CN117689373A - Maintenance decision support method for energy router of flexible direct-current traction power supply system - Google Patents
Maintenance decision support method for energy router of flexible direct-current traction power supply system Download PDFInfo
- Publication number
- CN117689373A CN117689373A CN202410153011.7A CN202410153011A CN117689373A CN 117689373 A CN117689373 A CN 117689373A CN 202410153011 A CN202410153011 A CN 202410153011A CN 117689373 A CN117689373 A CN 117689373A
- Authority
- CN
- China
- Prior art keywords
- maintenance
- data
- energy router
- real
- rule
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012423 maintenance Methods 0.000 title claims abstract description 150
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 34
- 238000007418 data mining Methods 0.000 claims abstract description 31
- 230000007613 environmental effect Effects 0.000 claims abstract description 20
- 238000004458 analytical method Methods 0.000 claims abstract description 19
- 238000005516 engineering process Methods 0.000 claims abstract description 19
- 230000000694 effects Effects 0.000 claims abstract description 16
- 238000012544 monitoring process Methods 0.000 claims abstract description 14
- 230000015556 catabolic process Effects 0.000 claims abstract description 11
- 238000006731 degradation reaction Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 10
- 230000002452 interceptive effect Effects 0.000 claims abstract description 7
- 238000001514 detection method Methods 0.000 claims abstract description 5
- 230000008569 process Effects 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 14
- 230000017525 heat dissipation Effects 0.000 claims description 9
- 230000002787 reinforcement Effects 0.000 claims description 9
- 230000003137 locomotive effect Effects 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000003066 decision tree Methods 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 230000003044 adaptive effect Effects 0.000 claims description 5
- 238000013461 design Methods 0.000 claims description 5
- 238000000513 principal component analysis Methods 0.000 claims description 5
- 238000012098 association analyses Methods 0.000 claims description 4
- 238000007621 cluster analysis Methods 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 238000005065 mining Methods 0.000 claims description 4
- 238000003909 pattern recognition Methods 0.000 claims description 4
- 238000000611 regression analysis Methods 0.000 claims description 4
- 238000013079 data visualisation Methods 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 238000012731 temporal analysis Methods 0.000 claims description 3
- 238000000700 time series analysis Methods 0.000 claims description 3
- 238000007726 management method Methods 0.000 abstract description 3
- 230000009471 action Effects 0.000 description 17
- 230000006870 function Effects 0.000 description 10
- 230000003993 interaction Effects 0.000 description 7
- 230000008901 benefit Effects 0.000 description 4
- 238000007405 data analysis Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 238000004146 energy storage Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000013439 planning Methods 0.000 description 2
- 238000012300 Sequence Analysis Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 230000000875 corresponding effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000005206 flow analysis Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000010223 real-time analysis Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Abstract
The invention relates to a maintenance decision support method for an energy router of a flexible direct current traction power supply system, which is characterized by monitoring and collecting key performance indexes and environmental data of the energy router in real time, carrying out anomaly detection, trend identification and performance degradation prediction by using a data processing algorithm based on a time sequence, establishing a maintenance database by combining the real-time data and a historical maintenance record, analyzing the database by a data mining technology to identify a correlation mode between maintenance activities and system performance, designing and implementing a rule-based reasoning system, generating maintenance decision advice, combining the advice with the real-time data, generating a specific operation guide and maintenance plan, carrying out fault risk prediction and management by applying a fault mode and an influence analysis technology, displaying the operation guide and the maintenance plan through an interactive interface, simultaneously collecting feedback to continuously optimize the reasoning system and the maintenance plan, implementing a self-adaptive learning algorithm, and adjusting and perfecting the maintenance database and the data mining model according to the running state and the maintenance effect of the system.
Description
Technical Field
The invention belongs to the technical field of flexible direct current traction power supply systems, and particularly relates to a maintenance decision support method for an energy router of a flexible direct current traction power supply system.
Background
In recent years, urban rail transit flexible direct current traction power supply systems are rapidly developed. In order to integrate a flexible direct current traction power supply system such as a new energy source, an energy storage or a low-voltage 400V system, an energy router becomes core equipment of the flexible direct current traction power supply system, and the performance and the reliability of the energy router are critical to the stable operation of the whole system. Conventional maintenance methods typically rely on periodic inspection and fault response, which is not only inefficient, but also unable to predict and prevent sudden faults. In addition, due to the lack of deep analysis and intelligent maintenance decision support on real-time data, the traditional method is difficult to cope with the problems of complex and changeable operation environments, equipment aging and the like.
With the development of technology, particularly in the field of data processing and machine learning, new opportunities are provided to improve the maintenance strategies of energy routers. However, existing solutions have shortcomings in integrating real-time data analysis, intelligent decision support, and user interaction. Many systems fail to take full advantage of the data collected to optimize maintenance planning, and also lack an efficient way to predict and manage potential failure risk of equipment.
