Mine equipment intelligent operation and maintenance management and prediction analysis system based on internet of things

Abstract

The present application relates to the technical field of intelligent operation and maintenance and management of mine equipment cluster, in particular to a mine equipment intelligent operation and maintenance management and prediction analysis system based on Internet of Things, comprising: a data fusion module for collecting equipment working condition data, positioning and speed data, operation scheduling data and historical maintenance data of heterogeneous equipment, and aligning them into unified space-time data flow; a health quantification module for determining real-time equipment health index of single equipment; a resilience evaluation module for calculating actual efficiency of the cluster and determining cluster task flexibility index; a state classification module for judging cluster resilience state and triggering reconstruction decision; a dynamic reconstruction module for generating degradation-compensation optimal instruction set and issuing it to heterogeneous equipment. The system overcomes the limitations of traditional single threshold alarm, can more accurately and smoothly reflect the progressive attenuation process of equipment from health to failure, and provides high-quality quantitative input for subsequent cluster resilience evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent operation and maintenance and management of mine equipment clusters, in particular to a mine equipment intelligent operation and maintenance management and prediction analysis system based on the Internet of Things. BACKGROUND

[0002] In large-scale mine operations, equipment clusters work in a strongly coupled manner. In order to ensure production continuity, enterprises need to perform real-time operation and maintenance management on heterogeneous devices such as excavators and trucks. These devices are closely related, and the progressive degradation of the performance of a single device will have a complex nonlinear impact on the overall efficiency of the entire cluster through coupling relationships such as queuing effects. Traditional operation and maintenance management methods rely mainly on alarm mechanisms based on single-device threshold values or post-fault maintenance, and lack fine quantitative means for device health status. These methods result in a serious lag in management intervention measures, often responding passively after the efficiency of the cluster has already decreased significantly, causing irreparable economic losses. Existing technologies lack a mechanism for deeply integrating real-time health status of devices with job flow models in scheduling decisions. In the face of high dynamics and complexity of mine operating environments, traditional static scheduling and operation and maintenance strategies lack foresight and adaptive adjustment capabilities.

[0003] Therefore, it is an urgent problem for those skilled in the art to provide an intelligent operation and maintenance management system that actively reconfigures before efficiency significantly decreases.

[0004] The above information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] To solve the above technical problems, the present application discloses a mine equipment intelligent operation and maintenance management and prediction analysis system based on the Internet of Things. Specifically, the technical solution of the present application comprises:

[0006] A data fusion module for collecting device operating condition data, positioning and speed data, job scheduling data, and historical maintenance data of heterogeneous devices, and aligning them into a unified space-time data stream;

[0007] A health quantification module for determining a real-time device health index of a single device based on real-time monitoring values in the unified space-time data stream and a preset health reference value;

[0008] A resilience evaluation module for calculating a cluster actual efficiency based on the real-time device health index, job scheduling data, and a preset cluster target efficiency, and determining a cluster task flexibility index;

[0009] a state grading module configured to determine a cluster resilience state according to the cluster task elasticity index and a preset resilience state threshold, and trigger a reconstruction decision when the cluster resilience state is a pre-warning state;

[0010] a dynamic reconstruction module configured to generate a degradation-compensation optimal instruction set and issue the same to the heterogeneous devices in response to the reconstruction decision, with a target of minimizing efficiency deviation.

[0011] Preferably, the health quantification module determines a real-time device health index of a single device, including:

[0012] collecting real-time monitoring values in a unified space-time data stream, and calling preset health benchmark values and weights;

[0013] calculating a dimensionless deviation degree of the real-time monitoring values relative to the health benchmark values;

[0014] determining the real-time device health index based on a weighted dimensionless deviation degree sum of all key performance indicators.

[0015] Preferably, the resilience evaluation module determines a cluster task elasticity index, including:

[0016] determining a shovel loading time based on the real-time device health index;

[0017] calculating a queuing waiting time based on the real-time device health index and a truck quantity in the job scheduling data;

[0018] calling positioning and speed data and the job scheduling data to obtain truck transportation time and unloading time;

[0019] calculating a truck cycle time in combination with the shovel loading time, the queuing waiting time, the truck transportation time and the unloading time;

[0020] determining a cluster actual efficiency according to the truck cycle time and preset average truck load and average truck distance;

[0021] obtaining the cluster task elasticity index by dividing the cluster actual efficiency by the cluster target efficiency by the real-time device health index.

[0022] Preferably, the state grading module determines a cluster resilience state, including:

[0023] determining the cluster resilience state as a stable state when the cluster task elasticity index is not less than a first preset threshold;

[0024] determining the cluster resilience state as a pre-warning state when the cluster task elasticity index is less than the first preset threshold and not less than a second preset threshold;

[0025] determining the cluster resilience state as an unstable state when the cluster task elasticity index is less than the second preset threshold.

