CN-122022773-A - Distributed intelligent maintenance and fault early warning system for pole tower machine nest

CN122022773ACN 122022773 ACN122022773 ACN 122022773ACN-122022773-A

Abstract

The invention discloses a distributed intelligent maintenance and fault early warning system for pole and tower machine nests, which relates to the technical field of equipment operation and maintenance, and comprises a sensing and control terminal, a centralized control maintenance platform and a communication network, wherein the sensing and control terminal is distributed and deployed in each pole and tower machine nest, the centralized control maintenance platform is deployed in a regional center, the communication network realizes data interaction of the sensing and control terminal and the centralized control maintenance platform, the sensing and control terminal collects machine nest multidimensional state data, preprocessing and feature extraction are completed, the centralized control maintenance platform executes fault diagnosis based on rules and performance attenuation trend prediction based on machine learning through the intelligent diagnosis and prediction engine, and then a maintenance decision and resource scheduling module generates a maintenance decision and optimally schedules maintenance resources. The invention integrates front end sensing, intelligent analysis and back end resource scheduling depth to form a closed-loop operation and maintenance system, improves the operation and maintenance efficiency and reliability of the distributed tower crane nest, and ensures the continuity of the inspection service of the unmanned aerial vehicle of the power line.

Inventors

DENG XIN
LIU TIANCI
FU XIAOWEI
YU XIANG

Assignees

国网湖北省电力有限公司随州供电公司

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. The distributed intelligent maintenance and fault early warning system for the tower machine nest is characterized by comprising a sensing and control terminal, a centralized control maintenance platform and a communication network; the sensing and control terminals are distributed and deployed in each tower machine nest; the centralized control maintenance platform is deployed in the center of the area; The communication network is connected with the sensing and control terminal and the centralized control maintenance platform; The system comprises a sensing and control terminal, a centralized control maintenance platform, a maintenance decision and resource scheduling module, a control and maintenance management module and a control and maintenance management module, wherein the sensing and control terminal is used for collecting state data of a tower machine nest and carrying out preprocessing and feature extraction, the centralized control maintenance platform comprises an intelligent diagnosis and prediction engine and the maintenance decision and resource scheduling module, the intelligent diagnosis and prediction engine is used for carrying out fusion analysis on data uploaded by the sensing and control terminal so as to execute fault diagnosis based on rules and performance attenuation trend prediction based on machine learning, and the maintenance decision and resource scheduling module is used for generating maintenance decisions and optimizing scheduling maintenance resources according to diagnosis results and prediction trends output by the intelligent diagnosis and prediction engine.
2. The tower-machine-nest-oriented distributed intelligent maintenance and fault early warning system according to claim 1, wherein the sensing and control terminal comprises a multi-sensor array and an edge computing unit; The multi-sensor array is deployed inside the tower crane nest and is used for collecting environmental parameters, mechanical state parameters and electrical parameters, wherein the environmental parameters comprise temperature and humidity, the mechanical state parameters comprise vibration amplitude and cabin door switch state, and the electrical parameters comprise charging current, charging voltage and internal resistance of a standby battery; The edge computing unit is used for carrying out filtering and compression preprocessing on the original data acquired by the multi-sensor array, extracting key characteristic values from the preprocessed data, and carrying out matching analysis on the key characteristic values based on a preset local diagnosis rule base so as to realize quick fault diagnosis and local alarm.
3. The system for intelligent maintenance and fault pre-warning for a tower machine nest according to claim 2, wherein the multi-sensor array specifically comprises a temperature and humidity sensor, a water immersion sensor, a vibration sensor, a door magnetic switch, a current sensor, a voltage sensor and a battery health monitoring sensor; The vibration sensor is arranged at the lifting motor of the charging platform of the tower crane nest and is used for monitoring the running vibration signal of the motor; the battery health monitoring sensor is arranged in a standby battery of the tower crane nest and is used for monitoring the internal resistance, terminal voltage and temperature of the standby battery.
4. The tower-machine-nest-oriented distributed intelligent maintenance and fault early warning system according to claim 1, wherein the centralized control maintenance platform further comprises a data aggregation and storage module and a digital twin model library; the data aggregation and storage module is used for receiving and storing state data, characteristic data and alarm information uploaded by each sensing and control terminal; the digital twin model library establishes corresponding digital twin models for each type of tower machine nest, and the digital twin models can simulate physical characteristics and operation logics of the tower machine nest and perform data synchronization with the entity machine nest; The intelligent diagnosis and prediction engine comprises a rule-based diagnosis module and a machine learning-based prediction module; The rule-based diagnosis module utilizes a preset expert rule chain to carry out deep fault analysis on the data in the data aggregation and storage module; The prediction module based on machine learning adopts a long-short-term memory network model to analyze the historical time sequence data in the data aggregation and storage module so as to predict the residual service life of specific equipment; The updating process of the unit state of the long-term and short-term memory network model is described by the following mathematical formula: Wherein, the Indicating the current time Is used to determine the cell state vector of (a), Representing the last time Is used to determine the cell state vector of (a), Indicating that the forgetting door is at the moment For controlling the degree of retention of the state of the previous cell, Indicating the time of the input door For controlling the degree of update of the current candidate state, Is shown at the moment The generated candidate cell state vector is then used to determine, Element-by-element multiplication representing a vector; And the maintenance decision and resource scheduling module is in communication connection with the centralized control maintenance platform through a software data bus or a service bus.