Therefore, it is necessary to study a new method capable of comprehensively utilizing real-time monitoring data, advanced data processing technology and intelligent decision support system to improve maintenance efficiency, reduce operation cost, and improve performance and reliability of the entire traction power supply system.
Disclosure of Invention
The invention provides a maintenance decision support method for an energy router of a flexible direct current traction power supply system, which aims to solve the limitation of the maintenance strategy of the energy router in the flexible direct current traction power supply system of urban rail transit, and particularly aims to solve the problems that sudden faults of equipment cannot be effectively predicted and prevented, and real-time data analysis and intelligent maintenance decision support are lacking. The method improves maintenance efficiency, reduces operation cost, and improves performance and reliability of the whole traction power supply system. By integrating real-time data monitoring, advanced data processing algorithms and an intelligent decision support system, the method provides a comprehensive maintenance strategy for the energy router.
The invention realizes the aim through the following technical scheme:
a maintenance decision support method for an energy router of a flexible direct current traction power supply system comprises the following steps:
s1: deploying an advanced sensor network system to monitor and collect key performance indexes and environmental data of the energy router in real time;
s2: analyzing the real-time data stream by using an anomaly detection, trend identification and performance degradation prediction data processing algorithm based on a time sequence;
s3: establishing a maintenance database based on the real-time operation data and the historical maintenance record;
s4: constructing a data mining model, analyzing and maintaining a database by adopting a data mining technology, and identifying a correlation mode between maintenance activities and the performance of the energy router system;
s5: designing and implementing a rule-based reasoning system for generating maintenance decision suggestions;
s6: combining the maintenance advice generated by the reasoning system with the real-time operation data to generate a specific operation instruction and a maintenance plan, and performing a judging process to evaluate the feasibility of the advice and the potential influence on the operation of the energy router system, generating the specific operation instruction and the maintenance plan, if the specific operation instruction and the maintenance plan are feasible, continuing the next step, otherwise, returning to the execution step S3;
s7: predicting and managing potential fault risks by applying fault modes and influence analysis technologies;
s8: displaying an operation guide and a maintenance plan through an interactive user interface, and collecting user feedback;
s9: continuously optimizing an inference system and a maintenance plan according to user feedback and energy router system operation data;
s10: and implementing a self-adaptive learning algorithm, continuously adjusting and perfecting a maintenance database and a data mining model according to the latest system running state and maintenance effect, and continuously optimizing and adjusting a maintenance strategy.
Further, the key performance indexes of the energy router in the step S1 include locomotive load fluctuation, energy efficiency ratio and heat dissipation efficiency, and the environmental data include environmental temperature and humidity.
Further, the advanced sensor network system deployed in step S1 further includes sensors for monitoring electrical characteristics of the energy router, including a current sensor, a voltage sensor, and a power sensor.
Further, the data mining technique applied in step S4 employs cluster analysis or neural network.
Further, the data mining model process constructed in the step S4 is as follows:
s4.1: feature selection and extraction: selecting key characteristics of equipment performance indexes and environmental parameters from a large amount of collected data, and carrying out characteristic extraction and dimension reduction by applying a principal component analysis technology;
s4.2: pattern recognition and association analysis: a clustering or association rule mining algorithm is applied to identify significant patterns and correlations in the data, and relationships among energy router fault patterns, performance trends and environmental factors are identified;
s4.3: predictive modeling: regression analysis, time series analysis, or neural network prediction algorithms are used to predict future states and potential risks of the energy router, taking into account environmental changes and the effects of operating conditions.
Further, the data mining in step S4 includes a similarity analysis process based on historical maintenance cases.
Further, the rule-based reasoning system in step S5 adopts a decision tree algorithm, and the construction of the reasoning system is specifically as follows:
s5.1: design of a rule system:
firstly, defining targets, and determining main targets of an inference system: generating maintenance decision suggestions of the energy router; then constructing rules, and constructing a group of maintenance decision rules based on expert knowledge and historical maintenance data, wherein the rules cover different scenes, and the specific scenes comprise specific types of faults, performance degradation and predictive maintenance requirements;
s5.2: data input:
inputting real-time data collected from an advanced sensor network system into an inference system through real-time data integration, wherein the input data comprises locomotive load fluctuation, energy efficiency ratio, heat dissipation efficiency, ambient temperature and humidity, and meanwhile, historical data is input as a reference;
s5.3: rule application:
developing rule matching, and according to the input data, the reasoning system applies a rule base to identify rules conforming to the current situation; executing decision logic, and generating maintenance decision suggestion by the reasoning system based on the matched rule;
s5.4: maintenance recommendation output:
generating maintenance decision suggestions and matching rules according to an inference system, wherein the inference system proposes specific maintenance operation suggestions, and at least one round of refinement is carried out on the suggestions according to specific conditions and running environments of equipment, so that the practicability and feasibility of the suggestions are ensured;
s5.5: system verification and tuning:
performing test verification, testing suggestions of the inference system in an actual environment, and verifying the validity and accuracy of the suggestions; and adjusting and optimizing the rule according to the test result and the user feedback through a feedback loop mechanism.