[0026] Preferably, the dynamic reconfiguration module generates a degradation-compensation optimal instruction set, including:

[0027] Construct an optimization function with the objective of minimizing efficiency deviation, where efficiency deviation is the difference between the cluster target efficiency and the predicted reconstruction efficiency;

[0028] A heuristic optimization algorithm is used to search for a degradation-compensation instruction set with the fitness function of minimizing efficiency deviation.

[0029] Among them, the predicted reconstruction efficiency is the efficiency value obtained by simulating the current iteration's degradation-compensation instruction set into the job flow coupling model during the search process of the heuristic optimization algorithm.

[0030] The search-generated instruction set includes: degradation strategies for degraded devices and compensation strategies for healthy devices. The compensation strategies include collaborative reconstruction, path reconstruction, and rate adjustment.

[0031] The optimal instruction set found by the heuristic optimization algorithm is output as the optimal instruction set.

[0032] Preferably, the system also includes:

[0033] The performance prediction module is used to predict the future performance degradation curve of the device based on historical real-time device health index sequences and employs a pre-trained long short-term memory network model.

[0034] Preferably, the system also includes:

[0035] Closed-loop correction module, used for:

[0036] The decision delay from triggering the reconfiguration decision to issuing the optimal instruction set is calculated by the dynamic reconfiguration module.

[0037] The actual cluster efficiency after executing the optimal instruction set is collected;

[0038] Based on the real-time equipment health index, a job flow coupling model is used to simulate and calculate the simulation efficiency of the optimal instruction set not being executed.

[0039] The economic benefits are determined based on the actual cluster efficiency, the simulated efficiency, and the preset economic value per unit efficiency.

[0040] An optimization algorithm that automatically corrects resilience state thresholds or dynamically reconfigures modules based on decision delay and economic benefits.

[0041] Preferably, the closed-loop correction module automatically corrects, including:

[0042] If the decision delay continues to exceed the preset delay threshold, the search space of the optimization algorithm is simplified.

[0043] If the economic benefits remain below the preset benefit threshold, the resilience threshold will be increased.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. This system integrates multi-source data from heterogeneous devices to construct a unified spatiotemporal data stream. By calculating the weighted dimensionless deviation of key performance indicators, a standardized real-time device health index is determined. This method overcomes the limitations of traditional single-threshold alarms, and can more accurately and smoothly reflect the gradual decline process of devices from health to failure, providing high-quality quantitative input for subsequent cluster resilience assessment.

[0046] 2. This system constructs a complete decoupled model from individual device health to cluster efficiency and pioneers a cluster task resilience index. This index not only monitors whether efficiency declines but also assesses the cluster's response to the degradation of individual device health, i.e., the amplification or absorption effect. This reveals and quantifies the nonlinear strong coupling relationship between individual device health and overall cluster efficiency, providing a deeper and more forward-looking basis for decision-making.

[0047] 3. By dividing the system into three states—stable, warning, and unstable—it creates a warning state where efficiency has just begun to deviate but has not yet collapsed, thus realizing a shift from passive response to proactive prevention. In response to the warning, the dynamic reconfiguration module aims to minimize efficiency deviation. Through heuristic algorithms and simulations, it rapidly seeks optimization among massive combinations to find the degradation-compensation balance point that maximizes cluster efficiency, ensuring the scientific and optimal nature of the decision-making.

[0048] 4. This system integrates prediction and self-correction capabilities to form a closed-loop intelligence. The performance prediction module can foresee the future performance degradation curve of the equipment, enabling proactive decision-making. Simultaneously, the unique closed-loop correction module can adaptively and self-optimize the resilience state threshold and optimization algorithm based on decision delays and economic benefits, ensuring that the system maintains the optimal balance between fast response and high performance throughout long-term operation. Attached Figure Description

[0049] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0050] Figure 1 This is a system structure diagram of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0052] Example 1:

[0053] Please seeFigure 1 An IoT-based intelligent operation and maintenance management and predictive analysis system for mining equipment includes:

[0054] The data fusion module is used to collect equipment operating condition data, positioning and rate data, job scheduling data and historical maintenance data from heterogeneous devices, and align them into a unified spatiotemporal data stream.

[0055] The health quantification module is used to determine the real-time device health index of a single device based on real-time monitoring values ​​in a unified spatiotemporal data stream and preset health benchmark values.

[0056] The resilience assessment module is used to calculate the actual efficiency of the cluster and determine the cluster task elasticity index based on real-time equipment health index, job scheduling data and preset cluster target efficiency.

[0057] The status classification module is used to determine the cluster resilience status based on the cluster task elasticity index and the preset resilience status threshold, and to trigger a reconstruction decision when the cluster resilience status is in the warning state.

[0058] The dynamic reconfiguration module is used to respond to reconfiguration decisions, aiming to minimize efficiency deviations, generate the optimal set of degradation-compensation instructions and distribute them to heterogeneous devices.

[0059] This embodiment provides an IoT-based intelligent operation and maintenance management and predictive analysis system for mining equipment. The system aims to assess the nonlinear impact of the health status of a single device on the efficiency of a tightly coupled cluster by quantifying the health status of the device in real time, and to intervene early before the cluster efficiency drops significantly, thereby maintaining the overall operational resilience of the cluster through dynamic reconfiguration decisions.

