Dynamic management method and system for optimizing full life cycle cost of equipment

Abstract

The invention discloses a dynamic management method and system for optimizing the full life cycle cost of equipment, and belongs to the technical field of equipment cost management, and the method specifically comprises the steps: collecting equipment operation data in real time, carrying out the preprocessing, generating first collection data, collecting equipment cost data in real time, carrying out the preprocessing, and generating second collection data, and performing evaluation and fault prediction on the health state of the equipment based on the first collection data, and constructing an equipment full life cycle cost model based on the second collection data and the health state evaluation and fault prediction result of the equipment to obtain the operation cost of the equipment in the full life cycle. Dynamically optimizing the life cycle cost of the equipment by using a multi-objective optimization algorithm according to the health condition evaluation and fault prediction result of the equipment and the operation cost of the equipment in the full life cycle; according to the invention, the cost optimization strategy can be dynamically adjusted according to the real-time health condition of the equipment, the fault prediction information and the change of the operation environment.

Description

Technical Field

[0001] The present invention belongs to the technical field of equipment cost management, and specifically relates to a dynamic management method and system for optimizing the total life cycle cost of equipment. Background Art

[0002] Equipment goes through multiple stages during its life cycle, including design, procurement, operation, maintenance, decommissioning, and scrapping. The costs at each stage have a significant impact on the overall life cycle cost of the equipment. Effective life cycle cost management can help enterprises optimize resource allocation, extend the service life of equipment, reduce operating costs, and thus improve production efficiency and economic benefits.

[0003] Although a large number of studies have focused on equipment life cycle management and optimization methods, existing technical solutions often have the following deficiencies: 1) Static optimization model: Traditional equipment life cycle cost optimization methods are mostly static models, usually based on pre-determined budgets and life cycle plans for cost allocation; 2) Single-objective optimization: Most existing technologies focus on optimizing one aspect (such as maintenance costs, operating costs, etc.), ignoring the interrelationships and overall benefits between different cost items; 3) Lack of dynamic adaptability: During the actual use of equipment, the health status and fault prediction results change continuously with time and the usage environment, etc.

[0004] For example, the Chinese patent with the authorization announcement number CN111882162B discloses a method for managing the total life cycle cost of equipment, including the following steps: Synchronize the equipment hierarchical structure in the production management system to the asset management system; Obtain maintenance work orders from the production management system and import them into the asset management system; Obtain the equipment associated with the maintenance work order according to the equipment hierarchical structure and allocate the maintenance cost to the equipment associated with the maintenance work order; Combine the maintenance costs of each equipment and the purchase costs of each equipment in the asset management system to achieve the total life cycle cost management of each equipment. This invention can manage the total life cycle cost of equipment without repeated information entry, and has high management efficiency.

[0005] As disclosed in the Chinese patent with the authorization announcement number CN118628097B, a method and system for equipment full - life - cycle cost management are provided. It collects equipment operation data through a sensor network, performs data cleaning and standardization processing; constructs a multi - dimensional cost analysis model using data mining techniques and machine learning algorithms; dynamically optimizes equipment costs through reinforcement learning algorithms and collaborative optimization mechanisms; constructs and optimizes a cost prediction model based on historical data and real - time data, and uses transfer learning and intelligent scheduling algorithms for real - time cost prediction to generate an optimal maintenance plan and resource allocation plan. Combining with a feedback optimization mechanism for dynamic adjustment, and finally storing the optimization results; through advanced data analysis and intelligent algorithms, the invention realizes accurate prediction and optimized scheduling of equipment operation and maintenance costs, significantly improves the efficiency and economic benefits of equipment management, and provides strong support for enterprises.

[0006] Defects of the above - mentioned patent: 1) Highly dependent on manual operations and repeated data entry, resulting in complex operations and low efficiency; 2) Do not consider cost fluctuations caused by unexpected situations, such as equipment status, equipment failures, etc., and the optimization and adjustment effect is poor. Summary of the Invention

[0007] Aiming at the deficiencies of the prior art, the present invention proposes a dynamic management method and system for optimizing the full - life - cycle cost of equipment, predicting the health status and faults of equipment, associating them with the full - life - cycle cost, and realizing the dynamic optimization of the full - life - cycle cost of equipment by combining real - time data analysis, prediction and decision support.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A dynamic management method for optimizing the full - life - cycle cost of equipment, including:

[0010] Collect equipment operation data in real - time and perform pre - processing to generate the first collected data. Collect equipment cost data in real - time and perform pre - processing to generate the second collected data;

[0011] Based on the first collected data, evaluate the health status of the equipment and predict faults;

[0012] Based on the second collected data, the equipment health status evaluation and fault prediction results, construct an equipment full - life - cycle cost model to obtain the operation cost within the full - life - cycle of the equipment;

[0013] According to the equipment health status evaluation and fault prediction results and the operation cost within the full - life - cycle of the equipment, use a multi - objective optimization algorithm to dynamically optimize the life - cycle cost of the equipment.

[0014] Specifically, the evaluating the health status of the equipment and predicting faults based on the first collected data includes:

[0015] Extract the features of the first acquisition data, including: mean, variance / standard deviation, frequency-domain features, time-domain features, and statistical features, fuse and reduce the dimensions of the extracted features to obtain a multi-dimensional feature vector;

[0016] Establish a multi-level time-series health index evaluation model, input the multi-dimensional feature vector, and evaluate the health status of the device;

[0017] Establish a fault prediction model, and predict the fault type and occurrence time of the device within a preset future time according to the multi-dimensional feature vector and historical data.

[0018] Specifically, the establishment of the multi-level time-series health index evaluation model, inputting the multi-dimensional feature vector, and evaluating the health status of the device includes:

[0019] Divide the device operation state into the first stage, the second stage, and the third stage, and establish an independent time-series model for each stage;

[0020] Convert the data of each stage in the first acquisition data into a spectrum or time-domain image, use the convolutional layer in the convolutional neural network to extract the time-series features of the spectrum or time-domain image, establish a dependency relationship model of the device operation state, and capture the health evolution trend of the device at different time points;

[0021] Convert the extracted time-series features into a health status evaluation result through a fully connected layer.

[0022] Specifically, the establishment of the fault prediction model, predicting the fault type and occurrence time of the device within a preset future time according to the multi-dimensional feature vector and historical data, includes:

[0023] Establish a device component connection graph G, G=(V, E), where V represents the set of nodes and E represents the set of edges;

[0024] Learn the non-linear relationship of device faults through a deep neural network, update the states of the nodes in the device component connection graph G, and construct a fault prediction model;

[0025] Combine evolution and the life cycle of the device, optimize and train the fault prediction model, input the multi-dimensional feature vector into the optimized and trained fault prediction model, and obtain the fault probability and fault occurrence time of the device.

[0026] Specifically, the combination of evolution and the life cycle of the device, and according to the device component connection graph G, predicting in advance the associated device components affected by abnormal device components, and optimizing and training the fault prediction model, includes:

[0027] Encode the parameters of the fault prediction model into gene sequences, where each individual represents a configuration of the fault prediction model;

[0028] According to the prediction error of the fault prediction model, select excellent individuals for crossover operation to generate new configurations of the fault prediction model;

[0029] Introduce new configurations of the fault prediction model through mutation operation, repeatedly generate new configurations of the fault prediction model until the optimal fault prediction model parameters are found, and obtain the optimized fault prediction model;

[0030] Combine the life cycle of the equipment, and according to the equipment component connection diagram G, predict in advance the associated equipment components affected by abnormal equipment components, and train the optimized fault prediction model.

[0031] Specifically, the construction of the equipment life cycle cost model includes:

[0032] Based on the second collection of data and the evaluation of equipment health status and fault prediction results, analyze the cost composition of the equipment, including: the first cost, the second cost, the third cost, the fourth cost, and the fifth cost;

[0033] Construct the equipment life cycle cost model, and the specific formula is:

[0034]

[0035] Among them, Lcc represents the equipment life cycle cost model, that is, the cost of the equipment during its entire life cycle, T represents the expected service life of the equipment, C1 represents the first cost, C 2,t 、C 3,t and C 4,t respectively represent the second cost, the third cost, and the fourth cost within the t time period, and C5 represents the fifth cost.

[0036] Specifically, according to the evaluation of equipment health status and fault prediction results and the operating costs during the equipment life cycle, use the multi-objective optimization algorithm to dynamically optimize the life cycle cost of the equipment, including:

[0037] Convert the optimization problem of the equipment life cycle cost into a multi-objective optimization model, construct the objective function of the multi-objective optimization model, which is the weighted sum of cost items;

[0038] Set the constraint conditions of the multi-objective optimization model, including: equipment health status constraint, fault prediction constraint, and budget constraint;

[0039] Use the multi-objective particle swarm optimization algorithm to solve the multi-objective optimization model, and through iterative update, find multiple optimal solutions of the multi-objective optimization model, that is, different life cycle cost objectives and strategies of the equipment;

[0040] Select the most suitable optimization solution according to actual requirements and budget.

[0041] Specifically, the device operation data includes: device physical data, operating condition data, and environmental data;

[0042] The preprocessing includes: data denoising to remove noise data, outliers, and missing values from the first and second collected data;

[0043] Data standardization is used to convert the first and second collected data into a unified format.

[0044] A dynamic management system for optimizing the total life cycle cost of a device, used to implement the dynamic management method for optimizing the total life cycle cost of a device, includes: a data collection module, an evaluation and prediction module, an operating cost module, and a dynamic management module;

[0045] The data collection module is used to collect device operation data in real time, perform preprocessing, generate the first collected data, collect device cost data in real time, and perform preprocessing to generate the second collected data;

[0046] The evaluation and prediction module is used to evaluate the health status of the device and predict faults based on the first collected data;

[0047] The operating cost module is used to construct a total life cycle cost model of the device based on the second collected data and the results of the device health status evaluation and fault prediction, and obtain the operating cost within the total life cycle of the device;

[0048] The dynamic management module is used to dynamically optimize the life cycle cost of the device according to the results of the device health status evaluation and fault prediction and the operating cost within the total life cycle of the device, using a multi-objective optimization algorithm.

[0049] Specifically, the evaluation and prediction module includes: a feature extraction unit, a health status evaluation unit, and a fault prediction unit;

[0050] The feature extraction unit is used to extract the features of the first collected data;

[0051] The health status evaluation unit is used to establish a multi-level time-series health index evaluation model, input a multi-dimensional feature vector, and evaluate the health status of the device;

[0052] The fault prediction unit is used to establish a fault prediction model, and predict the fault type and occurrence time of the device within a preset future time according to the multi-dimensional feature vector and historical data.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] 1. The present invention proposes a dynamic management method for optimizing the total life cycle cost of equipment, which can dynamically adjust the cost optimization strategy according to the real-time health status of the equipment, fault prediction information, and changes in its operating environment.

[0055] 2. The present invention proposes a dynamic management method for optimizing the total life cycle cost of equipment. The proposed multi-objective optimization method can simultaneously consider multiple life cycle cost objectives of the equipment, take into account the mutual influence and trade-off between different objectives, and through multi-objective optimization, can find the optimal balance among multiple costs, thereby achieving the overall minimization of the equipment life cycle cost.

[0056] 3. The present invention proposes a dynamic management method for optimizing the total life cycle cost of equipment, which can give early warnings and arrange repairs or replacements in time before a fault occurs, thus avoiding production downtime and high repair costs caused by sudden equipment failures. At the same time, it can formulate the most appropriate repair plan based on data such as the operating status and maintenance history of the equipment, reduce unnecessary repairs and replacements, and thereby reduce the operation and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flowchart of a dynamic management method for optimizing the total life cycle cost of equipment provided by the present invention;

[0058] Figure 2 It is an example diagram of the connection of equipment components provided by the present invention;

[0059] Figure 3 It is an architecture diagram of a dynamic management system for optimizing the total life cycle cost of equipment provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] The following will explain the present application in detail with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any form. It should be noted that those of ordinary skill in the art can make several modifications and improvements without departing from the concept of the present application. These all belong to the protection scope of the present application.

[0061] In order to make the purpose, technical solution, and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] It should be noted that if there is no conflict, the various features in the embodiments of the present application can be combined with each other, and all are within the protection scope of the present application. In addition, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different module division in the device or a different order in the flowchart. In addition, the terms "first", "second", "third", etc. used in the present application do not limit the data and the execution order, but only distinguish the same items or similar items with basically the same functions and roles.

[0063] Unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used in this specification in the description of the present application are only for the purpose of describing specific embodiments and are not used to limit the present application. The term "and / or" used in this specification includes any and all combinations of one or more of the related listed items.

[0064] Embodiment 1

[0065] Please refer to Figure 1 and Figure 2 , an embodiment provided by the present invention: a dynamic management method for optimizing the full life cycle cost of equipment, including the following specific steps:

[0066] Step S1: Real-time collect equipment operation data and perform preprocessing to generate first collected data, and real-time collect equipment cost data and perform preprocessing to generate second collected data;

[0067] The equipment operation data includes: equipment physical data (such as acceleration, torque, temperature, pressure, etc.), working condition data (such as equipment start-stop state, running speed, load rate, etc.) and environmental data (humidity, dust concentration, electromagnetic, etc.);

[0068] The equipment cost data includes: purchase cost, operation cost, maintenance cost, failure loss cost and retirement disposal cost;

[0069] The preprocessing includes: data denoising, removing noise data, outliers and missing values in the first collected data and the second collected data;

[0070] For example, for the temperature data collected by the sensor, if there is a value that significantly exceeds the normal range (such as the temperature sensor shows a negative temperature, while this situation is impossible in the actual environment), it is regarded as an outlier for processing. Mean filtering, median filtering and other methods can be used for smoothing processing, or reasonable correction can be made according to the distribution characteristics of the data and business knowledge; for missing values, interpolation methods (such as linear interpolation, polynomial interpolation) or methods based on statistical models can be used for filling;

[0071] Data standardization is used to convert the first collected data and the second collected data into a unified format to ensure data consistency.

[0072] Step S2: Based on the first collected data, evaluate the health status of the device and predict faults.

[0073] The specific steps of Step S2 are as follows:

[0074] Step S201: Extract the features of the first collected data, including: mean, variance / standard deviation, frequency-domain features, time-domain features, and statistical features, etc. Fuse and reduce the dimensions of the extracted features to obtain a multi-dimensional feature vector.

[0075] Frequency-domain features such as the spectrum of vibration data, etc., time-domain features such as maximum value, minimum value, peak value, peak factor, etc., and statistical features include volatility, entropy value, etc.

[0076] Step S202: Establish a multi-level time-series health index evaluation model, input the multi-dimensional feature vector, and evaluate the health status of the device.

[0077] Step S203: Establish a fault prediction model, and predict the fault type and occurrence time of the device within a preset future time according to the multi-dimensional feature vector and historical data.

[0078] The specific steps of Step S202 are as follows:

[0079] Step S2021: Divide the device operation status into the first stage, the second stage, and the third stage, and establish an independent time-series model for each stage.

[0080] In this embodiment, the first stage can be the startup stage, establish a time-series model based on the device startup process, analyze the changes in indicators such as temperature, vibration, and current during the startup process, and capture potential abnormalities in the startup stage; the second stage can be the stable stage, for the normal operation period of the device, establish a time-series model based on device load, vibration spectrum, etc., and analyze whether there are problems such as load fluctuations and efficiency decay; the third stage can be the shutdown stage, analyze the temperature and pressure changes during the shutdown process, and evaluate the recovery speed and health status of the device after shutdown.

[0081] Step S2022: Convert the data of each stage in the first collected data into a spectrum or time-domain image, use the convolutional layer in the convolutional neural network to extract the time-series features of the spectrum or time-domain image, establish a long-term dependence relationship model of the device operation status, and capture the health evolution trend of the device at different time points.

[0082] In this embodiment, a long-term dependence relationship model of the device operation state is established using an LSTM model. LSTM can selectively remember or forget historical information through its internal forget gate, input gate, and output gate mechanisms, and can effectively analyze the changes in time series features in the decline stage and capture long-term dependence relationships. Among them, the input layer: The data of each stage in the first acquisition data within each time window (data of the first stage such as vibration, temperature, current, etc.) is used as the input of the LSTM; LSTM layer: Stacking multiple LSTM layers can enhance the depth of the model, enabling it to capture complex patterns during the device decline process. The memory function of LSTM helps the model not only focus on the feature changes at the current time point but also capture the influence of historical data on the current state when analyzing the decline stage. Output layer: The output of the LSTM can be used to predict the health state of the device.

[0083] Step S2023: Convert the extracted time series features into a health condition assessment result through a fully connected layer.

[0084] The specific steps of step S203 are as follows:

[0085] Step S2031: Establish a device component connection graph G, G=(V, E), where V represents the set of nodes, that is, each component of the device, and E represents the set of edges, that is, the relationships between components;

[0086] Step S2032: Learn the non-linear relationship of device failures through a deep neural network, update the states of the nodes in the device component connection graph G, and construct a failure prediction model;

[0087] In this embodiment, using a graph neural network GNN, the state of the nodes in the device component connection graph G is updated through convolutional operations. The specific formula is:

[0088] Among them, represents the state of the i-th node in the (k + 1)-th layer, σ() represents the activation function, represents the state of the i-th node in the k-th layer, N(i) represents the set of neighbor nodes of node i, W k represents the weight matrix, b k represents the bias;

[0089] Step S2033: Combine evolution and the life cycle of the device, and based on the device component connection graph G, predict in advance the associated device components affected by abnormal device components, optimize and train the failure prediction model, and input the multi-dimensional feature vector into the optimized and trained failure prediction model to obtain the failure probability and failure occurrence time of the device.

[0090] In step S2033, the fault prediction model is optimized and trained by combining evolution and the life cycle of the device, including: encoding the fault prediction model parameters into gene sequences, where each individual represents a fault prediction model configuration;

[0091] According to the prediction error of the fault prediction model such as MSE (Mean Squared Error), select excellent individuals for crossover operation to generate new fault prediction model configurations;

[0092] Introduce new fault prediction model configurations through mutation operation, repeatedly generate new fault prediction model configurations until the optimal fault prediction model parameters are found, and obtain the optimized fault prediction model;

[0093] Combine the life cycle of the device to train the optimized fault prediction model.

[0094] In this embodiment, by combining the device component connection diagram and real-time monitoring data, the artificial intelligence algorithm is used to perform real-time analysis on the node status. When an abnormal trend appears in a certain node but no fault has occurred yet, based on the association between nodes in the connection diagram, other components that may be affected are predicted in advance. Then, an improved genetic algorithm is used to optimize the parameters of the fault prediction model. The evolutionary computing method can effectively search for the global optimal solution, avoid the problem of local optimal solutions, and at the same time facilitate better finding of device component faults.

[0095] Combine Figure 2 , where A1 to A13 in the figure are all nodes in the device component connection diagram, that is, the components of the device. The arrow connections in the figure are the associations between the components of the device. If node A1 is the starting node and node A5 is the fault node, it is necessary to find the nodes affected by the fault node A5. Starting from node A1, the reachable nodes are A2, A7, A6, A10, A12, A13. From Figure 2 it can be seen that node A2 needs to be selected. When optimizing the fault prediction model, select excellent individuals for crossover operation to generate new fault prediction model configurations, introduce new fault prediction model configurations through mutation operation, repeatedly generate new fault prediction model configurations until the optimal fault prediction model parameters are found, that is, including paths such as A1 - A2 - A3 - A4 - A5, A1 - A7 - A8, A1 - A6 - A9, A1 - A10 - A11, A1 - A12, and A1 - A13, and also include return paths, etc. Select the optimal path from them. After iteration, the optimal A1 - A2 - A3 - A4 - A5 can be obtained. Optimize according to the above content to obtain the optimized fault prediction model and perform training.

[0096] Step S3: Based on the second acquisition data and the evaluation of the device health status and the fault prediction results, construct a device life cycle cost model to obtain the operating cost within the device life cycle;

[0097] The specific steps of step S3 are as follows:

[0098] Step S301: Analyze the cost composition of the device based on the second acquisition data and the results of equipment health status assessment and fault prediction, including: the first cost, the second cost, the third cost, the fourth cost, and the fifth cost;

[0099] In this embodiment, the first cost may be the acquisition cost, including the purchase price of the device, transportation costs, installation and commissioning costs, taxes, etc. The acquisition cost is basically determined at the time of device procurement, but dynamic adjustments can be considered due to factors such as inflation;

[0100] The second cost may be the operating cost, including the energy consumption cost and the consumable cost. The energy consumption cost is calculated based on the power of the device, the operating time, and the real-time electricity price. The real-time electricity price can be dynamically adjusted according to the price fluctuations in the electricity market or the contract electricity price signed between the enterprise and the power supply company. The consumable cost is calculated based on the consumable consumption rate of the device and the consumable price. The consumable price may change due to factors such as market supply and demand relationships and raw material price fluctuations, and needs to be updated in real time;

[0101] The third cost may be the maintenance cost, including: the regular maintenance cost and the fault repair cost. The regular maintenance cost is to formulate a maintenance plan and maintenance items based on the health status of the device, and calculate the cost according to the maintenance plan and maintenance items; The fault repair cost is related to the fault probability and fault repair cost of the device, including spare part replacement cost, manual repair cost, etc.;

[0102] The fourth cost may be the fault loss cost, including production downtime losses, product quality losses, order default losses, etc. caused by device faults. The production downtime loss can be calculated based on the device downtime, the production line capacity, and the unit product profit. The product quality loss can be calculated based on the defective rate caused by the fault and the defective product processing cost. The order default loss is calculated according to the default terms agreed in the contract;

[0103] The fifth cost may be the retirement disposal cost, which is calculated based on factors such as the residual value of the device, the disassembly and recycling cost, and the waste treatment cost. The residual value of the device can be calculated through the device depreciation method (such as the straight-line depreciation method, the double-declining balance method, etc.). The disassembly and recycling cost includes the device disassembly cost, the recovery value of recyclable parts, etc. The waste treatment cost is determined according to the treatment method.

[0104] Step S302: Construct an equipment full-life cycle cost model, and the specific formula is:

[0105]

[0106] Among them, Lcc represents the equipment life cycle cost model, that is, the cost of the equipment during its entire life cycle. T represents the expected service life of the equipment, measured in years or other appropriate time units. C1 represents the first cost, C 2,t 、C 3,t and C 4,t respectively represent the second cost, the third cost, and the fourth cost within the t time period. C5 represents the fifth cost.

[0107] In this embodiment, the equipment life cycle cost model comprehensively considers the cost factors at each stage of the equipment during its entire life cycle. Based on the health status and failure probability of the equipment, the corresponding costs are obtained. And through dynamic data collection and analysis, it can reflect the changes in costs in real time. For example, as the operating time of the equipment increases, the aging of the equipment may lead to a decrease in operating efficiency, an increase in energy consumption costs, and at the same time, the maintenance costs may also rise due to component wear; changes in the market environment (such as electricity price adjustments, spare part price fluctuations, etc.) will also affect the life cycle cost of the equipment. Through this model, enterprises can more accurately grasp the equipment costs and formulate reasonable equipment management strategies, such as optimizing equipment procurement decisions, reasonably arranging equipment maintenance plans, improving equipment operating efficiency, and reducing failure risks.

[0108] Step S4: According to the equipment health status assessment and failure prediction results and the operating costs within the equipment's entire life cycle, use a multi-objective optimization algorithm to dynamically optimize the life cycle cost of the equipment.

[0109] The specific steps of Step S4 are as follows:

[0110] Step S401: Transform the optimization problem of the equipment life cycle cost into a multi-objective optimization model, and construct the objective function of the multi-objective optimization model, which is the weighted sum of cost items;

[0111] Step S402: Set the constraint conditions of the multi-objective optimization model, including: equipment health status constraint, failure prediction constraint, and budget constraint;

[0112] In this embodiment, the equipment health status constraint is specifically: the health status of the equipment needs to be maintained within a reasonable range to avoid the risk of excessive wear or exceeding the acceptable failure rate; the failure prediction constraint is specifically: based on the probability of the equipment failure output by the failure prediction model, ensure that the equipment will not fail within the expected life cycle; the budget constraint is specifically: the total cost within the equipment's entire life cycle should not exceed the preset budget;

[0113] Step S403: Use the multi-objective particle swarm optimization algorithm to solve the multi-objective optimization model. Through iterative update, find multiple optimal solutions of the multi-objective optimization model, that is, different life cycle cost objectives and strategies of the equipment;

[0114] In this embodiment, the optimization objectives and strategies need to be determined first. The optimization strategies for each objective are as follows: 1) Reduce the acquisition cost: Reduce the initial investment by selecting a suitable equipment supplier, optimizing the equipment configuration, etc.; 2) Reduce the operation cost. According to the operation status, environmental conditions, and operation strategy of the equipment, adjust the energy efficiency strategy of the equipment to reduce energy consumption; 3) Optimize the maintenance cost. Combine the health status assessment and the results of fault prediction to dynamically adjust the maintenance plan to avoid over-maintenance or late maintenance; 4) Reduce the cost of fault losses. By predicting faults and performing maintenance in advance, reduce the downtime caused by equipment faults and ensure production continuity; 5) Optimize the health management cost. Through a reasonable health monitoring and early warning mechanism, reduce ineffective health management expenditures. The first half is the objective, and the second half is the strategy. For example, in the first item: reducing the acquisition cost is the objective, and reducing the initial investment by selecting a suitable equipment supplier, optimizing the equipment configuration, etc. is the strategy;

[0115] In the multi-objective particle swarm optimization algorithm, each particle represents a possible solution, that is, a set of equipment life cycle cost optimization strategies. The particles find the optimal solution in the objective space through swarm search. The specific formula for the update process of each particle is as follows:

[0116] v p (t + 1) = ω × v p (t) + λ1 × rand1 × (P best -x p (t)) + λ2 × rand2 × (Q best -x p (t)),

[0117] x p (t + 1) = x p (t) + v p (t + 1),

[0118] where v p (t + 1) represents the velocity of particle p at time t + 1, v p (t) represents the velocity of particle p at time t, P best represents the best position in the particle's history, x p (t) represents the position of particle p at time t, x p (t + 1) represents the position of particle p at time t + 1, Q best represents the best solution in the swarm, ω represents the inertia weight, λ1 and λ2 represent the acceleration constants, and rand1 and rand2 represent random numbers.

[0119] The velocity here is expressed as (the velocity in the x direction, the velocity in the y direction), and corresponding addition and subtraction operations are performed with the position coordinates.

[0120] Step S404: Select the most suitable optimization solution according to actual requirements and budget.

[0121] Embodiment 2

[0122] Please refer to Figure 3 , another embodiment provided by the present invention: A dynamic management system for optimizing the full - life - cycle cost of equipment, including: a data acquisition module, an evaluation and prediction module, an operating cost module, and a dynamic management module;

[0123] The data acquisition module is used to collect equipment operation data in real time, perform pre - processing, generate first acquisition data, collect equipment cost data in real time, and perform pre - processing to generate second acquisition data;

[0124] The evaluation and prediction module is used to evaluate the health status of the equipment and predict faults based on the first acquisition data;

[0125] The operating cost module is used to construct a full - life - cycle cost model of the equipment based on the second acquisition data and the results of equipment health status evaluation and fault prediction, and obtain the operating cost within the full life cycle of the equipment;

[0126] The dynamic management module is used to dynamically optimize the life - cycle cost of the equipment according to the results of equipment health status evaluation and fault prediction and the operating cost within the full life cycle of the equipment, using a multi - objective optimization algorithm.

[0127] The evaluation and prediction module includes: a feature extraction unit, a health status evaluation unit, and a fault prediction unit;

[0128] The feature extraction unit is used to extract the features of the first acquisition data;

[0129] The health status evaluation unit is used to establish a multi - level time - series health index evaluation model, input a multi - dimensional feature vector, and evaluate the health status of the equipment;

[0130] The fault prediction unit is used to establish a fault prediction model, and predict the fault type and occurrence time of the equipment within a preset future time according to the multi - dimensional feature vector and historical data.

[0131] In addition, parts of the above - mentioned technical solutions provided in the embodiments of the present application that are the same as the corresponding technical solutions in the prior art in terms of implementation principles are not described in detail to avoid excessive elaboration.

[0132] The specific embodiments described above further elaborate in detail the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A dynamic management method for optimizing the total life cycle cost of a device, characterized in that, Including: Collecting device operation data in real time, preprocessing it to generate first collected data, collecting device cost data in real time, and preprocessing it to generate second collected data; Evaluating the health status and predicting faults of the device based on the first collected data; Constructing a full life cycle cost model of the device based on the second collected data and the results of device health status evaluation and fault prediction, and obtaining the operation cost within the full life cycle of the device; Dynamically optimizing the life cycle cost of the device according to the results of device health status evaluation and fault prediction and the operation cost within the full life cycle of the device by using a multi-objective optimization algorithm.

2. The dynamic management method for optimizing the full life cycle cost of a device according to claim 1, wherein The evaluating the health status and predicting faults of the device based on the first collected data includes: Extracting the features of the first collected data, including: mean value, variance / standard deviation, frequency domain features, time domain features and statistical features, fusing and dimension-reducing the extracted features to obtain a multi-dimensional feature vector; Establishing a multi-level time series health index evaluation model, inputting the multi-dimensional feature vector, and evaluating the health status of the device; Establishing a fault prediction model, and predicting the fault type and occurrence time of the device within a preset future time according to the multi-dimensional feature vector and historical data.

3. The dynamic management method for optimizing the full life cycle cost of a device according to claim 2, characterized in that, The establishing a multi-level time series health index evaluation model, inputting the multi-dimensional feature vector, and evaluating the health status of the device includes: Dividing the device operation state into a first stage, a second stage and a third stage, and establishing an independent time series model for each stage; Converting the data of each stage in the first collected data into a frequency spectrum or time domain image, using the convolutional layer in the convolutional neural network to extract the time series features of the frequency spectrum or time domain image, establishing a dependency relationship model of the device operation state, and capturing the health evolution trend of the device at different time points; Converting the extracted time series features into a health status evaluation result through a fully connected layer.

4. The dynamic management method for optimizing the total life cycle cost of a device according to claim 3, characterized in that, The establishing a fault prediction model, and predicting the fault type and occurrence time of the device within a preset future time according to the multi-dimensional feature vector and historical data includes: Establishing a device component connection graph G, G=(V, E), where V represents the set of nodes and E represents the set of edges; Learning the non-linear relationship of device faults through a deep neural network, updating the states of the nodes in the device component connection graph G, and constructing a fault prediction model; Combining evolution and the life cycle of the device, and according to the device component connection graph G, predicting in advance the associated device components affected by abnormal device components, optimizing and training the fault prediction model, inputting the multi-dimensional feature vector into the optimized and trained fault prediction model, and obtaining the fault probability and fault occurrence time of the device.

5. The dynamic management method for optimizing the full life cycle cost of a device according to claim 4, characterized in that, The combining evolution and the life cycle of the device, and according to the device component connection graph G, predicting in advance the associated device components affected by abnormal device components, optimizing and training the fault prediction model includes: Encoding the fault prediction model parameters into a gene sequence, and each individual represents a fault prediction model configuration; Selecting individuals for crossover operation according to the prediction error of the fault prediction model to generate new fault prediction model configurations; Introduce new fault prediction model configurations through mutation operations, repeatedly generate new fault prediction model configurations until the optimal fault prediction model parameters are found, and obtain an optimized fault prediction model; Combine the life cycle of the device, and based on the device component connection graph G, predict in advance the associated device components affected by abnormal device components, and train the optimized fault prediction model.

6. The dynamic management method for optimizing the total life cycle cost of a device according to claim 1, characterized in that The construction of the device life cycle cost model includes: Based on the second collection of data and the results of device health status assessment and fault prediction, analyze the cost composition of the device, including: the first cost, the second cost, the third cost, the fourth cost, and the fifth cost; Construct a device life cycle cost model, which is the sum of the total costs of the first cost, the fifth cost, and the second cost, the third cost, and the fourth cost within the expected service life.

7. The dynamic management method for optimizing the full life cycle cost of a device according to claim 1, characterized in that, According to the results of device health status assessment and fault prediction and the operating costs during the device life cycle, use a multi-objective optimization algorithm to dynamically optimize the life cycle cost of the device, including: Convert the optimization problem of the device life cycle cost into a multi-objective optimization model, and construct the objective function of the multi-objective optimization model, which is the weighted sum of the cost items; Set the constraint conditions of the multi-objective optimization model, including: device health status constraint, fault prediction constraint, and budget constraint; Use the multi-objective particle swarm optimization algorithm to solve the multi-objective optimization model, and through iterative updates, find multiple optimal solutions of the multi-objective optimization model, that is, the different life cycle cost objectives and strategies of the device; Select an optimization plan according to actual needs and budget.

8. A dynamic management method for optimizing the full life cycle cost of a device according to claim 1, characterized in that, The device operation data includes: device physical data, working condition data, and environmental data; The preprocessing includes: data denoising, removing noise data, outliers, and missing values in the first collection of data and the second collection of data; Data standardization is used to convert the first collection of data and the second collection of data into a unified format.

9. A dynamic management system for optimizing the full life cycle cost of a device, which is used to implement a dynamic management method for optimizing the full life cycle cost of a device according to any one of claims 1-8, characterized in that, Including: Data acquisition module, evaluation and prediction module, operating cost module, and dynamic management module; The data acquisition module is used to collect device operation data in real time and perform preprocessing to generate the first collection of data, collect device cost data in real time and perform preprocessing to generate the second collection of data; The evaluation and prediction module is used to evaluate the health status and predict faults of the device based on the first collection of data; The operating cost module is used to construct a device life cycle cost model based on the second collection of data and the results of device health status assessment and fault prediction, and obtain the operating costs during the device life cycle; The dynamic management module is used to dynamically optimize the life cycle cost of the device according to the results of device health status assessment and fault prediction and the operating costs during the device life cycle, using a multi-objective optimization algorithm.

10. A dynamic management system for optimizing the full life cycle cost of a device according to claim 9, characterized in that, The evaluation and prediction module includes: a feature extraction unit, a health status evaluation unit, and a fault prediction unit; The feature extraction unit is used to extract the features of the first collection of data; The health status evaluation unit is used to establish a multi-level time series health index evaluation model, input a multi-dimensional feature vector, and evaluate the health status of the device; The fault prediction unit is used to establish a fault prediction model, and predict the fault type and occurrence time of the device within a preset future time according to the multi-dimensional feature vector and historical data.

Images