Further, the interactive user interface in step S8 includes data visualization tools, in particular charts and dashboards.
Further, the process of continuous optimization in step S9 includes applying a feedback loop.
Further, the adaptive learning algorithm applied to the data mining model in step S10 adopts a reinforcement learning technique.
Compared with the prior art, the invention has the following beneficial effects:
firstly, the performance of the energy router under different running conditions is monitored in real time through the advanced sensor network system, including the performance under new energy and energy storage application and the stability in a low-voltage 400V system. The real-time monitoring not only improves the accuracy of fault detection, but also provides a reliable data base for predictive maintenance.
And secondly, the invention adopts a data processing algorithm based on a time sequence, so that the analysis of key performance indexes such as locomotive load fluctuation, energy efficiency ratio, heat dissipation efficiency and the like is more accurate, thereby effectively predicting equipment performance degradation and ensuring the efficient operation of the system.
In addition, the invention extracts the mode from the history maintenance record by utilizing the data mining technology, combines an intelligent reasoning system, provides scientific and reasonable maintenance decision advice for maintenance teams, and greatly improves the pertinence and the efficiency of maintenance work. The fault mode and the influence analysis technology are comprehensively utilized, the management capability of potential fault risks is further improved, and higher reliability and safety are brought to flexible direct current traction power supply equipment in the urban rail transit system.
In summary, through the application of the technology, the invention not only optimizes the operation and maintenance of the energy router, but also provides powerful support for the overall performance and stability of the urban rail transit system.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those skilled in the art from this disclosure that the drawings described below are merely exemplary and that other embodiments may be derived from the drawings provided without undue effort.
FIG. 1 is a flow chart of an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other. The invention will be described in detail below with reference to the drawings in connection with embodiments.
As shown in fig. 1, in a specific embodiment of the maintenance decision support method for an energy router of a flexible direct current traction power supply system of the present invention, the method includes the following steps:
s1: an advanced sensor network system is deployed on an energy router in a flexible direct current traction power supply system, and key performance indexes of the energy router and environmental temperature and humidity data are captured and collected in real time. These data are critical to assessing the operation of the energy router under different environmental conditions.
The deployed advanced sensor network system also includes current sensors, voltage sensors, and power sensors for monitoring the electrical characteristics of the energy router. By analyzing these electrical characteristic data, the energy router system is able to more accurately assess the operational efficiency, load handling capability, and performance issues that may exist of the energy router. In particular, in the context of data mining and fault diagnosis, current, voltage and power data are particularly critical to identify potential failure modes and performance degradation.
Key performance indicators for an energy router include:
the locomotive load fluctuates and this parameter is used to monitor the change in locomotive load.
The energy efficiency ratio, the parameter is used for measuring the energy efficiency performance of the energy router.
The heat dissipation efficiency, which is a parameter used to evaluate the heat dissipation capacity of the energy router.
The environmental data parameters include:
ambient temperature, the temperature data collected is monitored by an ambient monitoring sensor.
Ambient humidity, the humidity data collected is monitored by an ambient monitoring sensor.
Through the advanced sensor network, comprehensive data monitoring can be provided for the energy router, and the data provides a necessary real-time information basis for maintenance decision-making, which is vital for subsequent maintenance decision-making and risk management.
S2: performing real-time data flow analysis by using an anomaly detection, trend identification and performance degradation prediction data processing algorithm based on a time sequence; analysis helps identify potential performance problems and failure risks in advance, thereby making maintenance more proactive and preventative.
The performance degradation prediction data processing algorithm comprises the following steps:
step 1: data preparation and preprocessing are carried out, and historical and real-time equipment operation data are collected and preprocessed to ensure the quality and consistency of the data.
Step 2: and extracting features, selecting key features from the preprocessed data, and extracting features and reducing dimensions by methods such as principal component analysis to simplify a model and improve efficiency.
Step 3: and analyzing the data and the patterns, and analyzing important patterns and relations in the data by using algorithms such as cluster analysis, association rule mining and the like to reveal the interaction between the performance of the energy router and the operation condition.
Step 4: and (3) establishing a prediction model, and constructing a model for predicting the performance degradation of the energy router by using regression analysis, time sequence analysis or a neural network and other methods based on the selected characteristics and the mode analysis result.