[0060] In this embodiment, the system includes:

[0061] The data fusion module integrates multi-source and heterogeneous data from heterogeneous equipment within the mining area, providing a unified data foundation for subsequent health quantification and resilience assessment. In this embodiment, the specific function of this module is to collect: equipment operating condition data from vehicle ECUs, such as engine speed, oil pressure, and temperature; positioning and speed data from GPS / BeiDou systems; work scheduling data from dispatch systems, such as the excavator to which the truck belongs and shift plans; and historical maintenance data from the CMMS computerized maintenance management system. After collection, the module aligns the above multi-source data using a unified timestamp and equipment ID to construct a unified spatiotemporal data stream with a single device as the core and consistent in time sequence.

[0062] The health quantification module transforms complex equipment operating condition data into a standardized, quantifiable health indicator. Based on the unified spatiotemporal data stream output by the preceding data fusion module, this module obtains real-time monitoring values ​​of key performance indicators (KPIs). Simultaneously, this module calls pre-set health benchmark values, such as rated values ​​from the equipment manual or historical best operating condition values. Through specific algorithms, such as weighted deviation calculation, this module determines the real-time equipment health index of a single device.

[0063] Real-time device health index, also referred to in this invention , is a standardized, dimensionless index in the range [0,1], used to quantify the current health of device i at time t, where 1 represents the ideal health state, and a decrease in the index represents a deviation from the benchmark in performance;

[0064] The resilience assessment module is used to evaluate the real-time device health index. The performance degradation of a single device reflected is due to It reflects the nonlinear impact on the collaborative operation efficiency of the entire equipment cluster; this module first calculates the actual cluster efficiency based on real-time equipment health index and job scheduling data, such as the number of trucks currently in operation, through a job flow coupling model, such as a queuing theory model.

[0065] The actual efficiency of the cluster, also referred to in this invention, is... This refers to the overall efficiency of the cluster's actual output under the current equipment health status, such as ton-kilometer efficiency, which reflects the cluster's true output capacity.

[0066] This module incorporates preset cluster target efficiency, such as the rated efficiency required by the shift schedule. Based on the real-time device health index, the cluster task elasticity index is finally determined.

[0067] Cluster task elasticity index, also referred to in this invention Its core is to quantify the performance of the cluster on a single device, the denominator being... When it decreases, to maintain its overall efficiency, the molecule, i.e. Ability;

[0068] The state classification module assesses cluster stability in real time based on quantified FMEI values ​​and decides whether intervention measures are needed. This module categorizes cluster resilience into stable, warning, or unstable states based on real-time calculations of the cluster task resilience index and a set of preset resilience thresholds, such as a first threshold of 0.9 and a second threshold of 0.5. Intervention measures are only initiated when the cluster resilience state is determined to be in a warning state. This indicates that equipment degradation has begun to propagate through coupling relationships and affect cluster efficiency. At this point, the system will automatically trigger a reconfiguration decision.

[0069] The dynamic reconfiguration module responds to the reconfiguration decisions triggered by the status grading module and proactively adjusts the workflow to avoid a significant drop in efficiency. Once started, this module focuses on minimizing efficiency deviation as its core optimization objective, i.e., minimizing the gap between the predicted efficiency after reconfiguration and the target efficiency. This module uses heuristic algorithms to optimize and generate a set of degradation-compensation optimal instructions.

[0070] The optimal downgrade-compensation instruction set is a set that includes refactoring paths ( ), adjusted rate ( ) and restructured collaborative relationships ( The compound instruction is: ) Degradation refers to actively reducing the workload of degraded equipment, such as reducing the number of trucks assigned to it; Compensation refers to making up for lost efficiency by adjusting other healthy equipment, such as rearranging trucks or replanning routes.

[0071] This module uses IoT communication to send this optimal instruction set in a closed loop to relevant heterogeneous devices in the mining area, such as the on-board terminals of trucks and excavators, for execution.

[0072] This embodiment constructs a complete technical closed loop, from single device health quantification to cluster resilience nonlinear assessment, and then to proactive closed-loop reconstruction decision-making, through the collaborative work of the aforementioned data fusion, health quantification, resilience assessment, state classification, and dynamic reconstruction modules. It solves the problem in traditional operation and maintenance where it is impossible to quantify the nonlinear impact of single device failure on cluster efficiency. This allows the system to intervene in advance in the early warning state before a serious decline in cluster efficiency occurs, and proactively avoid bottlenecks through dynamic reconstruction, thereby greatly improving the overall operational resilience and comprehensive efficiency of mining equipment clusters in highly dynamic and strongly coupled operating environments.

[0073] Example 2:

[0074] The health quantification module determines the real-time health index of a single device, including:

[0075] Collect real-time monitoring values ​​from a unified spatiotemporal data stream and call preset health benchmark values ​​and weights;

[0076] Calculate the dimensionless deviation of real-time monitoring values ​​from health baseline values;

[0077] The real-time equipment health index is determined based on the weighted dimensionless deviation of all key performance indicators.

