Motorcade maintenance human resource allocation method integrated with health monitoring and prediction information
By integrating PHM system data and multi-objective optimization models, the maintenance human resource needs of airlines are dynamically assessed, solving the problem of rigid resource allocation in existing technologies. This enables timely and sufficient completion of maintenance tasks and cost reduction, thereby improving the operational economy and safety of airlines.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-15
- Publication Date
- 2026-04-14
AI Technical Summary
Existing technologies are difficult to integrate into PHM systems to optimize the allocation of maintenance human resources for airline fleets, resulting in rigid resource allocation and an inability to flexibly adapt to priority adjustments and resource bottlenecks in predictive maintenance, thus increasing the uncertainty and cost of maintenance tasks.
By integrating real-time monitoring data, historical maintenance records, and maintenance site supply capabilities from the PHM system, and combining them with a multi-objective optimization model, we can dynamically assess and optimize human resource needs, ensuring the flexibility and efficiency of resource allocation. We use the time accumulation method to calculate work hour requirements, classify task complexity, quantify personnel requirements, and construct a multi-objective optimization model to maximize availability and minimize costs.
This enables precise assessment of maintenance needs, optimization of human resource supply and demand balance, reduction of resource waste and overtime costs, improvement of operational economy, and timely and complete completion of maintenance tasks, while ensuring flight safety and reducing airworthiness and safety risks.
Smart Images
Figure SMS_15 
Figure SMS_93 
Figure QLYQS_5
Abstract
Description
Technical Field
[0001] This invention belongs to the technical field of aviation maintenance resource management, specifically relating to a method for allocating fleet maintenance human resources by integrating health monitoring and predictive information. Background Technology
[0002] Aircraft maintenance costs account for 15% to 25% of overall operating costs, and the maintenance costs incurred by airlines throughout the entire lifespan of an aircraft can exceed its ownership cost. Compared to foreign airlines, domestic airlines incur two to three times higher overall aircraft maintenance costs, including human resource costs, indicating that domestic airlines have significant room for improvement in controlling aircraft maintenance costs. Given the relatively small number of aircraft in an airline's fleet, low flight frequency, and a surplus of technical personnel in aircraft maintenance companies, airlines rely on historical maintenance data and experience to make judgments. Planners, based on their familiarity with maintenance operations and understanding of the manpower situation of frontline staff, can easily arrange various maintenance projects for each route. However, due to differences in maintenance experience and work methods among different planners, the work plans developed by different personnel vary considerably, resulting in a high degree of arbitrariness in route maintenance planning.
[0003] With the rapid growth of airline fleet types and aircraft numbers, increased flight density, and a continuous increase in the number and types of maintenance work, a shortage of various maintenance resources, such as maintenance technicians, maintenance downtime, and maintenance equipment, has emerged. Continuing to use traditional experience-based planning methods can no longer meet the changing maintenance environment. Arbitrary planning will miss many maintenance opportunities and waste a lot of maintenance resources. This often results in maintenance projects not being completed on time and within the deadline, creating significant pressure on airworthiness and safety.
[0004] In this context, based on an in-depth analysis of the two major factors in line maintenance production planning—maintenance man-hours and maintenance qualifications—optimizing human resource allocation is urgently needed to adapt to the ever-changing maintenance environment and reduce maintenance costs for maintenance companies. The rational allocation and scheduling of human resources helps reduce maintenance costs, ensure maintenance quality, and complete maintenance tasks on time, which is of great significance to improving the core competitiveness of airlines.
[0005] Considering that modern civil aircraft are designed with predictive and health management ( Prognostics and Health Management, PHM This system enables fault monitoring, location, and prediction, allowing for more precise predictive maintenance of aircraft. Predictive maintenance refers to... PHM Monitoring and forecasting results guide maintenance decisions, allowing for the advance planning of appropriate maintenance activities. This ensures flight safety while reducing overall maintenance costs and improving operational economics.