Step 5: and (3) evaluating and optimizing performance, evaluating a prediction result of the prediction model, analyzing the accuracy of the prediction model by comparing the prediction data with actual performance data, and adjusting the parameters of the prediction model according to feedback.
Step 6: and (3) carrying out iterative updating on the prediction model, and periodically updating the prediction model to adapt to new data and environmental changes, so as to ensure the accuracy and adaptability of the prediction model.
Step 7: and continuously monitoring and feeding back the circulation, continuously monitoring the running state of the energy router, and feeding back the monitoring result to the prediction model to form a continuous learning and adapting process.
For example, by analyzing the pattern of load fluctuations, an overload condition that an energy router may face may be predicted. These analysis results are used to guide maintenance decisions to ensure that appropriate measures are taken before the problem becomes a significant fault.
S3: and establishing a maintenance database according to the real-time operation data and the historical maintenance record. This database contains not only detailed information of the history maintenance activities, but also performance history data of the energy routers. These data are analyzed by data mining techniques to identify patterns of association between maintenance activities and system performance that help predict future likely maintenance needs and formulate more efficient maintenance plans. This step provides data support for developing an effective maintenance strategy.
S4: constructing a data mining model, analyzing a maintenance database through a data mining technology to identify a correlation mode between maintenance activities and the performance of the energy router system; the applied data mining technology comprises cluster analysis and neural network, and deep pattern recognition and association analysis are carried out.
Data mining includes similarity analysis based on historical maintenance cases to identify possible maintenance patterns and policies.
Data mining models are key technologies for analyzing large amounts of data, identifying patterns and associations, and predicting future trends. The construction process comprises the following steps:
s4.1: feature selection and extraction: and selecting key characteristics of equipment performance indexes and environmental parameters from a large amount of collected data, and performing characteristic extraction and dimension reduction by applying a Principal Component Analysis (PCA) technology.
S4.2: pattern recognition and association analysis: applying a clustering or association rule mining algorithm to identify significant patterns and correlations in the data, identifying relationships between energy router failure modes, performance trends, and environmental factors;
s4.3: predictive modeling: regression analysis, time series analysis, or neural network prediction algorithms are used to predict future states and potential risks of the device, taking into account environmental changes and the effects of operating conditions.
The constructed data mining model has adaptability, reliability, accuracy, real-time performance, user interaction and interpretation.
The adaptation model should be able to adapt to new data, continually updated to reflect the latest system state and operating conditions.
Reliability and accuracy guarantee high accuracy and reliability, false alarms and false misses are reduced, and cross verification and other technologies are implemented to verify the robustness of the model.
Real-time performance is that real-time data can be processed, maintenance decision suggestions can be generated rapidly, and calculation efficiency is optimized so as to support real-time or near real-time analysis.
User interaction and interpretation: a friendly interface is provided for a user, the model output is convenient to monitor and interpret, and the decision making process of the model is transparent enough to facilitate the user to understand and trust the suggestion of the model.
Through the design and the limitation, the data mining model can effectively support maintenance decisions, and help an operation team predict and prevent potential problems, so that the reliability and the performance of the energy router of the flexible direct current traction power supply system are improved.
S5: designing and implementing a rule-based reasoning system for generating maintenance decision suggestions; the rule-based reasoning system employs decision tree algorithms to support more accurate maintenance decision generation. The inference system combines insights from real-time data and historical data to provide more accurate and reliable maintenance recommendations.
Similarity analysis is used to identify possible maintenance patterns and policies from the historical maintenance data, and decision tree algorithms are used to construct an inference system based on these patterns and policies. The design and construction process of the inference system includes defining objectives, constructing maintenance decision rules, integrating real-time and historical data as inputs, and applying decision tree algorithms for rule matching and maintenance decision suggestion generation. In this process, the decision tree algorithm uses patterns and policies derived from the similarity analysis to help determine maintenance actions that should be taken in a particular situation. The similarity analysis provides the basis for data and insight for decision tree algorithms that use this information to generate specific maintenance decision suggestions. This combination allows the inference system to more effectively predict and resolve potential maintenance issues.
The specific process of constructing the rule-based reasoning system is as follows:
s5.1: design of rule system.
Firstly, defining a target, and determining a main target of an inference system, namely generating maintenance decision suggestions of an energy router; and then constructing rules, and constructing a group of maintenance decision rules based on expert knowledge and historical maintenance data, wherein the rules can cover different scenes and are specifically as follows: specific types of faults, performance degradation, predictive maintenance requirements.
S5.2: and (5) inputting data.
Inputting real-time data collected from an advanced sensor network system into an inference system through real-time data integration, wherein the data comprise locomotive load fluctuation, energy efficiency ratio, heat dissipation efficiency, ambient temperature and humidity; and based on historical data references, historical maintenance records and performance data are combined to provide a more comprehensive maintenance decision basis.