[0078] Based on Example 1, this embodiment specifies how the health quantification module determines the real-time device health index; the module aims to provide a standardized health quantification model that considers the importance of different indicators.

[0079] To perform quantization, this module collects data from a unified spatiotemporal data stream. The real-time monitoring values ​​of each key performance indicator (KPI) are expressed as follows: ;

[0080] Simultaneously, this module calls preset health benchmark values ​​that correspond one-to-one with these m KPIs, represented as... This can be represented by a CMMS or device manual, and the weights are expressed as follows: ;

[0081] This is achieved by analyzing historical failure datasets and statistically analyzing different KPI indicators, for example, using independent symbols. Indicates the final failure, for example, using The correlation coefficients, such as Pearson correlation coefficients or feature importance scores, are represented and obtained through normalization to enhance the model's sensitivity to critical failures and satisfy the following conditions: ;

[0082] This module calculates the dimensionless deviation of each KPI; in this embodiment, the deviation is specifically calculated through... This is used for calculation, reflecting the percentage deviation of the real-time value from the baseline value; to prevent the denominator from being zero, when In such cases, the deviation can be handled specially, for example, by directly using the absolute value. And normalize it, or define it as a preset maximum penalty value;

[0083] This module determines the final real-time device health index based on the weighted dimensionless deviation of all KPIs; in this embodiment, the calculation uses the following custom health status quantification model:

[0084] ;

[0085] in, The health index of device i at time t is the final calculated output of this module and is dimensionless.

[0086] The number of key performance indicators (KPIs) used for evaluation, which are pre-selected based on expert experience and historical failure data;

[0087] The weight of the j-th KPI;

[0088] : The real-time monitoring value of the j-th KPI, the ECU data stream in step 1.1;

[0089] The health baseline or rating for this KPI, CMMS, or equipment manual;

[0090] This formula establishes a standardized index within the range of [0,1]. By subtracting the weighted sum of the deviations of all KPIs from the ideal state 1 and taking the maximum value with 0, it ensures that the health index will not be negative. This invention realizes a method for uniformly quantifying the health status of heterogeneous devices of different manufacturers and models.

[0091] This embodiment provides a standardized, reproducible model that takes into account the importance (weight) of different performance indicators by introducing a weighted summation model of deviations. A health quantification method; compared to single threshold alarms, It can more accurately and smoothly reflect the gradual degradation process of equipment from health to failure, providing high-quality, time-consistent input data for subsequent resilience assessment and prediction.

[0092] Example 3:

[0093] The resilience assessment module determines the cluster task resilience index, including:

[0094] Determine the excavator loading time based on real-time equipment health index;

[0095] Calculate queuing time based on the real-time equipment health index and the number of trucks in the job scheduling data;

[0096] Call the location and speed data and job scheduling data to obtain the truck transportation time and unloading time;

[0097] Calculate the truck cycle time by combining the excavator loading time, queuing time, truck transportation time, and unloading time;

[0098] The actual efficiency of the cluster is determined based on the truck cycle time and the preset average truck load and average transport distance.

[0099] The cluster task elasticity index is obtained by dividing the ratio of the actual efficiency of the cluster to the target efficiency of the cluster by the real-time device health index.

[0100] This embodiment, based on Embodiment 2, further specifies in detail how the resilience assessment module determines the cluster task resilience index (FMEI); its core lies in constructing a job flow coupling model to deconstruct... How to transfer efficiency to the overall cluster through job flow coupling;

[0101] To perform the evaluation, this module constructs a workflow coupling model based on queuing theory, taking the excavator-truck work unit as an example, and analyzes the truck cycle time. The composition;

[0102] Determine the excavator loading time This module is based on the real-time device health index calculated by the previous module. Through a predetermined functional relationship The excavator loading time was calculated. The functional relationship Yes, by retrieving historical assignment datasets, a system containing multiple groups ( , ) data pairs, where For the equipment health index at a historical moment, The actual loading time corresponding to that moment was determined by fitting the data using regression analysis methods such as least squares; this step established the equipment health status. With the time spent on the task The first level of association;

[0103] Calculate queue waiting time This module further relies on real-time device health indices. Because it affected And the number of trucks in the current work flow obtained from the job scheduling data. The queuing time of the truck at excavator i can be calculated by applying queuing theory models, such as the M / M / 1 or M / G / c models. This step establishes health. Coupled with clustering effects, i.e., the second layer of nonlinear association in queuing;

[0104] Get other times ( and This module simultaneously accesses location and speed data as well as job scheduling data to directly obtain the truck's transport time. and uninstallation time ;

[0105] Calculate truck cycle time This module ultimately combines the four components—excavator loading time, queuing time, truck transportation time, and unloading time—to calculate the single truck cycle time for truck k. :

[0106] ;

