters of the buried pipe end of the energy pile as sample attribute data, establish an ID3 decision tree local model for the control parameters of unit equipment operation, determine the control parameters of the heat pump and the circulation pump and the valve opening, and encrypt and upload them to the central processor. The learning sample attributes for the ID3 decision tree local model for unit equipment operation control include the working liquid level of the condenser, the inlet temperature of the condenser, the working liquid level of the evaporator, the inlet temperature of the evaporator, the compressor pressure, the flow rate of the cooling pump, and the flow rate of the chilled water pump. In addition, generate a public key for encrypting and transmitting data according to the global shared model of the central server, and update the iteration results of the relevant clients at the indoor user end and the buried pipe end of the energy pile as new sample data attribute parameters. Figure 7 It is a flowchart of an ID3 decision tree for the control parameters of unit equipment operation in an embodiment. The specific steps of the s - th round of this client are as follows;

[0143] Step D2.2.1: Decrypt the public key for encrypting the transmitted data using the private key, and update the iteration results of the relevant clients of the energy pile buried pipe end control module, unit equipment control module, and indoor user control module according to the global sharing model of the central server as the new sample data attribute parameters. Calculate the information gain of all attributes (various unit parameters) of the training samples according to the information entropy, and sort all attributes according to the information gain.

[0144] Step D2.2.2: Select the attribute with the largest information gain as the optimal attribute, and perform sample partitioning based on the optimal attribute. Samples with the same value of the optimal attribute are used as the same sample set to form the root node.

[0145] Step D2.2.3: Then regard each root node as a complete data set, perform sample partitioning based on the sub-optimal attribute, use samples with the same value of the sub-optimal attribute as the same sample set to form leaf nodes, and perform decision tree pruning by the REP method, and form a decision tree through iterative steps in sequence.

[0146] Step D2.2.4: Compare the calculated unit equipment control parameters with the equipment control threshold. If the calculation result meets the equipment control threshold, convert the decision calculation result into encrypted parameters using the public key as the advanced prediction control information and upload it to the central server for the iteration of the global sharing model of federated learning. If the calculation result exceeds the control threshold, execute Step D2.2.5.

[0147] Step D2.2.5: Output the decision-making unit equipment control result and issue an alarm, convert the equipment control threshold into encrypted parameters using the public key as the advanced prediction control information and upload it to the central server for the iteration of the global sharing model of federated learning.

[0148] 3. Indoor user end control module: Figure 8 This is a structural diagram of a random forest local model for indoor user end control provided by an embodiment of the present invention. A random forest local model for indoor user end control parameters is established based on samples of indoor user end operating state parameters, unit operating state parameters, and energy pile buried pipe end operating state parameters, and the calculated indoor user end control parameters are used as an advanced control scheme and encrypted and uploaded to the central processor. The indoor user end control state parameters include: set temperature of the fan coil unit, opening degree of the indoor return water control valve, opening degree of the indoor supply water control valve, opening degree of the collector header control valve at the indoor user end, opening degree of the collector branch control valve, opening degree of the header control valve at the indoor user end of the water distributor, and opening degree of the branch control valve of the water distributor. Taking the ID3 algorithm as an embodiment of the present invention, Figure 9 This is a flowchart of a random forest local model for indoor user end control in an embodiment. The specific steps of the s-th round of this client are as follows:

[0149] Step D2.3.1: Decrypt the public key of the encrypted data using the private key. Update the iteration results of the relevant clients of the energy pile buried pipe control module, unit equipment control module, and indoor user control module according to the global shared model of the central server, and use them as the new sample data attribute parameters. Use the Bootstrap sampling method to randomly generate a subset of samples from the samples, which will be used as training samples for one of the decision tree models. Repeat the sampling k times to form k decision tree training samples.

[0150] Step D2.3.2: Train decision trees based on the attribute subsets from the k training samples to form k independent random decision trees.

[0151] Step D2.3.3: The random forest randomly selects a classifier to vote on the indoor user terminal control schemes predicted by the k decision trees, and the voting result is taken as the optimal control scheme for the indoor user terminal.