S5.3: rules are applied.
Developing rule matching, and according to the input data, the reasoning system applies a rule base to identify rules conforming to the current situation; executing the decision logic, the system generates maintenance decision suggestions based on the matched rules. If a decrease in heat dissipation efficiency is detected, the inference system may recommend checking the cooling system.
S5.4: maintaining a recommendation output.
The reasoning system generates maintenance decision suggestions, and according to the matched rules, the reasoning system proposes specific maintenance operation suggestions such as component replacement, adjustment setting or inspection; according to the specific condition and the running environment of the equipment, at least one round of refinement is performed on the advice, so that the practicability and the feasibility of the advice are ensured.
S5.5: and (5) system verification and optimization.
And performing test verification, and testing suggestions of the reasoning system in an actual environment to verify the effectiveness and accuracy of the system. And adjusting and optimizing the rule according to the test result and user feedback through a feedback loop mechanism so as to improve the decision accuracy and adaptability of the system.
The process of adjusting and optimizing rules involves the steps of:
step 1: and analyzing the test result. The performance of the inference system in the actual environment is evaluated, and in particular, the decision validity and accuracy of the inference system are concerned.
Step 2: user feedback is collected. Feedback is obtained from a user using the system to learn about the actual conditions of the system in terms of performance, accuracy, user experience, and the like.
Step 3: an optimization point is identified. Based on the test results and user feedback, the place in the rule system where improvement is needed is identified. This may include improving the data processing method, adjusting algorithm parameters, or reconstructing certain rules.
Step 4: and (5) performing adjustment. Actual adjustments to rules and algorithms are made, including programming changes, updating rule logic, or introducing new data analysis techniques.
Step 5: and (5) testing again. After the adjustment, the test is carried out again to verify the effect of the modification, so that the performance and decision accuracy of the system are improved.
The tuning process is iterative and requires constant tuning and optimization of the system based on feedback and test results until the desired performance level is reached.
Through the key links, the rule-based reasoning system can provide real-time and accurate maintenance decision support for the energy router of the flexible direct-current traction power supply system, so that the efficiency and effectiveness of equipment maintenance are improved.
S6: combining the maintenance advice of the reasoning system with the real-time operation data to generate a specific operation instruction and a maintenance plan, and performing a judging process to evaluate the feasibility of the advice and the potential influence on the operation of the energy router system, generating the specific operation instruction and the maintenance plan, if the specific operation instruction and the maintenance plan are feasible, continuing the next step, otherwise, returning to the step S3;
and (3) generating an operation guide and a maintenance plan, and combining the maintenance suggestion of the reasoning system with the real-time operation data to generate a specific operation guide and a specific maintenance plan. This step ensures timeliness and pertinence of maintenance activities.
And the reasoning system generates specific maintenance decision suggestions by using the analysis result. These recommendations combine real-time operational data with historical maintenance experience to provide specific operational guidelines and maintenance plans. For example, if the analysis shows an increase in the probability of failure of a particular component, the inference system may recommend checking or replacing the component.
S7: and predicting and managing potential fault risks by applying fault mode and influence analysis (FMEA) technology, so as to further improve the reliability and safety of the system. This analysis takes into account the various possible failure modes and their impact on system performance, helping to formulate a more comprehensive maintenance strategy.
S8: displaying an operation guide and a maintenance plan for a maintenance team through an interactive user interface, and collecting feedback of user experience and advice of the use of the reasoning system; the interactive user interface contains data visualization tools, in particular charts and dashboards, through which information such as key performance indicators, maintenance recommendations, schedule of plans, etc. should be presented clearly to facilitate a maintenance team's quick understanding of maintenance recommendations and plans. At the same time, user feedback is collected through the interface and used to continuously optimize the inference system and maintenance planning. This interaction mechanism increases the transparency and engagement of the maintenance process.
S9: continuously optimizing an inference system and a maintenance plan according to user feedback and energy router system operation data; wherein the process of continuous optimization includes applying a feedback loop to adjust parameters and logic of the inference system in real time, continuously improving performance of the inference system, ensuring continuous improvement and adaptability of maintenance strategies.
S10: and implementing a self-adaptive learning algorithm, continuously adjusting and perfecting a maintenance database and a data mining model according to the latest system running state and maintenance effect, and continuously optimizing and adjusting a maintenance strategy. The adaptive learning algorithm employs reinforcement learning techniques to automatically adjust the data mining model to accommodate new operating conditions and maintenance feedback. This ensures that the process evolves over time, adapting to new operating conditions and challenges.