[0107] In obtaining all Taiwan truck Then, the module calculates the average load per truck based on the truck cycle time and a preset average load per truck. and average transport distance Determine the actual efficiency of the cluster :

[0108] ;

[0109] This formula needs to be standardized to ton-kilometers per hour to account for the time efficiency of all trucks. The total tonne-kilometer efficiency of the cluster is summarized by this formula; Caused The changes were explicitly passed on to the overall cluster efficiency. superior;

[0110] After determining the index, the module evaluates the actual efficiency of the cluster. With cluster target efficiency The ratio, i.e., the normalized cluster efficiency, is divided by the real-time device health index, which serves as the normalized device performance index. Obtain the cluster task elasticity index :

[0111] ;

[0112] This formula represents the final derivation of FMEI; the elasticity of the quantization system: This means the system has compensatory properties; for example, scheduling automatically compensates for attenuation. This means that attenuation is amplified; for example, queuing effects can cause bottlenecks. To ensure the robustness of the calculation, a check is performed before executing the formula: if the health index in the denominator... If the value is below a very small threshold, such as 0.01, it indicates a serious equipment failure, and the system should directly classify the cluster resilience state as unstable, rather than calculating the FMEI value. Similarly, the system's evaluation and decision-making are based solely on the cluster's target efficiency. Operating under effective working conditions;

[0113] This embodiment constructs... A complete mathematical model was developed, revealing and quantifying the health of a single device for the first time. Overall cluster efficiency The nonlinear, strongly coupled relationship between them; as an innovative quantitative indicator, FMEI not only monitors... Whether it declines or not, more importantly, it is through its relationship with... In comparison, the extent to which the cluster amplifies or absorbs the decline was assessed, providing a more profound and forward-looking decision-making basis for subsequent state classification than simple monitoring.

[0114] Example 4:

[0115] The state grading module determines the cluster resilience status, including:

[0116] When the cluster task elasticity index is not less than the first preset threshold, the cluster resilience state is determined to be stable.

[0117] When the cluster task elasticity index is less than the first preset threshold and not less than the second preset threshold, the cluster resilience state is determined to be in a warning state.

[0118] When the cluster task elasticity index is less than the second preset threshold, the cluster resilience state is determined to be unstable.

[0119] This embodiment, based on embodiment 3, specifically illustrates how the state classification module determines the cluster resilience state based on the FMEI value and a preset resilience state threshold.

[0120] This module presets two key resilience state thresholds, namely the first preset threshold and the second preset threshold.

[0121] The first and second preset thresholds are derived from historical operation data statistics. In this embodiment, the first preset threshold is set to 0.9, which corresponds to the critical point where the cluster efficiency fluctuation does not exceed the standard deviation range caused by normal operating environment, such as weather and road conditions. The second preset threshold is set to 0.5, which corresponds to the critical point where the efficiency decline leads to the inability to complete the shift's planned output.

[0122] This module acquires the FMEI value output by the toughness assessment module in real time and executes the following judgment logic:

[0123] Steady state: When the cluster task elasticity index (FMEI) is not less than the first preset threshold, i.e. At this point, the system determines that the cluster resilience state is stable; in this state, the system can absorb performance fluctuations without intervention.

[0124] Warning state: When the cluster task elasticity index (FMEI) is less than the first preset threshold, i.e. And not less than the second preset threshold, i.e. At this time, the system determines the cluster resilience status to be in an early warning state; this state is a key signal that triggers subsequent dynamic reconfiguration.

[0125] Instability: When the cluster task elasticity index (FMEI) is less than the second preset threshold, i.e. At this point, the system determines that the cluster resilience state is unstable; this state indicates that the coupling effect has led to a severe decrease in efficiency, and an emergency reconfiguration must be performed immediately.

[0126] This embodiment achieves refined and hierarchical management of cluster risks by setting graded thresholds for FMEI (0.9 and 0.5). Unlike traditional post-fault repair or intervention after a significant drop in efficiency, the early warning state created by this invention identifies an intermediate state where efficiency has just begun to deviate but has not yet collapsed. This allows the system to trigger the dynamic reconstruction module at the optimal time, i.e., when FMEI first falls below 0.9, realizing a shift from passive response to proactive prevention and greatly improving the effectiveness and timeliness of intervention measures.

[0127] Example 5:

[0128] The dynamic reconfiguration module generates the optimal instruction set for degradation-compensation, including:

[0129] Construct an optimization function with the objective of minimizing efficiency deviation, where efficiency deviation is the difference between the cluster target efficiency and the predicted reconstruction efficiency;

[0130] A heuristic optimization algorithm is used to search for a degradation-compensation instruction set with the fitness function of minimizing efficiency deviation.

[0131] Among them, the predicted reconstruction efficiency is the efficiency value obtained by simulating the current iteration's degradation-compensation instruction set into the job flow coupling model during the search process of the heuristic optimization algorithm.

[0132] The search-generated instruction set includes: degradation strategies for degraded devices and compensation strategies for healthy devices. The compensation strategies include collaborative reconstruction, path reconstruction, and rate adjustment.