[0006] However, optimizing human resource allocation has become a core challenge in current maintenance resource management under the predictive maintenance model. Traditional experience-based methods are insufficient to address this issue. PHM The dynamic changes brought about by the real-time fault prediction and health monitoring data provided by the system, while capable of identifying potential faults in advance, increase the uncertainty and unpredictability of maintenance tasks. Existing configuration schemes often neglect the deep integration of predictive information with maintenance man-hours and qualification requirements, resulting in rigid allocation of human resources and an inability to flexibly adapt to priority adjustments and resource bottlenecks in predictive maintenance.
[0007] In summary, there is a lack of methods in the existing technology to integrate... PHM Information technology enables the allocation of human resources for civil aviation fleet maintenance, thereby improving the efficiency of human resource allocation while ensuring flight safety. Summary of the Invention
[0008] Technical Solution: To address the aforementioned shortcomings, this invention provides a method for allocating fleet maintenance human resources by integrating health monitoring and predictive information. Within the planning cycle, this method integrates... PHM The system monitors real-time data, historical maintenance records, and maintenance site supply capacity to accurately assess maintenance human resource needs. Based on this, and guided by the goals of availability and cost-effectiveness, it optimizes the balance between human resource supply and demand. To achieve the above objectives, this invention adopts the following technical solution: A method for allocating fleet maintenance human resources that integrates health monitoring and predictive information, comprising the following steps: S1. Within the planning period T, based on the current input of planned maintenance task requirements, calculate the daily maintenance man-hours and total maintenance man-hours within the planning period using the time accumulation method; divide the maintenance time into multiple time parts according to the activity composition of the maintenance task, quantify each time part based on task complexity, historical maintenance records and expert evaluation, calculate the daily man-hour requirements by accumulating all task man-hours, and obtain the total man-hour requirements for the planning period. S2. Obtain the current time in the integrated PHM system. Fault warning information f Identify the time of failure, i.e., the remaining lifespan of the component. RUL The corresponding predicted maintenance task execution time interval is [ , +RUL ; S3. Based on the outputs of S1 and S2, and combined with the skill and qualification requirements of maintenance tasks, calculate the daily human resource demand during the planning period and predict the dynamic human resource demand required for the execution of maintenance tasks; then, classify maintenance tasks into different categories, set as K categories, according to the required personnel professions and specific qualification levels, and quantify the human resource demand of each category by mapping the task skill requirements with the maintenance personnel qualification matrix. S4. Based on the daily maintenance man-hour requirements for various personnel given in S3, and considering the daily personnel configuration, and taking into account that insufficient human resources configuration will result in excessively high overtime or dispatch costs, calculate the supply-demand ratio of maintenance personnel, the guarantee rate, and the expected human resources cost within the planning period. S5. Construct a multi-objective optimization model with the objectives of maximizing maintenance support rate and minimizing expected human resource cost, and set constraints. S6. Solve the multi-objective optimization model of S5 and output the optimal human resource allocation plan within the planning period.
[0009] As an improvement, the time component includes preparation time, fault isolation time, repair time, and completion time; the repair time includes disassembly, replacement, reassembly, and debugging operations.
[0010] As an improvement, the statistical formulas for the daily maintenance man-hours and total maintenance man-hours required during the S1 planning period are as follows: (1) (2) In the formula: Maintenance man-hour requirements on day d within the planning period T. ; Total maintenance man-hours required within the planning period T; Maintenance tasks for day d; : No. d Heavenly i The preparation time for each task is quantified based on task complexity and takes a fixed value. : No. d The fault isolation time for the i-th task on day i is quantified based on historical maintenance records and taken as a fixed value; : No. d The repair time for the i-th task is quantified based on expert evaluation and taken as a fixed value. : No. d Heavenly i The completion time of each task is quantified based on historical maintenance records and a fixed value is taken.
[0011] As an improvement, S2 also includes predicting the execution of maintenance tasks from the day of the warning to the day the fault may occur, based on actual maintenance scheduling requirements. The task execution date is expressed as: In the formula: : Fault warning information f corresponds to the predicted maintenance task execution date; int(): Integer function.
[0012] As an improvement, the specific calculation process in S3 includes: (3) (4) (5) (6) In the above formula: : No. d Heavenly k Maintenance personnel working hours requirements S1 calculates the first d Daily maintenance labor demand ; The first in the historical maintenance data statistics k The proportion of demand for maintenance personnel ; Predict the man-hours required for maintenance tasks; When performing predictive maintenance tasks, the first k Maintenance personnel's working hours requirements ; : No. d Heavenly k Total demand for maintenance personnel's working hours , ; Total maintenance man-hours required within the planning period T.
[0013] As an improvement, the specific calculation process in S4 includes: (7) (8) (9) (10) (11) (12) formula: : No. d Heavenly k The supply and demand ratio of maintenance personnel : No. d Heavenly k The availability of maintenance personnel is provided in the following quantities: : No. d Heavenly k The number of maintenance personnel required is [number missing]. : No. d Heavenly k Daily planned maintenance hours for maintenance personnel : No. d Heavenly k The configuration of maintenance personnel meets the demand indicator variable, where 1 indicates completion and 0 indicates incompleteness. ; : No. d Heavenly k The required number of maintenance personnel is insufficient to cover the required maintenance man-hours. Maintenance personnel availability rate within the planning period T; TC Expected human resource costs within the planning period T; : No. k The unit labor cost of maintenance personnel; The unit labor cost incurred due to overtime or scheduling when human resources are insufficient.
[0014] As an improvement, the objective function in S5 is set as follows: (13) The target weight coefficient; Maintenance personnel availability rate within the planning period T; TC Expected human resource costs within the planning period T.
[0015] As an improvement, the constraints in S5 include skill adaptation constraints and fatigue balance constraints; among them, the skill adaptation constraint requires that the assigned tasks must strictly match the maintenance personnel's profession and specific qualification level, and is set as follows: (14) The fatigue balance constraint is based on historical work records and real-time monitoring to ensure that employee fatigue, measured by the number of consecutive working days, does not exceed a threshold, and is set as follows: (15) in, Indicates task i The decision variable for assigning to person j, either 0 or 1. For personnel j Qualification categories, For the task i Qualification requirements For the first d The day's maintenance tasks; For personnel j No. d Actual working hours per day max_workdays This is the maximum allowed number of working days within the cycle, a fixed value. This represents the upper limit of fatigue, a fixed constant.
[0016] As an improvement, the specific steps for solving the optimization model in S6 include: (6.1) Initialize model parameters Set target weight coefficients α、β According to airline priorities, safety takes precedence. α If cost is a priority, then β should be the largest value; determine the constraint parameters: max_workdays Maximum number of working days allowed within the cycle Fatigue limit; Input historical statistical parameters: the proportion of demand for various maintenance personnel Unit labor cost Overtime / scheduling costs ; (6.2) Prepare input data Collect dynamic data within the planning period T: S3 obtains the total daily demand for various personnel work hours. Number of suppliers Daily planned working hours Organize qualification information: Task i Qualification requirements ,personnel j Qualification Category Obtain PHM output: Predict the execution date of maintenance tasks. Required working hours ; (6.3) Selecting a solution algorithm For the multi-objective optimization model constructed for S5, the solver is determined according to the length of the planning period and the data scale. A mixed-integer linear programming solver is adopted, and an approximate optimal solution is obtained through a genetic algorithm. (6.4) Embedded Constraints Set constraints such as skill adaptation, fatigue balance, and high-priority task constraints; (6.5) Running the solution and iterating Input the objective function and constraints, and iterate until convergence; when resource conflicts are encountered, trigger the equalization strategy, including shifting non-critical tasks, calling cross-site resources, and reducing the weight of low-priority tasks; (6.6) Output and Verification Generate a daily human resource allocation matrix, including task assignment, timestamps, and utilization rate; verify the guarantee rate. and cost If the standard is not met, adjust the parameters and recalculate. (6.7) Dynamic Scrolling Optimization Repeat the above steps to update the configuration at regular intervals or when the PHM data deviation exceeds the tolerance, in order to adapt to changes in real-time fault prediction.
[0017] Beneficial Effects: This invention proposes a fleet maintenance human resource allocation method that integrates health monitoring and predictive information. Within the planning cycle, by integrating real-time monitoring data from the PHM system, historical maintenance records, and maintenance site supply capacity, it achieves accurate assessment of maintenance human resource needs. Based on this, and guided by the goals of availability and economy, it optimizes the balance between human resource supply and demand. Furthermore, compared with existing methods, this invention has the following advantages: 1. By dynamically optimizing human resource allocation, resource waste and overtime scheduling costs caused by arbitrary planning can be reduced, directly lowering aircraft maintenance labor costs and improving operational economics.
[0018] 2. By integrating real-time health monitoring data and predictive information, we can accurately assess maintenance needs, optimize the balance between manpower supply and demand, ensure that maintenance tasks are completed on time and in full, and reduce airworthiness and safety risks.
[0019] 3. Dynamically adapt to predictive maintenance priorities and unexpected tasks, and achieve efficient resource allocation through a multi-objective optimization model of availability and cost, avoiding the rigidity problem of traditional methods. Detailed Implementation
[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below, so that those skilled in the art can better understand the advantages and features of the present invention, thereby making a clearer definition of the scope of protection of the present invention. The embodiments described in this invention are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Example 1
[0021] 1. Task Scenario A medium-sized airline has a fleet of 50 B737-800 narrow-body passenger aircraft, with maintenance sites covering three hub airports: Beijing, Shanghai, and Guangzhou. The planned maintenance cycle is T=7 days. On the 3rd day, the PHM system issues a real-time warning: the high-pressure turbine blades of one aircraft operating the Beijing-Guangzhou route have a remaining life (RUL) of 48 hours (replacement must be completed within 2 days, otherwise an AOG grounding will be triggered); at the same time, there are 10 aircraft scheduled for routine Category A maintenance (2 aircraft per day); in addition, two engine maintenance personnel at the Shanghai site are suddenly absent, resulting in a shortage of 14 engine-related personnel at that site.
[0022] 2. Sample Data (1) S1 work hour statistics: Daily A-class maintenance work hour requirements for planned tasks =120 person-hours (2 aircraft × 60 person-hours / aircraft, each aircraft includes 30 person-hours for fuselage inspection, 20 person-hours for electronic testing, and 10 person-hours for routine engine inspection); predicted mission engine blade replacement man-hours =80 person-hours (20 person-hours for dismantling + 30 person-hours for replacement + 30 person-hours for debugging).
[0023] (2) S2 prediction task execution: d (days 3-4) = 1, RUL = 48 hours.
[0024] (3) S3 personnel requirements: K = 3 types of personnel (engine / airframe / electronics), demand percentage =45%, =30% =25%; Total man-hour requirement for Day 3 = Planned 120 man-hours + Forecasted 80 man-hours = 200 man-hours, Engine-related requirements =200 × 45% = 90 people, fuselage type ==200 × 30% = 60 people, Electronics =200 × 25% = 50 person-hours.
[0025] (4) S4 supply and demand parameters: Beijing station engine supply =70 person-hours (10 people × 7 person-hours / person), Shanghai station has a shortage of 14 person-hours for engine-related personnel, Guangzhou station can allocate 5 person-hours (10 person-hours) for engine-related personnel; unit labor cost =500 yuan / person-hour, overtime cost =800 yuan / person / hour, cross-site dispatch cost =200 yuan per person per hour.
[0026] (5) S5 target weights: α=0.7 (safety first), β=0.3; constraint parameter max_workdays=5, =0.8.
[0027] 3. Data Results (1) Before optimization: Supply and demand ratio of engines at Beijing station on day 3 =70 / 90≈0.778 (when there is a shortage of 20 people), the Shanghai station has a shortage of 14 people, and the total shortage is 34 people; the original supply cost = 70×500 = 35,000 yuan, the overtime cost = 34×800 = 27,200 yuan, the total cost on the 3rd day = 35,000 + 27,200 = 62,200 yuan; the cycle guarantee rate R_T = 85%, the objective function value = 0.7×0.85 -0.3×(120000 / 10000) = -3.005; (2) After optimization: 5 engine personnel from Guangzhou cross-site (cost 10 person-hours, 10 × (500 + 200) = 7000 yuan) + 4 person-hours of overtime work from Shanghai site (cost 4 × 800 = 3200 yuan) + adjustment of 1 aircraft for Class A maintenance to day 5 (reducing engine demand by 9 person-hours); After optimization, the supply from Beijing site = 70 + 10 = 80 person-hours, and the overtime work from Shanghai is 4 person-hours, totaling 90 person-hours of demand; Total cost TC = 95,000 yuan (25,000 yuan less than the traditional method of 120,000 yuan), and the guarantee rate =98%, objective function value = 0.7×0.98-0.3×(95000 / 10000) = -2.164 (increase of 0.841); (3) Dynamic rolling: On the 4th day, the PHM system updated the engine blade remaining life RUL=36 hours (originally 48 hours), triggering event-level optimization: the replacement task was brought forward from the original 4th day to the afternoon of the 3rd day, the emergency response time was shortened from 120 minutes to 80 minutes (shortened by 33%), avoiding AOG downtime losses of about 500,000 yuan; 4. Effect Comparison Table 1 Comparison of Effects
[0028] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.
Claims
1. A method for allocating fleet maintenance human resources by integrating health monitoring and predictive information, characterized in that, The specific steps include: S1. Within the planning period T, based on the current input of planned maintenance task requirements, calculate the daily maintenance man-hours and total maintenance man-hours within the planning period using the time accumulation method; divide the maintenance time into multiple time parts according to the activity composition of the maintenance task, quantify each time part based on task complexity, historical maintenance records and expert evaluation, calculate the daily man-hour requirements by accumulating all task man-hours, and obtain the total man-hour requirements for the planning period. S2. Obtain the current time in the integrated PHM system. Fault warning information f Identify the time of failure, i.e., the remaining lifespan of the component. RUL The corresponding predicted maintenance task execution time interval is [ , +RUL ; S3. Based on the outputs of S1 and S2, and combined with the skill and qualification requirements of maintenance tasks, calculate the daily human resource demand during the planning period and predict the dynamic human resource demand required for the execution of maintenance tasks; then, classify maintenance tasks into different categories, set as K categories, according to the required personnel professions and specific qualification levels, and quantify the human resource demand of each category by mapping the task skill requirements with the maintenance personnel qualification matrix. S4. Based on the daily maintenance man-hour requirements for various personnel given in S3, and considering the daily personnel configuration, and taking into account that insufficient human resources configuration will result in excessively high overtime or dispatch costs, calculate the supply-demand ratio of maintenance personnel, the guarantee rate, and the expected human resources cost within the planning period. S5. Construct a multi-objective optimization model with the objectives of maximizing maintenance support rate and minimizing expected human resource cost, and set constraints. S6. Solve the multi-objective optimization model of S5 and output the optimal human resource allocation plan within the planning period.
2. The fleet maintenance human resource allocation method incorporating health monitoring and prediction information according to claim 1, characterized in that, The time component includes preparation time, fault isolation time, repair time, and completion time; the repair time includes disassembly, replacement, reassembly, and debugging operations.
3. The fleet maintenance human resource allocation method incorporating health monitoring and prediction information according to claim 1 or 2, characterized in that, The statistical formulas for the daily maintenance man-hours and total maintenance man-hours required during the S1 planning period are as follows: (1) (2) In the formula: The first [number]th ... d Daily maintenance labor demand ; Total maintenance man-hours required within the planning period T; ： The maintenance task for day d; : No. d Heavenly i The preparation time for each task is quantified based on task complexity and takes a fixed value. : No. d The fault isolation time for the i-th task on day i is quantified based on historical maintenance records and taken as a fixed value; : No. d Heavenly i The repair time for each task is quantified based on expert evaluation and taken as a fixed value; : No. d Heavenly i The completion time of each task is quantified based on historical maintenance records and a fixed value is taken.
4. The fleet maintenance human resource allocation method incorporating health monitoring and prediction information according to claim 1, characterized in that, S2 also includes predicting the execution of maintenance tasks from the day of the warning to the day the fault may occur, based on actual maintenance scheduling requirements. The task execution date is expressed as follows: In the formula: The fault warning information f corresponds to the predicted date of the maintenance task execution; int() : Integer function.
5. The fleet maintenance human resource allocation method incorporating health monitoring and predictive information according to claim 3, characterized in that, The specific calculation process in S3 includes: (3) (4) (5) (6) In the above formula: : No. d Heavenly k Maintenance personnel working hours requirements , ; S1 calculates the first d Daily maintenance labor demand ; The first in the historical maintenance data statistics k The proportion of demand for maintenance personnel ; Predict the man-hours required for maintenance tasks; When performing predictive maintenance tasks, the first k Maintenance personnel's working hours requirements ; : No. d Heavenly k Total demand for maintenance personnel's working hours , ; Total maintenance man-hours required within the planning period T.
6. The fleet maintenance human resource allocation method incorporating health monitoring and prediction information according to claim 1, characterized in that, The specific calculation process in S4 includes: (7) (8) (9) (10) (11) (12) formula: : No. d Heavenly k The supply and demand ratio of maintenance personnel , No. d Heavenly k The availability of maintenance personnel is provided in the following quantities: No. d Heavenly k The number of maintenance personnel required is [number missing]. : No. d Heavenly k Daily planned maintenance hours for maintenance personnel No. d The configuration of maintenance personnel for category k in this case meets the demand indicator variable, where 1 indicates completion and 0 indicates incompleteness. : No. d Heavenly k The required number of maintenance personnel is insufficient to cover the required maintenance man-hours. Maintenance personnel availability rate within the planning period T; TC Expected human resource costs within the planning period T; : No. k The unit labor cost of maintenance personnel; The unit labor cost incurred due to overtime or scheduling when human resources are insufficient.
7. The fleet maintenance human resource allocation method incorporating health monitoring and prediction information according to claim 1, characterized in that, The objective function in S5 is set as follows: (13) The target weight coefficient.
8. The fleet maintenance human resource allocation method incorporating health monitoring and predictive information according to claim 1 or 7, characterized in that, The constraints in S5 include skill adaptation constraints and fatigue balance constraints; the skill adaptation constraint requires that assigned tasks must strictly match the maintenance personnel's profession and specific qualification level, and is set as follows: (14) The fatigue balance constraint is based on historical work records and real-time monitoring to ensure that employee fatigue, measured by the number of consecutive working days, does not exceed a threshold, and is set as follows: （15） in, This represents the decision variable for assigning task i to person j, and can be either 0 or 1. For personnel j Qualification categories, For the task i Qualification requirements For the first d The day's maintenance tasks; For personnel j No. d Actual working hours per day max_workdays This is the maximum allowed number of working days within the cycle, a fixed value. This represents the upper limit of fatigue, a fixed constant.
9. The fleet maintenance human resource allocation method incorporating health monitoring and predictive information according to claim 1, characterized in that, The specific steps for solving the optimization model in S6 include: (6.1) Initialize model parameters Set target weight coefficients α and β: based on airline priorities, α is maximized for safety priority and β is maximized for cost priority; determine constraint parameters: max_workdays is the maximum allowed number of working days within the period. Fatigue limit; Input historical statistical parameters: the proportion of demand for various maintenance personnel Unit labor cost Overtime / scheduling costs; (6.2) Prepare input data Collect dynamic data within the planning period T: S3 obtains the total daily demand for various personnel work hours. Number of suppliers Daily planned working hours Organize qualification information: Task i Qualification requirements ,personnel j Qualification Category Obtain PHM output: Predict the execution date of maintenance tasks. Required working hours ; (6.3) Selecting a solution algorithm For the multi-objective optimization model constructed for S5, the solver is determined according to the length of the planning period and the data scale. A mixed-integer linear programming solver is adopted, and an approximate optimal solution is obtained through a genetic algorithm. (6.4) Embedded Constraints Set constraints such as skill adaptation, fatigue balance, and high-priority task constraints; (6.5) Running the solution and iterating Input the objective function and constraints, and iterate until convergence; when resource conflicts are encountered, trigger the equalization strategy, including shifting non-critical tasks, calling cross-site resources, and reducing the weight of low-priority tasks; (6.6) Output and Verification Generate a daily human resource allocation matrix, including task assignment, timestamps, and utilization rate; verify the guarantee rate. and cost If the standard is not met, adjust the parameters and recalculate. (6.7) Dynamic Scrolling Optimization Repeat the above steps to update the configuration at regular intervals or when the PHM data deviation exceeds the tolerance, in order to adapt to changes in real-time fault prediction.