For example, if new operating conditions or maintenance strategies prove to be effective, the learning algorithm will automatically incorporate this information into future maintenance decisions. Such an adaptive approach ensures that the overall system is able to continue to evolve over time, adapting to new challenges and conditions.
Reinforcement learning is a machine learning method in which a learning agent learns how to take action to maximize a certain jackpot by interacting with the environment. The aim is to enable the data mining model to automatically adjust to new operating conditions and maintenance feedback, thereby improving the accuracy and efficiency of the model.
The adaptive learning algorithm employs a reinforcement learning technique, which is centered on learning optimal behavior or strategies through interactions with the environment to maximize the jackpot over a period of time. The technique is applied to dynamically adjust the data mining model to accommodate new operating conditions and maintenance feedback. The following is a specific description of reinforcement learning techniques:
first, description is made of the basic elements of reinforcement learning including:
agent: an entity making a decision in an enhanced learning environment is referred to herein as a data mining model.
Environment: the system with which the agent interacts, i.e., the energy router and its operating environment.
Status: representing the current situation of the environment, such as the operational data and maintenance history of the energy router.
The actions are as follows: the agent may take action in a given state, such as adjusting model parameters or changing prediction strategies.
Rewarding: the feedback that the agent gets for the action it takes is typically a numerical value indicating the effect of the action.
Strategy: from a mapping of states to actions, policies define actions that should be taken in a particular state.
The specific process of learning is as follows:
step 1: implementing exploration and utilization policies balance. The agent needs to balance between exploring new strategies (i.e., trying unknown actions to find potentially valid solutions) and utilizing known best strategies (i.e., applying the most efficient methods currently known). This balance is critical to ensure that new strategies that may be better are discovered while maintaining current availability.
Step 2: the state-action policy mapping is implemented. The agent selects the corresponding action based on the current state. This involves analyzing each specific state and determining the most appropriate course of action in this case. This process is a dynamic decision process in which the agent continually evaluates the current state and makes the best action selection based on this information.
Step 3: the cost function is used to predict the expected rewards that may be obtained after taking a particular action or strategy.
Step 4: the strategy and cost function are adjusted according to the reward feedback to optimize future decisions.
In reinforcement learning, a cost function is used to estimate the expected rewards that can be achieved under a given state or after taking a particular action. Such estimations help guide agents in making optimal decisions when facing different situations. The cost functions are generally divided into two types: a state cost function and an action cost function.
A state-cost function describing the expected return from a state under a particular policy. The formula is generally expressed as:
wherein,is in policy->Lower state->Value of->Is at the time oftIs awarded (1)>Is at the time oftIs a state of (2).
An action cost function describing the expected return from the current state after taking a particular action. The formula is generally expressed as:
wherein,is in policy->In the state->Take action->Value of->Is at the time oftAction taken.
The goal of reinforcement learning algorithms is to learn these cost functions through interactions with the environment so that long-term benefits of taking different actions can be predicted and optimal decisions made accordingly. In the context of dynamic security assessment models, a cost function may help the model learn of the long-term benefits of taking different maintenance decisions in a particular state.
The invention aims to solve the limitation of an energy router maintenance strategy in the existing rail transit system, and particularly aims to solve the problems that sudden faults of equipment cannot be effectively predicted and prevented, and real-time data analysis and intelligent maintenance decision support are lacking. The invention has the advantages of improving the maintenance efficiency, reducing the operation cost and improving the performance and the reliability of the whole traction power supply system. By integrating real-time data monitoring, advanced data processing algorithms and an intelligent decision support system, a comprehensive maintenance strategy is provided for the energy router.
The invention not only enhances the maintenance efficiency and the preventive property of the energy router, but also obviously improves the performance and the reliability of the whole traction power supply system through the intelligent decision support system.
The embodiments of the present invention have been described in detail above, but the present invention is not limited to the described embodiments. It will be apparent to those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, and yet fall within the scope of the invention.
Claims (10)
1. The maintenance decision support method for the energy router of the flexible direct current traction power supply system is characterized by comprising the following steps of:
s1: deploying an advanced sensor network system to monitor and collect key performance indexes and environmental data of the energy router in real time;
s2: analyzing the real-time data stream by using an anomaly detection, trend identification and performance degradation prediction data processing algorithm based on a time sequence;
s3: establishing a maintenance database based on the real-time operation data and the historical maintenance record;
s4: constructing a data mining model, analyzing and maintaining a database by adopting a data mining technology, and identifying a correlation mode between maintenance activities and the performance of the energy router system;
s5: designing and implementing a rule-based reasoning system for generating maintenance decision suggestions;
s6: combining the maintenance advice generated by the reasoning system with the real-time operation data to generate a specific operation instruction and a maintenance plan, and performing a judging process to evaluate the feasibility of the advice and the potential influence on the operation of the energy router system, generating the specific operation instruction and the maintenance plan, if the specific operation instruction and the maintenance plan are feasible, continuing the next step, otherwise, returning to the execution step S3;
s7: predicting and managing potential fault risks by applying fault modes and influence analysis technologies;
s8: displaying an operation guide and a maintenance plan through an interactive user interface, and collecting user feedback;
s9: continuously optimizing an inference system and a maintenance plan according to user feedback and energy router system operation data;
s10: and implementing a self-adaptive learning algorithm, continuously adjusting and perfecting a maintenance database and a data mining model according to the latest system running state and maintenance effect, and continuously optimizing and adjusting a maintenance strategy.