[0133] Output the optimal instruction set found by the heuristic optimization algorithm as the optimal instruction set;

[0134] Based on Example 4, this embodiment specifies in detail how the dynamic reconfiguration module generates the optimal set of degradation-compensation instructions after receiving the warning state trigger signal;

[0135] After the module starts, its optimization objective is constructed as an optimization function aimed at minimizing efficiency deviation; a quantitative standard for evaluating the quality of refactoring instructions is defined; where efficiency deviation is defined as the absolute value of the difference between the cluster's target efficiency and the predicted refactoring efficiency:

[0136] ;

[0137] in, : represents the efficiency deviation to be minimized;

[0138] : This represents the preset target efficiency for the cluster;

[0139] : A function for reconstructing the prediction efficiency of decision variables;

[0140] (P,V,C): These are decision variables, representing the reconstructed path instruction set, respectively. ), Rate instruction set ( ) and collaborative relationship instruction set ( For example, truck k switches from excavator i to j;

[0141] To achieve this goal, the module employs heuristic optimization algorithms, such as, but not limited to, improved ant colony optimization or genetic algorithms, to minimize efficiency bias. As a fitness function, the optimal (P,V,C) combination is searched under the premise of satisfying the constraints of mine roads, safety and equipment health.

[0142] During the algorithm search process, predicting reconstruction efficiency The value is dynamically calculated; specifically, in each search iteration of the heuristic optimization algorithm, the system will use the current iteration's downgrade-compensation instruction set. Together with the current attenuation equipment The input is fed into the workflow coupling model established in Example 3, and simulation is performed to derive the instruction set. Efficiency value ;

[0143] The search-generated instruction set must include:

[0144] Degradation strategies for degraded equipment: For example, automatically reducing the operating rate of the equipment, such as excavator i, or reducing the number of trucks it is matched with, to avoid it becoming a bottleneck;

[0145] The compensation strategies for healthy equipment specifically include: collaborative reconfiguration (C), which dynamically assigns trucks to other healthy excavators with light loads and low coupling; path reconfiguration (P), which replans truck routes to avoid congestion points; and rate adjustment (V), which increases the rate of healthy trucks within safety constraints.

[0146] After the algorithm converges, this module outputs the optimal instruction set found by the heuristic optimization algorithm, which satisfies... The (P,V,C) combination is used as the optimal instruction set for issuance;

[0147] This embodiment defines... The optimization function is defined with respect to the objective, and a job flow coupling model is used for simulation and derivation to calculate... The tool achieves a deep integration of decision-making and evaluation; instead of blindly executing preset rules, it uses heuristic search and simulation to quickly find a degradation-compensation balance point that maximizes the overall efficiency of the cluster from a massive number of (P,V,C) combinations; this ensures the scientific and optimal nature of the reconstruction decision and can effectively cope with the high dynamism of the operating environment.

[0148] Example 6:

[0149] This system also includes:

[0150] The performance prediction module is used to predict the future performance degradation curve of the device based on historical real-time device health index sequences and employs a pre-trained long short-term memory network model.

[0151] Based on Example 1, this embodiment also includes a performance prediction module; the purpose of this module is to extrapolate the future health status of the equipment using existing health data, providing a longer lead time for operation and maintenance management.

[0152] In this embodiment, the module first obtains the historical real-time device health index sequence continuously output by the health quantification module, i.e. ;

[0153] This module uses a pre-trained Long Short-Term Memory (LSTM) network model;

[0154] Long Short-Term Memory (LSTM) networks are a type of deep learning network suitable for processing and predicting time series data; in this embodiment, the pre-trained LSTM model has the following network parameters. It is represented by retrieving one or more complete historical health index sequences of the same model of equipment from performance degradation to failure recorded in CMMS. It was obtained as training data through offline training;

[0155] This module takes the aforementioned historical real-time health index sequence as input and uses a pre-trained LSTM model to calculate and predict the device's future performance degradation curve, i.e., the health index for the next k time steps. :

[0156] ;

[0157] This embodiment adds an LSTM performance prediction module, enabling the system to evaluate performance based on its current state. It has been upgraded to predict future trends (based on) This can be used not only for earlier maintenance warnings, but also for predicting... The inputs are fed into the resilience assessment module and the dynamic reconfiguration module, enabling the system to anticipate future cluster resilience states and formulate more forward-looking, proactive reconfiguration strategies, rather than just reactive reconfiguration in response to early warning states.

[0158] Example 7:

[0159] This system also includes:

[0160] Closed-loop correction module, used for:

[0161] The decision delay from triggering the reconfiguration decision to issuing the optimal instruction set is calculated by the dynamic reconfiguration module.

[0162] The actual cluster efficiency after executing the optimal instruction set is collected;

[0163] Based on the real-time equipment health index, a job flow coupling model is used to simulate and calculate the simulation efficiency of the optimal instruction set not being executed.

[0164] The economic benefit rate brought about by the reconstruction is determined based on the actual cluster efficiency, the simulated efficiency, and the preset economic value per unit efficiency.