[0152] Step D2.3.4: Compare the determined indoor user terminal control parameters with the equipment control thresholds. If the calculation result meets the equipment control threshold, the decision calculation result is used as advanced predictive control information, converted into encrypted parameters using a public key, and uploaded to the central server for iteration of the globally shared model in federated learning. If the calculation result exceeds the control threshold, proceed to step D2.3.5.

[0153] Step D2.3.5: Output the determined indoor user-end control results and issue an alarm. Use the equipment control threshold as the advanced predictive control information, convert it into encrypted parameters using a public key, and upload it to the central server for iteration of the globally shared model in federated learning. Simultaneously, arrange for a heat exchanger tube maintenance device based on the principle of electromagnetic induction to perform precise detection and location.

[0154] Step D3: The three clients—the energy pile buried pipe end control module, the unit equipment control module, and the indoor user control module—upload encrypted and de-identified parameters to the central server in the form of public keys.

[0155] Step D4: The central server uses its private key to decrypt the encrypted and anonymized parameters uploaded by the three clients, aggregates the decrypted parameters of all client models based on knowledge distillation using federated learning, and updates the global model weights of the M clients according to the following formula: Among them W s+1 For the global model parameters in round s, The client sub-model weights are uploaded to the server by client i in round s. After each round of model weight updates, the central server calculates the global model error and accuracy. The central server can also control the transmission speed and the shutdown of model training.

[0156] Step D5: The central server generates a public key for encrypting transmitted data using the globally shared model and distributes it to each client. Based on the globally shared model, each local client updates the iteration results of other relevant clients as the new sample data attribute parameters. For example, the support vector regression model client for the energy pile buried pipe control module updates the unit equipment control parameters and indoor user terminal equipment control parameters based on the globally shared model.

[0157] Step D6: Repeat steps D2 to D5 iteratively until the globally shared model is robust. Finally, the client calculates the corresponding result based on the global model. The matching control states in a green low-carbon building with energy piles in Dalian City are calculated, including the set temperature of the fan coil units, the opening degree of the indoor return water control valve, the opening degree of the indoor supply water control valve, the opening degree of the indoor user-end water collector manifold control valve, the opening degree of the water collector branch pipe control valve, the opening degree of the indoor user-end water distributor manifold control valve, the opening degree of the water distributor branch pipe control valve, the opening degree of the makeup water pump control valve, the opening degree of the cooling circulation pump control valve, the opening degree of the chilled water circulation pump control valve, the opening degree of the expansion valve, the opening degree of the buried pipe end water collector manifold control valve, the opening degree of the water collector branch pipe control valve, the opening degree of the buried pipe end water distributor manifold control valve, the opening degree of the water distributor branch pipe control valve, the opening degree of the energy pile return water control valve, and the opening degree of the energy pile supply water control valve, as shown in Table 2.

[0158] Table 2. Control parameter results of a bridge de-icing system using energy-powered piles in Jiangyin City.

[0159]

[0160]

[0161] This invention provides a complete set of devices and methods for detecting, clearing, and repairing blockages and damage to embedded pipes within energy pile structures. Specific implementation methods are described below.

[0162] like Figure 10 The diagram shows the schematic of a heat exchanger tube maintenance device based on the principle of electromagnetic induction. This device mainly includes a double-layer intelligent heat exchanger tube structure 1, a miniature magnetoelectric sensor 3 with a clearing module, a cable tray 4, an automatic data acquisition unit 5, and a control terminal 6. The detection principle is as follows: When an electrical signal is applied to the magnetoelectric sensor, the electromagnetic frequency changes as it passes through the magnetic induction ring 2. The electrical signal is transmitted to the automatic data acquisition unit 5 via a connected cable. The automatic data acquisition unit 5 automatically records the time the sensor passes through each magnetic induction ring sequentially and transmits this information to the control terminal 6. If no electrical signal is transmitted to the next measuring point for an extended period after passing a certain measuring point, it is determined that there may be a blockage between the two measuring points.

[0163] like Figure 11 and Figure 12As shown, the double-layer intelligent heat exchange tube structure 1 includes an inner polyethylene tube 1-A, an outer polyethylene tube 1-B, and a magnetic induction ring 2. The magnetic induction ring 2 is heat-fused and embedded in the wall of the outer polyethylene tube 1-B, with a ring width of 5-8 mm. It is integrally formed with the outer polyethylene tube 1-B and is arranged at equal intervals along the axial direction of the double-layer intelligent heat exchange tube structure 1. The spacing is determined according to the design requirements of the energy pile and is between 0.3 and 0.5 m. In this embodiment, it is 0.5 m. The inner wall of the outer polyethylene tube 1-B and the outer wall of the inner polyethylene tube 1-A are threaded. Lubricating oil is applied between the two polyethylene tubes, and the two polyethylene tubes are tightened using the threads. Figure 12 for Figure 11 A schematic diagram of the cross-section of a double-layer intelligent heat exchanger tube structure.

[0164] The structural diagram of the miniature magnetoelectric sensor 3 containing the unblocking module is shown below. Figure 13 As shown, Figure 14 for Figure 13 The cross-sectional view of AA includes a dredging module, a monitoring module, a magnetoelectric sensor 3-6, and a moving module. The dredging module includes a spiral metal wire 3-1 and / or a thermoplastic rod, a connecting component 3-2, a gear 3-3, and a series-wound motor 3-5; the monitoring module includes a surround camera 3-4 and a lighting lamp 3-9; and the moving module includes a drive wheel 3-7 and a corresponding motor.

[0165] This invention proposes a method for detecting blockages and damage in heat exchanger pipelines using a heat exchanger tube inspection device based on the principle of electromagnetic induction. The detection method steps are as follows:

[0166] Step S1: After the Energy Pile Green Building Federation Learning Operation Control issues a preliminary fault alarm, first turn on the power switch of the automatic data acquisition instrument 5, and use a magnetic induction ring to surround and move the magnetic electric sensor 3-6. When the magnetic induction ring encounters the sensing point of the magnetic electric sensor 3-6, an audible and visual alarm will be issued, and the instrument will show an indication, indicating that the acquisition instrument is working normally.

[0167] Step S2: During testing, the miniature magnetoelectric sensor 3 containing the unblocking module is inserted into the pipeline along the magnetoelectric sensor 3-6. The pipeline is vertically arranged within the energy pile. The magnetoelectric sensor 3-6 is slowly inserted using its own gravity and the rotating cable reel 4. When the center of the magnetoelectric sensor 3-6 intersects with the magnetic induction ring 2, the instrument emits a beep, accompanied by a light indicator. Simultaneously, the surround camera 3-4 and the lighting 3-9 are activated, transmitting data in real-time to the control terminal 6, ensuring the camera is aimed at the blockage. The automatic data acquisition instrument 5 automatically records the number of each magnetic induction ring measuring point, the time interval between adjacent measuring points, real-time image data, and sound waveform information.

[0168] Step S3: The automatic data acquisition device 5 transmits the automatically recorded data to the control terminal 6. The control terminal 6 can use the isolated forest algorithm to analyze the collected data and accurately locate the damage point. Based on the analyzed abnormal time and location of the data, the location of the blockage point can be accurately located.

[0169] The Isolation Forest algorithm is used to analyze real-time data. An isolation tree is built in the attribute space of the sample set to isolate attributes. Within the same attribute space, outliers are relatively distant from the attribute values ​​of most samples. Through model training, outliers are separated earlier, while normal values ​​are located in deeper child nodes of the isolation tree. The algorithm structure is as follows: Figure 15 The specific steps are as follows:

[0170] Step S3.1: Organize the current signal, sound wave signal, and electromagnetic signal transmitted from the unblocking module to the control center as sample attributes to form a sample set, which can be represented as X=(x1,x2,x3,…,x…) n ), where x n =(x n1 ,x n2 ,x n3 ,…,x nm ), x nm Let n be the attribute signal of the nth class at time m.

[0171] Step S3.2: Use the attribute dimension q as the attribute cut point to segment the attribute types, forming different isolated trees. For each isolated tree, use the attribute value cut point p as the outlier cut point to segment the attribute values ​​and divide the sample space. Q is a random value belonging to the range (0, n), and p is a random value belonging to the range min(x). n ) < p < max(x) n () random values ​​within the range.

[0172] Step S3.3: Based on the splitting of attribute cut points and outlier cut points, a hyperplane is formed in the sample space. Samples smaller than the outlier cut points are classified into one class, forming the left node of the isolated tree, and vice versa.

[0173] Step S3.4: Repeat steps S3.2 to S3.3 until each isolated tree leaf node contains only one sample, thus completing the training of the isolated tree model.

[0174] Step S3.5: Take the sample data x at each time step i By traversing and calculating the data in an isolated tree model, the average height h(x) of each sample data can be obtained. i ), and the anomaly score S(x,ψ) of the sample data. Determine the anomaly signal based on the table below.

[0175] Table 3

[0176] S(x, ψ) value Signal evaluation ＞0.75 Abnormal signal 0.4～0.75 Suspected abnormal signal ＜0.4 Normal signal

[0177] This invention proposes a method for unblocking and repairing heat exchanger pipelines using a heat exchanger tube maintenance device based on the principle of electromagnetic induction. The specific steps are as follows:

[0178] Step T1: Turn on the surround camera 3-4 and the lighting 3-9, and transmit the data to the control terminal 6 in real time, so that the camera is aimed at the blockage.

[0179] Step T2: Begin the unblocking work. If the blockage is found to be caused by pipe heat fusion, turn on the series motor 3-5 controlling the spiral metal wire 3-1. After the spiral metal wire 3-1 rotates to the working speed, continue to use the gravity of the magnetoelectric sensor 3-6 to slowly lower the sensor cable. The spiral metal wire 3-1 stirs the blockage until the blockage is cleared. If the blockage is caused by pipe heat fusion, the stainless steel metal wire 3-1 and connecting component 3-2 of the unblocking module can be replaced with a heat fusion rod, and the magnetoelectric sensor 3-6 can be placed at the blockage again. Turn on the heating rod to the heat fusion temperature of the polyethylene material, and use the gravity of the magnetoelectric sensor 3-6 to slowly lower the magnetoelectric sensor 3-6 cable until the blockage is cleared.

[0180] Step T3: After removing the miniature magnetoelectric sensor from the pipe, perform pipe pressure testing and flushing to clean any remaining foreign objects inside the pipe, thus completing the unblocking process.

[0181] Step T4: After accurately locating the damaged point of the intelligent heat exchange tube, since there are pre-made threads and lubricating oil to reduce friction between the double-layer intelligent heat exchange tube structure, the tube section where the damaged point is located can be unscrewed along the threads to replace the damaged tube section.

Claims

1. A method for the inspection and intelligent operation and maintenance control of heat exchanger tubes in a green and low-carbon building with energy piles, characterized in that, The heat exchanger tube maintenance and intelligent operation and maintenance control method of this energy pile green low-carbon building includes a control operation method and a heat exchanger tube maintenance device based on the principle of electromagnetic induction; a preliminary judgment is made and a preliminary alarm location is made based on the control operation method, and on this basis, the heat exchanger tube maintenance device based on the principle of electromagnetic induction is arranged to carry out refined repair work. The control operation method establishes a federated learning algorithm framework for the operation control of green low-carbon buildings with energy piles based on a client-server architecture; the learning and calculation of all sub-modules are implemented on the client, including three major modules: the energy pile buried pipe end control module, the unit equipment control module, and the indoor user end control module; the desensitized parameters calculated by the client are aggregated to the central server for calculation, and then distributed to each client to update its local model until the globally shared model is robust. The specific steps are as follows: D1: Acquire indoor user terminal operating status parameters, indoor user terminal control parameters, unit operating status parameters, unit control parameters, energy pile buried pipe terminal operating status parameters, and energy pile buried pipe terminal control parameters; preprocess the data of each parameter, analyze the data characteristics, and allocate them to the corresponding clients; D2: Each client updates its local model with the following parameters: the local support vector regression model for energy pile buried pipe control, the local decision tree model for unit equipment control, and the local random forest model for indoor user terminal control. D3: The three clients—the energy pile buried pipe end control module, the unit equipment control module, and the indoor user control module—upload encrypted and desensitized parameters to the central server in the form of a public key. D4: The central server uses its private key to decrypt the encrypted and anonymized parameters uploaded by the three clients, performs secure aggregation, and then updates the global shared model; the central server decodes the encrypted and anonymized parameters; the federated learning on the central server side obtains the global shared model using one or more of the following synchronization methods: gradient averaging, federated averaging, and knowledge distillation. D5: The central server generates a public key for encrypting data transmission using the globally shared model and distributes it to each client; it updates the iteration results of other clients as new sample data attribute parameters based on the globally shared model and each client's local model. D6: Repeat steps D2 to D5 iteratively until the global shared model is robust. Finally, the client calculates the corresponding result based on the global shared model. The control states that match each other in the green and low-carbon building of energy piles are calculated, including the control parameters of the buried pipe end of the energy pile, the control parameters of the unit equipment, and the control parameters of the indoor user end.

2. The method for overhauling and intelligently controlling the heat exchanger tubes of energy pile green low-carbon buildings according to claim 1, characterized in that, The energy pile buried pipe end control module uses indoor user terminal operating status parameters, unit operating status parameters, and energy pile buried pipe end operating status parameters as sample attribute data to establish a support vector regression local model for energy pile buried pipe end control, predict and calculate the control scheme, and encrypt and upload it to the central processing unit; the specific steps are as follows: D2.1.1: Use the private key to decrypt the public key of the encrypted data transmission, update the iteration results of the energy pile buried pipe end control module, unit equipment control module, and indoor user control module client according to the global shared model of the central server as the new sample data attribute parameters, and use the energy pile buried pipe end control parameters as the learning target. D2.1.2: To resolve the relationship between the operating status parameters and control parameters of the energy pile buried pipe end, a slack variable is introduced. , and penalty coefficient C A nonlinear segmentation support vector classifier considering soft margins is constructed; the prediction accuracy and stability of the local support vector regression model for the control parameters of the buried pipe end of the energy pile are represented by a loss function; the loss function is converted into a conditional extremum function. in, w Let be the normal vector of the hyperplane. C The penalty coefficient, ξ、 As a relaxation factor, The hyperparameter that determines the boundary width; These are the actual test results for the training samples; The classification hyperplane for the support vector regression local model of energy pile buried pipe end control; i Number the training samples; N The number of samples in the training set; It is a nonlinear mapping function; D2.1.3: The above conditional extremum function is transformed into a multivariate function for solution using the Lagrangian function, allowing the Lagrangian function to be applied to the optimization objective. w , b , ξ The partial derivatives are 0, so we obtain the Lagrange multipliers, which transform the conditional extremum function into the dual function, thereby finding the minimum value of the prediction boundary. D2.1.4: Nonlinear mapping function contained in the classification hyperplane of the local support vector regression model of the control end of the energy pile buried pipe. Its inner volume The combined kernel function is used for processing: the combined kernel function is one or more of the following: Gaussian kernel, linear kernel, polynomial kernel, and Sigmoid kernel; D2.1.5: Optimizing model parameters in the local support vector regression model for energy pile buried pipe end control: insensitive loss function 、 Penalty coefficient C Hyperparameters in kernel functions γ , λ, α, c, d Choose one of the following methods: simulated annealing, grid search, particle swarm optimization, PSO algorithm, or genetic algorithm. D2.1.6: Input the operating status parameters of the target energy pile buried pipe end, and use the trained support vector regression model to predict and calculate the control parameters of the target building's energy pile buried pipe end; D2.1.7 Compare the calculation result of step D2.1.6 with the device control threshold. If the calculation result meets the control threshold, the predicted calculation result is used as advance prediction control information and converted into encrypted parameters using the public key and uploaded to the central server for iteration of the globally shared model of federated learning. If the calculation result exceeds the control threshold, then step D2.1.8 is executed. D2.1.8: Output the calculated control parameters of the buried pipe end of the energy pile and issue an alarm. Use the equipment control threshold as the advanced prediction control information and convert it into encrypted parameters using the public key to upload to the central server for the global shared model iteration of federated learning. At the same time, arrange the heat exchange tube maintenance device based on the principle of electromagnetic induction to carry out fine detection and positioning. The unit equipment control module uses indoor user terminal operating status parameters, unit operating status parameters, and energy pile buried pipe terminal operating status parameters as sample attribute data to establish a decision tree local model of unit equipment operating control parameters. This model determines the control parameters and valve openings of the heat pump and circulating pump, and then encrypts and uploads them to the central server. The decision tree model is generated based on one or more of the ID3, C4.5, and CART algorithms. The specific steps are as follows: D2.2.1: Decrypt the public key of the encrypted data using the private key, update the iteration results of the energy pile buried pipe terminal control module, unit equipment control module, and indoor user control module client according to the global shared model of the central server as the new sample data attribute parameters; calculate the information gain of the unit operation status parameters of the training samples according to the information entropy, and sort all unit operation status parameters according to the information gain. D2.2.2: Select the unit operating state parameter with the largest information gain as the optimal attribute, divide the samples based on the optimal attribute, and form the root node by taking the samples with the same optimal attribute value as the same sample set. D2.2.3: Treat each root node as a complete dataset, divide the samples according to the suboptimal attribute, and form leaf nodes by taking samples with the same suboptimal attribute value as the same sample set. Use the REP method, PEP method or MEP method to prune the decision tree, and iterate to form a decision tree. D2.2.4: Compare the determined unit equipment control parameters with the equipment control thresholds. If the calculation result meets the equipment control threshold, the decision calculation result is used as advanced predictive control information, which is converted into encrypted parameters using a public key and uploaded to the central server for iteration of the globally shared model of federated learning. If the calculation result exceeds the equipment control threshold, proceed to step D2.2.

5. D2.2.5: Output the unit equipment control results of the decision and issue an alarm. Use the equipment control threshold as the advanced predictive control information, convert it into encrypted parameters using the public key, and upload it to the central server for the global shared model iteration of federated learning. The indoor user terminal control module, based on the indoor user terminal operating status parameters and the unit operating status parameters, establishes an indoor user terminal control random forest local model on the client side, using the indoor user terminal operating status parameters as sample attributes. This model calculates the indoor equipment set temperature and chilled water control parameters as the optimal control scheme and uploads it to the central server in encrypted form. The specific steps are as follows: D2.3.1: Decrypt the public key of the encrypted data using the private key, and update the iteration results of the energy pile buried pipe terminal control module, unit equipment control module, and indoor user control module client according to the global shared model of the central server as the new sample data attribute parameters; use the Bootstrap sampling method to randomly generate a subset of samples from the indoor user terminal control training samples as the training samples of one of the decision tree models, and repeat the sampling k times to form k decision tree training samples. D2.3.2: Train decision trees based on a subset of attributes from k decision tree training samples to form k independent random decision trees; D2.3.3: The random forest randomly selects a classifier to vote on the indoor user terminal control schemes predicted by k decision trees, and the voting result is taken as the optimal control scheme for the indoor user terminal. D2.3.4: Compare the indoor user terminal control parameters determined in the optimal control scheme with the equipment control threshold. When the calculation result meets the equipment control threshold, the decision calculation result is used as advanced predictive control information, which is converted into encrypted parameters using a public key and uploaded to the central server for iteration of the globally shared model of federated learning. When the calculation result exceeds the equipment control threshold, proceed to step D2.3.

5. D2.3.5: Output the indoor user terminal control results and issue an alarm. Use the equipment control threshold as the advanced predictive control information, convert it into encrypted parameters using the public key, and upload it to the central server for global shared model iteration in federated learning. At the same time, arrange a heat exchange tube maintenance device based on the principle of electromagnetic induction to perform fine detection and positioning.

3. A heat exchanger tube maintenance device based on the principle of electromagnetic induction, characterized in that, The heat exchanger tube maintenance device based on the principle of electromagnetic induction includes a double-layer intelligent heat exchanger tube structure (1), a miniature magnetoelectric sensor (3) with a dredging module, a cable tray (4), an automatic data acquisition device (5), and a control terminal (6); the double-layer intelligent heat exchanger tube structure (1) is U-shaped, with its two ends connected to the automatic data acquisition device (5) via the cable tray (4), and the automatic data acquisition device (5) is connected to the control terminal (6); the double-layer intelligent heat exchanger tube structure (1) is an inner polyethylene tube (1-A). The outer polyethylene pipe (1-B) and the magnetic induction ring (2); the inner polyethylene pipe (1-A) is located inside the inner wall of the outer polyethylene pipe (1-B), and the two are filled with lubricating oil and threaded together; the magnetic induction ring (2) is embedded in the wall of the outer polyethylene pipe (1-B) and is arranged at equal intervals along the axial direction of the double-layer intelligent heat exchange pipe structure (1); the micro magnetoelectric sensor (3) containing the unblocking module moves inside the inner polyethylene pipe (1-A), and includes an unblocking component, a series excitation motor (3-5), and a drive wheel (3- 7) and cable (3-8); one end of the series motor (3-5) is connected to the cable (3-8), and the other end is fixed to the unblocking component through the mounting gear (3-3) and the connecting component (3-2); multiple drive wheels (3-7) are connected to the side of the series motor (3-5) and slide on the inner wall of the inner polyethylene pipe (1-A); the magnetoelectric sensor (3-6), the surround camera (3-4) and the lighting lamp (3-9) are all installed on the surface of the series motor (3-5); the magnetoelectric sensor (3-6) is powered. When the signal passes through the magnetic induction ring (2), the electromagnetic frequency changes and the electrical signal is transmitted to the data acquisition instrument (5) through the connected cable (3-8). The data acquisition instrument (5) automatically records the time when the magnetoelectric sensor (3-6) passes through each magnetic induction ring (2) in sequence and transmits this information to the control terminal (6). When there is no electrical signal transmission from the next magnetic induction ring (2) after a set time has elapsed after passing through a certain magnetic induction ring (2), it is determined that there is a blockage between the two magnetic induction rings (2).

4. The heat exchanger tube maintenance device based on the principle of electromagnetic induction according to claim 3, characterized in that, The magnetic induction ring (2) has a ring width of 5~8mm and is integrally formed with the outer polyethylene pipe (1-B); the spacing of the magnetic induction ring (2) is 0.3~0.5m.

5. The heat exchanger tube maintenance device based on the principle of electromagnetic induction according to claim 3 or 4, characterized in that, The unblocking component is a spiral metal wire (3-1) and / or a hot melt rod.

6. A heat exchanger tube inspection device based on the principle of electromagnetic induction for detecting blockages and damage in heat exchanger tubes, characterized in that, The steps include the following: S1: After the initial alarm of the fault is issued by the Energy Pile Green Building Federation Learning Operation Control, before the test begins, turn on the automatic data acquisition instrument (5) switch, use a magnetic induction ring to surround the magnetic electric sensor (3-6) and move it. When the magnetic induction ring encounters the sensing point of the magnetic electric sensor (3-6), it will issue an audible and visual alarm. At the same time, the automatic data acquisition instrument (5) will have an indication. The automatic data acquisition instrument (5) is working normally and the subsequent steps can be carried out. S2: During the test, the pipeline is arranged vertically in the energy pile. The miniature magnetoelectric sensor (3) with the unblocking module is placed in the pipeline. The magnetoelectric sensor (3-6) is inserted by its own gravity and the cable tray (4) is rotated. When the center of the magnetoelectric sensor (3-6) intersects with the magnetic induction ring (2), the data acquisition instrument (5) emits a beeping sound and is accompanied by light indication. The surround camera (3-4) is turned on to monitor the image and sound waveform in the pipeline in real time. The data acquisition instrument (5) automatically records the number of each magnetic induction ring (2) measurement point and the time interval between the two adjacent magnetic induction rings (2) measurement points, as well as the real-time image and sound waveform information. S3: The automatic data acquisition instrument (5) transmits the automatically recorded data to the control terminal (6) and uses one or more of the following methods to analyze the acquired data and accurately locate the damage point: isolated forest algorithm, X-ray digital imaging technology, ultrasonic detection, ultrasonic guided wave, ultrasonic C-scan, ultrasonic phased array, high temperature thickness measurement method; the control terminal (6) accurately locates the location of the blockage point based on the abnormal time and location of the analyzed data.

7. The method for detecting blockages and damages in heat exchanger pipelines using the heat exchanger tube inspection device based on the electromagnetic induction principle according to claim 6, characterized in that, In step S3, when analyzing the collected data using the isolated forest algorithm, an isolated tree is built in the attribute space of the sample set to isolate attributes. Within the same attribute space, outliers are separated through model training, while normal values ​​are located in deeper child nodes of the isolated tree. The specific steps are as follows: S3.1: The current signal, acoustic signal, and electromagnetic signal transmitted from the magnetoelectric sensor (3-6) to the control terminal (6) are used as sample attributes to form a sample set, represented as X=(x1, x2, x3,…,x n ), where x n =(x n1 , x n2 ,x n3 ,…,x nm ), x nm The nth type attribute signal at time m; S3.2: Use the attribute dimension q as the attribute cut point to segment attribute types, forming different isolated trees. For each isolated tree, use the attribute value cut point p as the outlier cut point to segment the attribute values ​​within the attribute space; Q is a random value belonging to the range (0, n), and p is a random value belonging to the range min(x... n ) < p < max(x) n Random values ​​within a given range; S3.3: Based on the split of attribute cut point q and outlier cut point p, a hyperplane is formed in the sample space. Samples smaller than the outlier cut point p are classified into one class, forming the left node of the isolated tree, and vice versa. S3.4: Repeat steps S3.2 to S3.3 until each isolated tree leaf node contains only one sample, thus completing the training of the isolated tree model; S3.5: Take the sample data x at each time step i By traversing and calculating the data in an isolated tree model, the average height h(x) of each sample data can be obtained. i The outlier score of the sample data is calculated using the following formula: Where h(x) is the height of x in each isolated tree, c(n) is the average path length for a given number of samples n, and E(h(x)) is the expected path length of x in multiple isolated trees; When the score S(x,ψ) is greater than 0.75, the signal corresponding to the sample at that moment is considered an abnormal signal; When the score S(x,ψ) is between 0.4 and 0.75, the signal corresponding to the sample at that time is considered a suspected abnormal signal; When the score S(x,ψ) is less than 0.4, the sample at that moment is considered a normal signal.

8. A heat exchanger tube maintenance device based on the principle of electromagnetic induction for unblocking and repairing heat exchanger pipelines, characterized in that, The steps include the following: T1: Turn on the surround camera (3-4) and the lighting (3-9) and transmit the data to the control terminal (6) in real time, so that the surround camera (3-4) is aimed at the blockage. T2: Begin the unblocking operation. If the blockage is found to be caused by pipe heat fusion, turn on the series motor (3-5) controlling the spiral metal wire (3-1). After the spiral metal wire (3-1) rotates to the working speed, continue to use the gravity of the magnetoelectric sensor (3-6) to slowly lower the sensor cable (3-8). The spiral metal wire (3-1) stirs the blockage until the blockage is cleared. If the blockage is caused by pipe heat fusion, replace the spiral metal wire (3-1) and connecting component (3-2) of the unblocking module with a heat fusion rod and place the magnetoelectric sensor (3-6) at the blockage again. Turn on the heating rod to the heat fusion temperature of the polyethylene material and use the gravity of the magnetoelectric sensor (3-6) to slowly lower the sensor cable (3-8) until the blockage is cleared. T3: After removing the miniature magnetoelectric sensor (3) containing the unblocking module from the pipe, perform pipe pressure testing and flushing to clean the foreign matter remaining in the pipe and complete the unblocking work; T4: After accurately locating the damage point, repair the heat exchange pipeline. The section of pipe where the damage point is located is unscrewed along the thread to replace the damaged section, or an automatic repair agent is used to repair the damaged section.

9. The method for unblocking and repairing heat exchanger pipelines using the heat exchanger tube maintenance device based on the electromagnetic induction principle according to claim 8, characterized in that, The automatic repair agent is one or more of polyethylene powder, polyethylene granules, nylon fiber filaments, ethylene glycol, and adhesive; when the automatic repair agent is nylon fiber filaments, the fineness is 200~300D and the length of the nylon fiber filaments is 3~5mm.

10. The method for unblocking and repairing heat exchanger pipelines using the heat exchanger tube maintenance device based on the electromagnetic induction principle according to claim 9, characterized in that, The repair process and method of the automatic repair agent are as follows: P1: After locating the damage point, seal either end of the pipe inlet / outlet according to the pressure test standard, and inject an automatic repair agent into the pipe through the other pipe opening; P2: After injecting the repair agent, pressurize the pipe, but the pressure should not exceed the rated pressure that the polyethylene pipe used can withstand. P3: When the pressure gauge reading is stable, the initial repair of the damaged point is completed; after the automatic repair agent has fully reached the strength requirements, the pipeline is pressure tested. Under the test pressure, the pressure should be stabilized for 15~20 minutes. After stabilization, the pressure drop should be 1%~3%, and there should be no leakage to be considered as the repair work is completed.

Images