5. The system for intelligent maintenance and fault pre-warning for a tower machine nest according to claim 4, wherein the maintenance decision and resource scheduling module is specifically configured to: Grading the event according to the fault type output by the rule-based diagnosis module and the equipment failure risk probability output by the machine learning-based prediction module; Automatically generating a maintenance work order containing event description, positioning information and required spare part information according to the event level; Establishing an optimization model taking the geographical position of a fault event, the real-time geographical position of an available maintainer, the degree of matching of personnel skills and warehouse spare part inventory as constraint conditions, and solving the optimization model to generate a dispatching scheme with minimum weighted sum of total response time and cost; and dispatching the maintenance work order to a maintenance personnel terminal appointed by the dispatching scheme.
6. The distributed intelligent maintenance and fault early warning method for the tower machine nest is applicable to the distributed intelligent maintenance and fault early warning system for the tower machine nest, and is characterized by comprising the following specific steps of: Acquiring multidimensional state data of the machine nest through a sensing and control terminal arranged in each tower machine nest, and performing edge side preprocessing and feature extraction on the multidimensional state data to obtain feature data; Uploading the characteristic data to a centralized control maintenance platform through a communication network; Performing fusion analysis on the uploaded characteristic data through an intelligent diagnosis and prediction engine of the centralized control maintenance platform, and executing fault diagnosis based on rules and performance attenuation trend prediction based on machine learning; and generating a maintenance decision and optimizing and dispatching maintenance resources according to the fault diagnosis result and the performance decay trend prediction result by the maintenance decision and resource dispatching module of the centralized control maintenance platform.
7. The method for intelligent maintenance and fault pre-warning for a tower machine nest according to claim 6, wherein the steps of performing edge-side preprocessing and feature extraction on the multidimensional state data specifically comprise: Performing moving average filtering on the original sensor data to eliminate high-frequency noise; data compression is carried out on the filtered data so as to reduce the transmission data quantity; extracting time domain statistical features and frequency domain spectrum features from the compressed data as the feature data; And matching the extracted characteristic data with rule conditions in a preset local diagnosis rule base in real time, and immediately triggering local audible and visual alarm if the matching is successful.
8. The method for distributed intelligent maintenance and fault pre-warning for a pole and tower machine nest of claim 6, wherein the step of performing machine learning based performance decay trend prediction comprises: Constructing a time sequence of equipment health indexes; inputting the time sequence into a pre-trained long-period and short-period memory network prediction model; The long-term and short-term memory network prediction model outputs a predicted value of the equipment health index in a specific future time window and the risk probability of equipment failure ; The risk probability Calculated by the following mathematical formula: Wherein, the Representing a predicted probability of risk of failure of the device, Representing the Sigmoid activation function, The weight vector is represented by a weight vector, A hidden state vector representing the output of the long-short term memory network model at the last time step, The term of the bias is indicated, The transpose operation of the vector is represented, Representing a vector dot product operation; The equipment health index is formed by weighting and fusing vibration characteristic amplitude, battery internal resistance increasing rate and temperature drift amount.
9. The method for distributed intelligent maintenance and fault pre-warning for a tower machine nest according to claim 6, wherein the steps of generating a maintenance decision and optimizing the scheduled maintenance resources specifically comprise: identifying an explicit fault of the fault diagnosis result and the risk probability Exceeding a preset threshold Combining the early warning events of the event to be processed into an event set to be processed; Based on event type and said risk probability Grading each event in the event set to be processed; automatically creating a corresponding maintenance work order for each of the rated events; And aiming at minimizing the total moving distance and time delay of all maintenance personnel to finish the distributed work orders, solving a multi-objective optimization problem under the constraint of meeting the requirement of matching the skills of the maintenance personnel with the event requirement and sufficient inventory of required spare parts, distributing optimal maintenance personnel for each maintenance work order and planning a routing inspection path.
10. The method for distributed intelligent maintenance and fault pre-warning for a tower machine nest according to claim 6, further comprising the step of remote repair: after the maintenance decision and resource scheduling module distributes a maintenance work order and before corresponding maintenance personnel arrive at the site of the target tower machine nest, the centralized control maintenance platform sends a remote control instruction to a sensing and control terminal of the target tower machine nest through the communication network; The remote control instructions are for performing at least one of: Restarting the designated functional module of the target tower machine nest, resetting the circuit breaker of the target tower machine nest, or switching the charging system of the target tower machine nest to a standby working mode.

Description

Distributed intelligent maintenance and fault early warning system for pole tower machine nest Technical Field The invention relates to the technical field of equipment operation and maintenance, in particular to a distributed intelligent maintenance and fault early warning system for a tower machine nest. Background Along with the large-scale application of unmanned aerial vehicle inspection technology in a power system, a pole tower machine nest for providing automatic take-off and landing, charging and storage for unmanned aerial vehicles is deployed along a power transmission and distribution line in a large quantity. The nests are in a field severe environment for a long time and are unattended, and the health state of the nests is directly related to the reliability and availability of the whole unmanned aerial vehicle inspection system. Therefore, the intelligent state monitoring and maintenance of the distributed pole tower machine nest group are realized, and the intelligent state monitoring and maintenance becomes a key technical requirement for guaranteeing the continuity of the power line inspection service. At present, the technical development of the related field mainly focuses on two aspects, namely monitoring of the state of a power tower body and predictive maintenance of industrial equipment based on the Internet of things. For example, some prior arts realize millimeter-level monitoring of tower deformation through a multivariate parameter data fusion technology, and predict the inclination trend of the iron tower by means of a multisource coupling deformation dynamic early warning technology, so as to realize the transition from passive rush repair to active defense. On the other hand, in the field of wider industrial Internet of things, some prior arts can realize early fault detection and diagnosis of complex equipment by deploying multiple types of sensors, adopting a high-precision information autonomous fusion technology and an early warning mechanism based on multilayer combination reasoning. These techniques represent an advanced level of current device state awareness and fault diagnosis. However, when the prior art is applied to the operation and maintenance scene of the wide area distributed tower unmanned aerial vehicle nest group, the prior art still has the defect that the prior scheme focuses on monitoring, diagnosis or single-point early warning of equipment states in a plurality of ways and lacks a closed-loop system capable of integrating front-end sensing, intelligent prediction and dynamic scheduling depth of rear-end maintenance resources. Specifically, although the prior art can find potential faults or performance attenuation trends of the machine nest, maintenance personnel, vehicles and spare part resources with dispersed geographic positions cannot be automatically and efficiently co-scheduled based on the predictive information so as to generate maintenance decisions with optimal cost and fastest response. The early warning information is disjointed from the maintenance action, the operation and maintenance mode still stays in the passive stage of monitoring, alarming and manual dispatching, and the severe requirements of the large-scale distributed infrastructure on operation and maintenance efficiency, cost and response speed are difficult to meet. In summary, the prior art has not effectively solved the technical problem of how to implement global automatic optimization of maintenance decision and resource scheduling based on health prediction of the distributed tower crane nest under complex geographic and resource constraints. Disclosure of Invention The invention aims to make up the defects of the prior art, provides a distributed intelligent maintenance and fault early warning system for a tower crane nest, and can realize real-time health monitoring, intelligent fault diagnosis and predictive maintenance decision of a wide-area tower crane nest group by constructing a cooperative operation and maintenance framework of distributed perception, cloud intelligent and resource scheduling closed loop, thereby effectively improving operation and maintenance efficiency, reducing maintenance cost and guaranteeing continuous and reliable operation of unmanned aerial vehicle inspection business. On one hand, the distributed intelligent maintenance and fault early warning system facing the tower machine nest comprises a sensing and control terminal, a centralized control maintenance platform and a communication network; the sensing and control terminals are distributed and deployed in each tower machine nest; the centralized control maintenance platform is deployed in the center of the area; The communication network is connected with the sensing and control terminal and the centralized control maintenance platform; The system comprises a sensing and control terminal, a centralized control maintenance platform, a maintenance decision and resource scheduling module,