[0165] An optimization algorithm that automatically corrects resilience state thresholds or dynamically reconfigures modules based on decision delay and economic benefits;

[0166] Based on Embodiment 1, this embodiment also includes a closed-loop correction module; it evaluates the actual effect of the reconstruction instructions and optimizes the system itself in reverse to achieve self-adaptation;

[0167] To initiate the correction, this module calculates the response speed of the reconstruction decision, i.e., the decision latency. ;

[0168] Decision delay The moment defined as when the dynamic reconfiguration module issues the optimal instruction set ( The state grading module triggers a reconfiguration decision, i.e., the moment when the FMEI first falls below 0.9. The time difference between ) All the times mentioned above are system clocks;

[0169] Meanwhile, after one instruction execution cycle, this module collects the actual cluster efficiency after executing the optimal instruction set, denoted as... ;

[0170] At the same time, the module performs a counterfactual simulation: it is based on the current real-time device health index. Using a job flow coupling model, the simulation efficiency is calculated under the condition that the optimal instruction set is not executed, i.e., it is left to develop unchecked, denoted as . ;

[0171] Based on the above actual and simulated values, this module adjusts the actual cluster efficiency. Simulation efficiency Compared with the preset unit efficiency economic value For example, the value is yuan / (ton·km / hour), which is a preset financial parameter used to determine the hourly economic efficiency rate of this restructuring. :

[0172] ;

[0173] This formula is used to quantify the hourly economic benefits of dynamic reconfiguration instructions, that is, how much economic value the decision recovers per hour, and its unit is yuan / hour.

[0174] This module continuously monitors and And based on decision delay With economic benefits Automatically correct key parameters of the system, such as the resilience state threshold, for optimization algorithm parameters of the state grading module or dynamic reconstruction module;

[0175] This embodiment introduces second-order learning capability, namely adaptive and self-optimizing capabilities, into the system by adding a closed-loop correction module; the system is no longer a static model, but can instead learn through... Assess how fast and The evaluation quantifies the system's performance and uses this to correct its core algorithms, such as optimization algorithms and key thresholds, such as resilience thresholds. This enables the system to maintain optimal operating conditions over the long term and adapt to changes in the mining environment, such as equipment aging or the addition of new equipment.

[0176] Example 8:

[0177] The closed-loop correction module automatically corrects, including:

[0178] If the decision delay continues to exceed the preset delay threshold, the search space of the optimization algorithm is simplified.

[0179] If the economic benefits continue to be lower than the preset benefit threshold, then the resilience threshold will be increased.

[0180] Based on Example 7, this embodiment specifically defines the internal logic of the automatic correction of the closed-loop correction module, aiming to achieve a dynamic balance between system efficiency and effectiveness.

[0181] Correction based on decision delay:

[0182] This module sets a preset delay time threshold; for example, this threshold could be set to 50% of the average truck cycle time. ,in This information can be obtained from the resilience assessment module to ensure that the directives take effect before the next major change in the environment.

[0183] If the detected decision delay If the delay time is consistently higher than the preset threshold, it indicates that the optimization speed of the dynamic reconfiguration module is too slow, causing the instructions to fail.

[0184] At this point, the system will automatically simplify the search space of the optimization algorithm. For example, it will automatically reduce the number of iterations of the heuristic optimization algorithm or the range of values ​​of the decision variables (P,V,C) to sacrifice some optimality in exchange for the timeliness of the decision.

[0185] Correction based on economic benefits:

[0186] This module sets a preset economic benefit threshold, for example, The corresponding efficiency recovery loss is less than 10%, and this 10% is an acceptable lower limit set based on historical data statistics or expert experience;

[0187] If the monitored economic benefits If the value remains below the preset economic benefit threshold, it indicates that the current resilience threshold is too low. For example, if 0.9 is set too low, the system will trigger an early warning too late and miss the best time for intervention.

[0188] At this point, the system will automatically increase the resilience state threshold. For example, the first preset threshold of the state classification module will be increased from 0.9 to 0.95, making it more sensitive to small deviations in cluster efficiency and thus triggering refactoring earlier.

[0189] This embodiment clarifies and The correction logic achieves a dynamic balance between system efficiency and effectiveness; through By adjusting the algorithm's complexity, the timeliness of decision-making was ensured; through The revised warning threshold ensures the effectiveness of decision-making; this dual feedback mechanism ensures that the system can intelligently adjust its internal parameters in complex real-world operating environments, keeping its operation and maintenance decisions at the optimal balance point of fast response and good results.

[0190] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A smart operation and maintenance management and predictive analysis system for mining equipment based on the Internet of Things, characterized in that, include: The data fusion module is used to collect equipment operating condition data, positioning and rate data, job scheduling data and historical maintenance data from heterogeneous devices, and align them into a unified spatiotemporal data stream. The health quantification module is used to determine the real-time device health index of a single device based on real-time monitoring values ​​in a unified spatiotemporal data stream and preset health benchmark values. The resilience assessment module is used to calculate the actual efficiency of the cluster based on the real-time equipment health index, job scheduling data and the preset cluster target efficiency. The cluster task elasticity index is obtained by dividing the ratio of the actual efficiency of the cluster to the cluster target efficiency by the real-time equipment health index. The status classification module is used to determine the cluster resilience status based on the cluster task elasticity index and the preset resilience status threshold, and to trigger a reconstruction decision when the cluster resilience status is in the warning state. The dynamic reconfiguration module is used to respond to reconfiguration decisions, aiming to minimize efficiency deviations, generate the optimal set of degradation-compensation instructions and distribute them to heterogeneous devices. The dynamic reconfiguration module generates a degradation-compensation optimal instruction set, including: Construct an optimization function with the objective of minimizing efficiency deviation, where efficiency deviation is the difference between the cluster target efficiency and the predicted reconstruction efficiency; A heuristic optimization algorithm is used to search for a degradation-compensation instruction set with the fitness function of minimizing efficiency deviation. Among them, predicting reconstruction efficiency is to obtain the efficiency value by inputting the current iteration's degradation-compensation instruction set into the queuing theory-based workflow coupling model during the search process of the heuristic optimization algorithm; using the queuing theory-based workflow coupling model, the excavator loading time is determined based on the real-time equipment health index, the queuing waiting time is calculated based on the real-time equipment health index and the number of trucks in the job scheduling data, the truck cycle time is calculated by combining the obtained truck transportation time and unloading time, and the actual efficiency of the cluster is determined based on the truck cycle time and the preset average truck load and average transportation distance. The search-generated instruction set includes: a degradation strategy for attenuated devices and a compensation strategy for healthy devices. The compensation strategy includes collaborative reconfiguration, path reconfiguration, and rate adjustment. The collaborative reconfiguration includes dynamically assigning trucks to healthy excavators with light loads and low coupling. The optimal instruction set found by the heuristic optimization algorithm is output as the optimal instruction set.

2. The IoT-based intelligent operation and maintenance management and predictive analysis system for mining equipment according to claim 1, characterized in that, The health quantification module determines the real-time device health index of a single device, including: Collect real-time monitoring values ​​from a unified spatiotemporal data stream and call preset health benchmark values ​​and weights; Calculate the dimensionless deviation of real-time monitoring values ​​from health baseline values; The real-time equipment health index is determined based on the weighted dimensionless deviation of all key performance indicators.

3. The IoT-based intelligent operation and maintenance management and predictive analysis system for mining equipment according to claim 2, characterized in that, The resilience assessment module determines the cluster task resilience index, including: Determine the excavator loading time based on real-time equipment health index; Calculate queuing time based on the real-time equipment health index and the number of trucks in the job scheduling data; Call the location and speed data and job scheduling data to obtain the truck transportation time and unloading time; Calculate the truck cycle time by combining the excavator loading time, queuing time, truck transportation time, and unloading time; The actual efficiency of the cluster is determined based on the truck cycle time and the preset average truck load and average transport distance. The cluster task elasticity index is obtained by dividing the ratio of the actual cluster efficiency to the target cluster efficiency by the real-time device health index.

4. The IoT-based intelligent operation and maintenance management and predictive analysis system for mining equipment according to claim 3, characterized in that, The state grading module determines the cluster resilience state, including: When the cluster task elasticity index is not less than the first preset threshold, the cluster resilience state is determined to be stable. When the cluster task elasticity index is less than the first preset threshold and not less than the second preset threshold, the cluster resilience state is determined to be in a warning state. When the cluster task elasticity index is less than the second preset threshold, the cluster resilience state is determined to be unstable.

5. The IoT-based intelligent operation and maintenance management and predictive analysis system for mining equipment according to claim 1, characterized in that, Also includes: The performance prediction module is used to predict the future performance degradation curve of the device based on historical real-time device health index sequences and employs a pre-trained long short-term memory network model.

6. The IoT-based intelligent operation and maintenance management and predictive analysis system for mining equipment according to claim 1, characterized in that, Also includes: Closed-loop correction module, used for: The decision delay from triggering the reconfiguration decision to issuing the optimal instruction set is calculated by the dynamic reconfiguration module. The actual cluster efficiency after executing the optimal instruction set is collected; Based on the real-time device health index, a job flow coupling model based on queuing theory is used to simulate and calculate the simulation efficiency of the optimal instruction set that has not been executed. The economic benefits are determined based on the actual cluster efficiency, the simulated efficiency, and the preset economic value per unit efficiency. An optimization algorithm that automatically corrects resilience state thresholds or dynamically reconfigures modules based on decision delay and economic benefits.

7. The IoT-based intelligent operation and maintenance management and predictive analysis system for mining equipment according to claim 6, characterized in that, The closed-loop correction module automatically corrects, including: If the decision delay continues to exceed the preset delay threshold, the search space of the optimization algorithm is simplified. If the economic benefits remain below the preset benefit threshold, the resilience threshold will be increased.

Images