2. The method of claim 1, wherein the key performance indicators of the energy router in step S1 include locomotive load fluctuation, energy efficiency ratio, and heat dissipation efficiency, and the environmental data includes environmental temperature and humidity.
3. The method of claim 1, wherein the advanced sensor network system deployed in step S1 further comprises sensors for monitoring electrical characteristics of the energy router, including current sensors, voltage sensors, and power sensors.
4. The method according to claim 1, wherein the data mining technique applied in step S4 employs cluster analysis or neural networks.
5. The method according to claim 1, wherein the data mining model process constructed in step S4 is:
s4.1: feature selection and extraction: selecting key characteristics of equipment performance indexes and environmental parameters from a large amount of collected data, and carrying out characteristic extraction and dimension reduction by applying a principal component analysis technology;
s4.2: pattern recognition and association analysis: a clustering or association rule mining algorithm is applied to identify significant patterns and correlations in the data, and relationships among energy router fault patterns, performance trends and environmental factors are identified;
s4.3: predictive modeling: regression analysis, time series analysis, or neural network prediction algorithms are used to predict future states and potential risks of the energy router, taking into account environmental changes and the effects of operating conditions.
6. The method according to claim 1, wherein the data mining in step S4 comprises a similarity analysis process based on historical maintenance cases.
7. The method according to claim 6, wherein the rule-based reasoning system in step S5 employs a decision tree algorithm, and the reasoning system is constructed specifically as follows:
s5.1: design of a rule system:
firstly, defining targets, and determining main targets of an inference system: generating maintenance decision suggestions of the energy router; then constructing rules, and constructing a group of maintenance decision rules based on expert knowledge and historical maintenance data, wherein the rules cover different scenes, and the specific scenes comprise specific types of faults, performance degradation and predictive maintenance requirements;
s5.2: data input:
inputting real-time data collected from an advanced sensor network system into an inference system through real-time data integration, wherein the input data comprises locomotive load fluctuation, energy efficiency ratio, heat dissipation efficiency, ambient temperature and humidity, and meanwhile, historical data is input as a reference;
s5.3: rule application:
developing rule matching, and according to the input data, the reasoning system applies a rule base to identify rules conforming to the current situation; executing decision logic, and generating maintenance decision suggestion by the reasoning system based on the matched rule;
s5.4: maintenance recommendation output:
generating maintenance decision suggestions and matching rules according to an inference system, wherein the inference system proposes specific maintenance operation suggestions, and at least one round of refinement is carried out on the suggestions according to specific conditions and running environments of equipment, so that the practicability and feasibility of the suggestions are ensured;
s5.5: system verification and tuning:
performing test verification, testing suggestions of the inference system in an actual environment, and verifying the validity and accuracy of the suggestions; and adjusting and optimizing the rule according to the test result and the user feedback through a feedback loop mechanism.
8. The method according to claim 1, wherein the interactive user interface in step S8 comprises data visualization tools.
9. The method according to claim 1, wherein the process of continuous optimization in step S9 comprises applying a feedback loop.
10. The method according to claim 1, wherein the adaptive learning algorithm applied to the data mining model in step S10 employs reinforcement learning techniques.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410153011.7A CN117689373A (en) | 2024-02-04 | 2024-02-04 | Maintenance decision support method for energy router of flexible direct-current traction power supply system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410153011.7A CN117689373A (en) | 2024-02-04 | 2024-02-04 | Maintenance decision support method for energy router of flexible direct-current traction power supply system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117689373A true CN117689373A (en) | 2024-03-12 |
Family
ID=90128663
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410153011.7A Pending CN117689373A (en) | 2024-02-04 | 2024-02-04 | Maintenance decision support method for energy router of flexible direct-current traction power supply system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117689373A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103605771A (en) * | 2013-11-28 | 2014-02-26 | 东莞中国科学院云计算产业技术创新与育成中心 | Intelligent assistant decision and maintenance system and method for operating same |
| CN104573062A (en) * | 2015-01-23 | 2015-04-29 | 桂林电子科技大学 | Intelligent learning method based on description logic and case-based reasoning |
| CN106017551A (en) * | 2016-05-16 | 2016-10-12 | 国网河南省电力公司电力科学研究院 | Intelligent transmission line integrated monitoring analysis and early warning method |
| CN112131212A (en) * | 2020-09-29 | 2020-12-25 | 合肥城市云数据中心股份有限公司 | Hybrid cloud scene-oriented time sequence data anomaly prediction method based on ensemble learning technology |
| CN113112051A (en) * | 2021-03-11 | 2021-07-13 | 同济大学 | Production maintenance joint optimization method for serial production system based on reinforcement learning |
| CN116862199A (en) * | 2023-08-17 | 2023-10-10 | 浙江建设职业技术学院 | Building construction optimizing system based on big data and cloud computing |
| CN117196066A (en) * | 2023-09-15 | 2023-12-08 | 北京红山信息科技研究院有限公司 | Intelligent operation and maintenance information analysis model |
-
2024
- 2024-02-04 CN CN202410153011.7A patent/CN117689373A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103605771A (en) * | 2013-11-28 | 2014-02-26 | 东莞中国科学院云计算产业技术创新与育成中心 | Intelligent assistant decision and maintenance system and method for operating same |
| CN104573062A (en) * | 2015-01-23 | 2015-04-29 | 桂林电子科技大学 | Intelligent learning method based on description logic and case-based reasoning |
| CN106017551A (en) * | 2016-05-16 | 2016-10-12 | 国网河南省电力公司电力科学研究院 | Intelligent transmission line integrated monitoring analysis and early warning method |
| CN112131212A (en) * | 2020-09-29 | 2020-12-25 | 合肥城市云数据中心股份有限公司 | Hybrid cloud scene-oriented time sequence data anomaly prediction method based on ensemble learning technology |
| CN113112051A (en) * | 2021-03-11 | 2021-07-13 | 同济大学 | Production maintenance joint optimization method for serial production system based on reinforcement learning |
| CN116862199A (en) * | 2023-08-17 | 2023-10-10 | 浙江建设职业技术学院 | Building construction optimizing system based on big data and cloud computing |
| CN117196066A (en) * | 2023-09-15 | 2023-12-08 | 北京红山信息科技研究院有限公司 | Intelligent operation and maintenance information analysis model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Vrignat et al. | Sustainable manufacturing, maintenance policies, prognostics and health management: A literature review | |
| Compare et al. | Challenges to IoT-enabled predictive maintenance for industry 4.0 | |
| Pattison et al. | Intelligent integrated maintenance for wind power generation | |
| Álvarez et al. | A condition-based maintenance model including resource constraints on the number of inspections | |
| Choo et al. | Adaptive multi-scale prognostics and health management for smart manufacturing systems | |
| Wang et al. | Reliability analysis for automobile engines: Conditional inference trees | |
| CN116720324A (en) | Traction substation key equipment fault early warning method and system based on prediction model | |
| De Simone et al. | LSTM-based failure prediction for railway rolling stock equipment | |
| Eltyshev et al. | Intelligent diagnostic control and management of the condition of electrotechnical equipment | |
| Wu et al. | Adaptive sequential predictive maintenance policy with nonperiodic inspection for hard failures | |
| Cruz | Evaluating record history of medical devices using association discovery and clustering techniques | |
| CN117114454B (en) | DC sleeve state evaluation method and system based on Apriori algorithm | |
| Pisica et al. | Feature selection filter for classification of power system operating states | |
| Schenkelberg et al. | Supervised machine learning for knowledge-based analysis of maintenance impact on profitability | |
| KR102656115B1 (en) | Remaining useful life prediction method of fuel cell system and digital twin device performing the same | |
| CN117689373A (en) | Maintenance decision support method for energy router of flexible direct-current traction power supply system | |
| Merkt | Predictive models for maintenance optimization: an analytical literature survey of industrial maintenance strategies | |
| Urmeneta et al. | A methodology for performance assessment at system level—Identification of operating regimes and anomaly detection in wind turbines | |
| Aizpurua et al. | Prognostics & health management methods & tools for transformer condition monitoring in smart grids | |
| Zhang et al. | Adaptive alarm prediction in optical network based on model generalization in cross-layer AI | |
| Aggarwal et al. | Automation of Water Distribution System by Prediction of Water Consumption and Leakage Detection Using Machine Learning and IoT | |
| CN116136926A (en) | Underwater robot health management system based on Bayesian network | |
| Velasco-Gallego et al. | Mar-RUL: A remaining useful life prediction approach for fault prognostics of marine machinery | |
| Cvijic et al. | Need for AI in Transformer Diagnostics and Prognostics | |
| CN117713221B (en) | Micro-inversion photovoltaic grid-connected optimization system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |