An internet of things intelligent management system for construction machinery

Abstract

The application provides an Internet of Things intelligent management system for construction machinery, and relates to the technical field of construction machinery management, comprising: a sensing and collecting module for collecting device state parameters and environmental parameters of the construction machinery in real time; a space-time data processing module for processing, storing and analyzing the device state parameters and environmental parameters; an intelligent decision module for generating intermediate decision instructions; a decision execution module for issuing the intermediate decision instructions to a target construction machinery control system; and an ecological coordination module for interacting with an external ecological supervision system and executing an ecological compensation strategy. The application realizes Pareto optimal solution among conflicting targets such as energy consumption, efficiency and failure rate through an intelligent decision module based on a multi-objective genetic algorithm; through an ecological red line dynamic matching algorithm, high-precision identification of the boundary of a work prohibited area is realized, and in combination with the migration cycle data of migratory birds, the interference rate of key ecological areas is reduced, so that the protection effect of biological diversity is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of construction machinery management, and particularly relates to a construction machinery Internet of Things intelligent management system. BACKGROUND

[0002] Construction machinery refers to heavy machinery equipment used in earthwork engineering, construction, mining, material handling and other fields, mainly including excavators, loaders, cranes, bulldozers, road rollers and the like. Its technical development relies on the comprehensive application of mechanical engineering, hydraulic transmission, electrical control, material science and intelligent technology. In terms of mechanical structure, construction machinery adopts high-strength steel and welding technology to ensure the stability and durability of the equipment under heavy load working conditions. The power system is usually driven by a diesel engine or an electric motor, combined with a transmission and a transmission device, to realize efficient energy conversion and power output. Hydraulic technology is one of the cores of construction machinery, which realizes precise force and motion control through hydraulic pumps, valves, cylinders and pipeline systems to meet the needs of different working scenarios. Modern construction machinery gradually integrates electronic control technology, such as electro-hydraulic proportional control, CAN bus communication and sensor monitoring, to improve the operation accuracy and automation level. In addition, remote monitoring, fault diagnosis and unmanned operation technology are becoming the development trend of the industry, which optimizes equipment management and operation efficiency with the help of Internet of Things (IoT) and artificial intelligence (AI). The progress of material science and manufacturing technology further improves the reliability of construction machinery, such as the use of wear-resistant coatings, lightweight design and modular structure, to adapt to complex working conditions and prolong the service life. In the future, with the in-depth application of new energy power, intelligent control and automation technology, construction machinery will continue to develop towards high efficiency, environmental protection and intelligence.

[0003] Generally, the construction machinery management system adopts a single time series or spatial analysis model, which is difficult to effectively capture the spatio-temporal coupling characteristics of the equipment state and environmental parameters, resulting in generally low accuracy of abnormal working condition recognition and long response delay. Moreover, the ecological protection measures often use fixed fences or post-event monitoring mode, lack of intelligent cooperation with real-time construction data, have high misentry rate in ecological sensitive areas, and the building material supply plan is disconnected with environmental protection requirements, causing resource waste.

[0004] In order to solve the defects in the prior art, the technical scheme provides a construction machinery Internet of Things intelligent management system. SUMMARY

[0005] The present application provides a construction machinery Internet of Things intelligent management system to solve the defects in the prior art.

[0006] In one aspect, the present application provides a construction machinery Internet of Things intelligent management system, comprising:

[0007] A perception collection module is configured to collect equipment state parameters and environmental parameters of the construction machinery in real time.

[0008] A space-time data processing module is configured to process, store and analyze the equipment state parameters and the environmental parameters, and output standard parameter data.

[0009] An intelligent decision module is configured to generate intermediate decision instructions based on the standard parameter data.

[0010] A decision execution module is configured to issue the intermediate decision instructions to a target construction machinery control system.

[0011] An ecological coordination module is configured to interact with an external ecological supervision system, execute an ecological compensation strategy, update the intermediate decision instructions, and output final decision instructions.

[0012] According to the construction machinery Internet of Things intelligent management system provided by the application, the space-time data processing module comprises a data processing unit, a storage unit and an analysis unit; the data processing unit is configured to preprocess the equipment state parameters and the environmental parameters to obtain preprocessed fusion parameter data; the storage unit is configured to store the preprocessed fusion parameter data to obtain fusion parameter storage data; and the analysis unit is configured to perform space-time analysis on the fusion parameter storage data to obtain standard parameter data.

[0013] According to the construction machinery Internet of Things intelligent management system provided by the application, the step of performing space-time analysis by the analysis unit comprises:

[0014] Retrieving the fusion parameter storage data to obtain key parameter data;

[0015] Performing time series analysis and spatial analysis on the key parameter data to output time dimension analysis results and spatial dimension analysis results;

[0016] Associating and fusing the time dimension analysis results and the spatial dimension analysis results to generate multi-dimensional analysis results;

[0017] Normalizing and aggregating the multi-dimensional analysis results to output standard parameter data.

[0018] According to the construction machinery Internet of Things intelligent management system provided by the application, the intelligent decision module comprises a rule engine unit, an optimization unit and a decision generation unit; the rule engine unit is configured to evaluate the standard parameter data based on predefined business rules to generate preliminary decision instructions; the optimization unit is configured to perform multi-objective optimization on the preliminary decision instructions to generate high-level decision instructions; and the decision generation unit is configured to confirm, conflict resolution and encapsulation of the high-level decision instructions to output intermediate decision instructions.

[0019] The step that the rule engine unit generates the preliminary decision instruction according to the engineering machinery Internet of Things intelligent management system provided by the application includes:

[0020] Loading the trigger condition and action mapping table of the pre-defined business rule library;

[0021] Matching the standard parameter data with the trigger condition of the rule based on the business rule library;

[0022] Generating the initial action instruction set according to the trigger condition;

[0023] Generating the preliminary decision instruction according to the initial action instruction set.

[0024] The step that the optimization unit performs multi-objective optimization according to the engineering machinery Internet of Things intelligent management system provided by the application includes:

[0025] Parsing the preliminary decision instruction into a decision variable vector and constructing a multi-objective function based on the decision variable vector;

[0026] Optimizing the initial instruction set through a multi-objective genetic algorithm based on the multi-objective function and removing solutions with a sensitivity higher than a threshold value to obtain an optimized instruction set;

[0027] Encoding the decision variables of the optimized instruction set and outputting a high-level decision instruction.

[0028] The step that the optimization unit obtains the optimized instruction set through a multi-objective genetic algorithm according to the engineering machinery Internet of Things intelligent management system provided by the application includes:

[0029] Initializing the optimization parameters of the multi-objective genetic algorithm and constructing an initial solution set by taking the decision variable vector as input;

[0030] Executing the multi-objective genetic algorithm to iteratively optimize the initial solution set and generate a Pareto frontier candidate solution set;

[0031] Performing sensitivity analysis on the Pareto frontier candidate solution set, calculating the sensitivity index of each solution, and generating a sensitivity score;

[0032] Comparing the sensitivity score with a pre-set sensitivity threshold value, removing solutions with a sensitivity score higher than the pre-set sensitivity threshold value, and obtaining the optimized instruction set.

[0033] The decision execution module of the engineering machinery Internet of Things intelligent management system provided by the application includes an instruction formatting unit and an instruction issuing unit; the instruction formatting unit is used to convert the intermediate decision instruction into a compatible format of the engineering machinery control system and output a formatted instruction; and the instruction issuing unit is used to issue the formatted instruction to the target engineering machinery control system and process response feedback and monitor the execution state.

[0034] According to the engineering machinery Internet of Things intelligent management system provided by the application, the ecological coordination module comprises a data interaction unit, an ecological compensation unit and a decision correction unit; the data interaction unit is used for data interaction with an external ecological supervision system, and sends state data of the engineering machinery and receives external ecological data; the ecological compensation unit is used for combining the state data and the external ecological data to generate an ecological compensation instruction; and the decision correction unit is used for modifying the intermediate decision instruction according to the ecological compensation instruction to obtain a final decision instruction.

[0035] According to the engineering machinery Internet of Things intelligent management system provided by the application, the step of modifying the intermediate decision instruction by the strategy updating unit comprises:

[0036] The control parameters in the intermediate decision instruction are dynamically modified based on the constraint conditions in the ecological compensation instruction;

[0037] When the external ecological data exceeds a preset threshold, a forced strategy replacement mechanism is activated to generate a replacement decision instruction;

[0038] A priority mapping rule of the ecological compensation instruction and the intermediate decision instruction is established, and a final decision instruction is output according to a rule conflict resolution result.

[0039] According to the engineering machinery Internet of Things intelligent management system provided by the application, the spatio-temporal correlation analysis accuracy of the device state and the environmental parameters is improved through the ARIMA-GIS hybrid analysis of the spatio-temporal data processing module. Through the intelligent decision module based on the NSGA-III multi-objective genetic algorithm, the Pareto optimal solution is realized among the conflicting targets such as energy consumption, efficiency and failure rate. Through the establishment of the priority mapping rule of the ecological compensation instruction and the intermediate decision instruction, a closed loop of 'perception-analysis-decision-execution-feedback' is formed, so that the system can not only optimize the current decision with future data, but also dynamically balance the relationship between construction and ecology. Based on the GIS ecological red line dynamic matching algorithm, the high-precision identification of the operation prohibited area boundary is realized, and the interference rate of the key ecological area is reduced by cooperating with the migratory cycle data of the migratory birds, so that the protection effect of the biological diversity is improved. BRIEF DESCRIPTION OF DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0041] Figure 1 It is a structure schematic diagram of the engineering machinery Internet of Things intelligent management system provided by the embodiment of the application;

[0042] Figure 2 is a step diagram that an optimization unit of an engineering machinery Internet of Things intelligent management system provided by an embodiment of the present application obtains an optimized instruction set;

[0043] Figure 3 is a step diagram that a strategy updating unit of an engineering machinery Internet of Things intelligent management system provided by an embodiment of the present application modifies a middle-level decision instruction. DETAILED DESCRIPTION

[0044] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0045] Embodiment one:

[0046] The present application is described below with reference to Figures 1-3 An engineering machinery Internet of Things intelligent management system.

[0047] As shown in Figure 1 , an engineering machinery Internet of Things intelligent management system provided by an embodiment of the present application comprises a perception collection module, a space-time data processing module, an intelligent decision module, a decision execution module and an ecological coordination module.

[0048] The perception collection module is used for collecting device state parameters and environmental parameters of the engineering machinery in real time. The device state parameters include engine speed, hydraulic system pressure, fuel level, working device position coordinates, driving speed and fault codes, etc.; the environmental parameters include working area temperature and humidity, dust concentration, noise decibel, air quality index and geographic fence information, etc.

[0049] The space-time data processing module is used for processing, storing and analyzing the device state parameters and the environmental parameters, and outputting standard parameter data. The space-time data processing module comprises a data processing unit, a storage unit and an analysis unit. The data processing unit is used for pre-processing the device state parameters and the environmental parameters to obtain pre-processed fusion parameter data. The pre-processing steps include data cleaning (such as eliminating sensor drift data), format conversion, outlier detection and multi-source data space-time alignment (through time stamp synchronization and GPS coordinate calibration). For outlier detection, the standard method is Z-score detection, and the formula is as follows:

[0050]

[0051] Wherein, x is the original parameter value (for example, engine speed or temperature), μ is the sample mean, and σ is the sample standard deviation. The specific process is: assuming that the parameter data point x obeys the normal distribution, the sample mean μ and the sample standard deviation σ are calculated. The z measurement data point deviates from the mean, when ∣z∣>3 (i.e. more than 3 times the standard deviation), it is considered as an outlier and is removed.

[0052] Let the reference position coordinates be s ref , the original acquisition coordinates s raw contain errors. After timestamp synchronization, a time deviation compensation term ϵ t is added. The corrected coordinates can ensure that the multi-source data are aligned in space and time. The formula for timestamp synchronization and GPS coordinate calibration is:

[0053]

[0054]

[0055]

[0056] Wherein, s is the spatial coordinate vector, s ref is the reference position coordinates (ground truth point), s raw is the original acquisition coordinates, ϵ t is the time synchronization deviation compensation (time-dependent term), t gps is the GPS timestamp, t sensor is the sensor acquisition timestamp, and k is the correction factor.

[0057] The storage unit is used to store the preprocessed fusion parameter data to obtain fusion parameter storage data. The storage unit adopts a hybrid architecture of distributed database cluster and time series database, supports TB-level data parallel writing and millisecond-level historical data retrieval.

[0058] The analysis unit is used for spatio-temporal analysis of the fusion parameter storage data to obtain standard parameter data.

[0059] The steps of the analysis unit for spatio-temporal analysis include:

[0060] The fusion parameter storage data is retrieved to obtain key parameter data. The key parameter data includes device core component temperature trend, operation area energy consumption distribution, and environmental pollutant concentration gradient, etc.

[0061] The key parameter data is subjected to time series analysis and spatial analysis, and the time dimension analysis result and the spatial dimension analysis result are output. The time series analysis adopts an ARIMA model to identify the periodic fluctuation characteristics of the parameters, and the spatial analysis generates a parameter spatial distribution heat map through a GIS spatial interpolation technique. ARIMA (Autoregressive Integrated Moving Average) is a classic time series prediction method. The ARIMA model consists of three parts: AR (autoregressive): a linear combination of historical values is used to predict the current value, and the order p represents the dependence on the past p values. I (difference): make the non-stationary sequence stationary through difference, and the order d represents the number of differences. MA (moving average): a linear combination of historical error terms is used to predict the current value, and the order q represents the dependence on the past q errors. Therefore, the model is denoted as ARIMA (p, d, q).

[0062] Among them, the formula for identifying the periodic fluctuation characteristics of the parameters using the ARIMA (p, d, q) model is:

[0063]

[0064] Among them, ϕ (B) = 1 - ϕ1B -... - ϕ p B p is the autoregressive operator, θ (B) = 1 + θ1B +... + θ q B q is the moving average operator, B is the backshift operator, , d is the difference order. represents the value of the time series data at time point τ, τ represents the time index, ε τ represents a white noise error term (with mean 0 and noise variance σ 2 ).

[0065] The specific process includes:

[0066] First, perform the stationary treatment, i.e., perform d-order difference on the original sequence to obtain the stationary sequence, denoted as:

[0067]

[0068] Among them, d is determined according to the ADF test, for example, if there is a trend, take d = 1. ADF test (Augmented Dickey-Fuller Test) is a statistical method for testing the stationarity of a time series, which is a key step in determining the difference order (d) in ARIMA modeling.

[0069] Secondly, on the stationary sequence, fit the ARMA (p, q) model, denoted as:

[0070]

[0071] where, is the differenced sequence.

[0072] Finally, the coefficients ϕ i and θ j are solved by maximum likelihood estimation. i ϕ j represents the autoregressive coefficients (order i ∈ [1, p]) and θ b represents the moving average coefficients (order j ∈ [1, q]). The autocorrelation function (ACF) and partial autocorrelation function (PACF) of the residuals are analyzed to detect periodicity (e.g., fluctuation period of engine speed). Autocorrelation function (ACF) and partial autocorrelation function (PACF) are the core tools for analyzing the structure of time series and determining the parameters (p and q) of ARIMA model. ACF is used to measure the linear correlation of time series with its own lag. PACF is used to measure the pure direct correlation of time series with its own lag after controlling intermediate lag terms.

[0073] For spatial analysis (GIS spatial interpolation), the inverse distance weighting (IDW) method is used to generate a heat map, represented by the formula:

[0074]

[0075]

[0076] The unknown value of the target position s0 is estimated. Based on the values z(s b ) of the nearby n known points s b , the weight w b is inversely proportional to the p-th power of the distance d(s0, s b ). The distance can be the Euclidean distance, represented as:

[0077]

[0078] where s is the spatial coordinate, s0 is the target position point, s b represents the known position point, index b ∈ [1, n], z(s b ): parameter value (e.g., pollutant concentration) of position s b , d is the Euclidean distance, w b is the weight factor, and α is the power parameter, used to control the spatial influence range, usually α = 2.

[0079] GIS (Geographic Information System) is a technical system for collecting, storing, analyzing, managing and displaying geospatial data. Spatial interpolation is a core analysis method in GIS (Geographic Information System), which is used to predict the values of unknown areas from the attribute values of known discrete points (such as temperature, precipitation, pollutant concentration, etc.), and generate continuous surfaces (such as grids or contour lines). Inverse distance weighting (Inverse Distance Weighting, IDW) is one of the most commonly used deterministic spatial interpolation methods in GIS, which is suitable for generating continuous surfaces (such as elevation, temperature, pollutant concentration distribution) from discrete point data. It has the advantages of high computational efficiency, easy implementation, suitable for most continuous data, and more accurate interpolation results in dense data point areas.

[0080] The time dimension analysis result and the space dimension analysis result are associated and fused to generate a multi-dimensional analysis result. The association and fusion establish a time-space correlation matrix of the parameters through a time-space indexing technique, and identify cross-dimensional abnormal patterns.

[0081] The formula of the time-space correlation matrix is:

[0082]

[0083] where h(γ,s) represents the value of the parameter at time γ and location s, γ e and γ v represent different time points, s e and s v represent different location points, cov represents covariance, and σ(γ) or σ(s) is the standard deviation of time γ or location s.

[0084] A time-space correlation matrix R is constructed, and each element calculates the Pearson correlation coefficient of the parameter h at different time-space points and . The normalized variance is represented as:

[0085]

[0086] where μ e ,μ v is the local mean. The time-space index is used to speed up the calculation, and when ∣R∣ is close to 1, it indicates strong correlation, and deviation identifies abnormal patterns. E is the expectation operator.

[0087] The multi-dimensional analysis result is normalized and aggregated to output standard parameter data. The normalization uses the min-max normalization method to map the parameters to the [0,1] interval, and the aggregation calculation realizes the fusion of multi-parameter time-space features through a sliding window algorithm.

[0088] The intelligent decision module is configured to generate intermediate decision instructions based on the standard parameter data. The intelligent decision module comprises a rule engine unit, an optimization unit, and a decision generation unit. The rule engine unit is configured to evaluate the standard parameter data based on predefined business rules to generate preliminary decision instructions. The predefined business rules include equipment maintenance threshold rules (e.g., triggering a cooling instruction when the engine temperature exceeds 85°C), energy consumption warning rules (e.g., triggering an optimization instruction when the unit earthwork oil consumption exceeds 30%), and environmental compliance rules (e.g., triggering a work period adjustment when the noise exceeds 70 decibels).

[0089] The optimization unit is configured to perform multi-objective optimization on the preliminary decision instructions to generate advanced decision instructions. The decision generation unit is configured to confirm, resolve conflicts, and encapsulate the advanced decision instructions to output the intermediate decision instructions. The conflict resolution adopts a priority-based negotiation mechanism, and when a device protection instruction conflicts with an efficiency improvement instruction, the device protection instruction is given priority.

[0090] The steps for the rule engine unit to generate preliminary decision instructions include:

[0091] A trigger condition and action mapping table of a predefined business rule library is loaded. The trigger condition includes parameter absolute value threshold, change rate threshold, and multi-parameter association threshold, and the action mapping table stores the condition-action correspondence in XML format.

[0092] Based on the business rule library, the standard parameter data is matched with the trigger conditions of the rules. The matching algorithm adopts the Rete pattern matching algorithm, which supports fast matching of complex logical conditions. The Rete algorithm is a high-efficiency pattern matching algorithm mainly used in rule engine systems, which can effectively handle the matching problem of a large number of rules and facts. The core idea of the Rete algorithm is to avoid repeated computation by constructing a network structure, and its main features include: state preservation (remembering part of the past matching results), incremental matching (only re-computing the affected rules), and shared nodes (sharing computation among common conditions of different rules).

[0093] According to the trigger conditions, an initial action instruction set is generated. The initial action instruction set includes device control instructions (e.g., idle speed adjustment), warning notification instructions (e.g., SMS alarm), and data collection instructions (e.g., encrypted upload).

[0094] According to the initial action instruction set, preliminary decision instructions are generated.

[0095] The steps for the optimization unit to perform multi-objective optimization include:

[0096] The preliminary decision instruction is parsed into a decision variable vector, and a multi-objective function is constructed based on the decision variable vector. The multi-objective function includes an energy consumption minimization function, an efficiency maximization function, and a failure rate minimization function. The way of constructing the multi-objective function is represented as:

[0097]

[0098]

[0099]

[0100]

[0101] wherein f en represents an energy consumption target function value, f ef represents an efficiency target function value, and f fa represents a failure rate target function value. represents a decision variable vector, such as hydraulic pressure. represents an energy consumption weight coefficient, is a device energy consumption component, such as fuel consumption. t task represents a task execution time, represents a work output indicator, such as earthwork volume, etc. Pr is a failure probability statistical model.

[0102] wherein the multi-objective includes:

[0103] Energy consumption minimization: total energy consumption of the device weighted sum.

[0104] Efficiency maximization: task completion time t task and the negative value ratio of work output (maximizing efficiency is equivalent to minimizing its negative value).

[0105] Failure rate minimization: the probability model Pr(fault | x sig ) is regressed based on historical failure data. The target function is processed by normalization to ensure dimensional consistency, represented as:

[0106]

[0107] wherein k0 is a target index, k0 ∈ {en, ef, fa}. , from the range of historical data. represents the normalized value of the k0th target function, represents the original value of the k0th target function, represents the historical minimum value of the k0th target function, represents the historical maximum value of the k0th target function.

[0108] Based on the multi-objective function, the initial instruction set is optimized by the multi-objective genetic algorithm, and the solutions with sensitivity higher than the threshold are removed to obtain the preferred instruction set. The multi-objective genetic algorithm adopts the NSGA-III algorithm framework. NSGA-III is the latest improved version of the NSGA series algorithm, which is a multi-objective optimization algorithm based on reference points. NSGA-III retains the basic framework of NSGA-II, but improves the selection mechanism, and the main features include: using a set of predefined reference points to maintain population diversity; solving the problem of inconsistent different target scales; preserving excellent individuals in the non-dominated solution set.

[0109] The process of multi-objective genetic algorithm (NSGA-III) is as follows:

[0110] Randomly generate an initial population P0 with a size of N=100. The individual is represented as a vector r c .

[0111] SBX is used to simulate binary crossover, and the crossover operation is performed with a probability P c =0.8, and the formula is:

[0112]

[0113] where η is a random factor, i.e. the crossover distribution index, based on the distribution parameter. SBX is a commonly used crossover operator for real-coded genetic algorithms. The main features are: reproducing the search characteristics of binary crossover in real number coding; controlling the distribution of offspring solutions around the parents through parameters; and automatically adjusting the search range according to the distance of the parent solutions.

[0114] Random perturbation is used for mutation operation, represented as:

[0115]

[0116] where Δr is the mutation perturbation, Δr~N(0,σ m ), and σ m is the standard deviation of the individual vector.

[0117] The non-dominated sorting assigns the rank, and the crowding distance calculation is represented as:

[0118]

[0119] The final result retains the top 20% as the elite solution.

[0120] In the above formula, N represents the population size, P c is the crossover probability, P m is the mutation probability, and d c is the crowding distance, which is used to measure the sparsity of the individual.

[0121] The decision variables of the preferred instruction set are encoded, and high-level decision instructions are output. The encoding method adopts a hybrid strategy of real number encoding and binary encoding to ensure efficient representation of continuous variables and discrete variables.

[0122] As shown in Figure 2 , wherein the step of obtaining the preferred instruction set by the optimization unit through the multi-objective genetic algorithm comprises:

[0123] Initialize the optimization parameters of the multi-objective genetic algorithm, and construct the initial solution set by taking the decision variable vector as input. The optimization parameters include the maximum number of iterations, the crossover operator, and the mutation operator.

[0124] Perform the multi-objective genetic algorithm to iteratively optimize the initial solution set and generate a Pareto front candidate solution set. During the iteration process, the crowding distance is used to maintain the diversity of the solution set, and 20% of the elite solutions are retained in each generation.

[0125] Perform sensitivity analysis on the Pareto front candidate solution set, calculate the sensitivity index of each solution, and generate a sensitivity score. The sensitivity index is the change rate of the objective function under a small perturbation of the decision variable, which is calculated through Monte Carlo simulation. It is represented as:

[0126]

[0127] where δ m is the random perturbation, f() is the objective function vector, ||·||2 is the L2 norm, S is the sensitivity score, M is the number of Monte Carlo simulations, δ norm is the perturbation normalization factor, which is usually ||δ m ||2 or a fixed small constant.

[0128] For each solution r of the Pareto front, a small perturbation δ m ~U(−0.01,0.01) (uniform distribution) is applied. Calculate the change in the target value after perturbation: the L2 norm measures the change amplitude. Through M times of Monte Carlo simulation (e.g., M=1000), the average sensitivity score S is calculated. Remove high sensitivity solutions (S>S threshold =0.15) to ensure robustness. The sensitivity threshold is based on historical decision data and is represented as:

[0129]

[0130] where μ S, σ S is the historical sensitivity mean and historical sensitivity standard deviation, S threshold is the sensitivity threshold.

[0131] The sensitivity score is compared with a preset sensitivity threshold, and solutions with a sensitivity score higher than the preset sensitivity threshold are removed to obtain a preferred instruction set. The preset sensitivity threshold is determined by the 3σ rule of historical decision data, and is usually set to 0.15.

[0132] The decision execution module is configured to issue the intermediate decision instruction to the target construction machinery control system. The decision execution module includes an instruction formatting unit and an instruction issuing unit. The instruction formatting unit is configured to convert the intermediate decision instruction into a compatible format of the construction machinery control system, and output a formatted instruction. The compatible format includes a CAN bus protocol frame (for chassis control), a Modbus protocol message (for a hydraulic system), and OPCUA protocol data (for an intelligent terminal).

[0133] The instruction issuing unit is configured to issue the formatted instruction to the target construction machinery control system, and process response feedback and monitor execution status. The instruction issuing adopts a 5G industrial module to realize low-latency transmission, and the response timeout threshold is set to 500 ms. The execution status is monitored in real time through a heartbeat message.

[0134] As shown in Figure 3 The ecological coordination module is configured to interact with an external ecological supervision system and execute an ecological compensation strategy, and update the intermediate decision instruction to output a final decision instruction. The ecological coordination module includes a data interaction unit, an ecological compensation unit, and a decision correction unit. The data interaction unit is configured to interact with the external ecological supervision system, and send state data of the construction machinery and receive external ecological data. The data interaction adopts an MQTT protocol to realize asynchronous communication, and the transmission data is encrypted using a national SM4 algorithm. The state data is sampled at a frequency of 1 Hz. MQTT is a lightweight publish / subscribe message transmission protocol designed for low-bandwidth, high-latency, or unreliable network environments, and is particularly suitable for Internet of Things (IoT) application scenarios.

[0135] The ecological compensation unit is configured to generate an ecological compensation instruction in combination with the state data and the ecological data. The ecological compensation strategy includes: synchronizing the typhoon path and rainstorm warning from the meteorological bureau, and adjusting the equipment operation period and parking position accordingly; receiving the migration notification of migratory birds and the ecological protection red line data, and delimiting the equipment exclusion area and the limited operation range accordingly; predicting the demand for building materials according to the equipment operation progress, and automatically placing orders with suppliers to ensure that the material supply is adapted to the requirements of ecological protection.

[0136] The decision correction unit is configured to modify the intermediate decision instruction according to the ecological compensation instruction to obtain the final decision instruction.

[0137] The steps of modifying the intermediate decision instruction by the strategy updating unit include:

[0138] Based on the constraint condition in the ecological compensation instruction, the control parameter in the intermediate decision instruction is dynamically modified. The control parameter modification adopts a proportional regulation algorithm. When the ecological compensation coefficient is L, the energy consumption parameter is multiplied by (1+L) for modification, that is, energy new =energy old ×(1+L). Wherein, energy old is the original energy consumption parameter, energy new is the modified energy consumption parameter, and L is the ecological compensation coefficient. When the ecological compensation coefficient L (for example, L=0.1 corresponds to 10% adjustment), the energy consumption parameter of the intermediate decision instruction is dynamically modified.

[0139] When the external ecological data exceeds the preset threshold, the forced strategy replacement mechanism is activated to generate the final decision instruction.

[0140] The priority mapping rule of the ecological compensation instruction and the intermediate decision instruction is established, and the final decision instruction is output according to the rule conflict resolution result. The priority rule is: ecological red line protection instruction> ecological compensation instruction> device operation instruction> efficiency optimization instruction.

[0141] In summary, the engineering machinery Internet of Things intelligent management system provided by the application improves the spatio-temporal correlation analysis accuracy of device state and environmental parameters through the ARIMA-GIS hybrid analysis of the spatio-temporal data processing module. Through the intelligent decision module based on the NSGA-III multi-objective genetic algorithm, the Pareto optimal solution is realized among the conflicting targets such as energy consumption, efficiency and failure rate. Through the establishment of the priority mapping rule of the ecological compensation instruction and the intermediate decision instruction, a closed loop of "perception-analysis-decision-execution-feedback" is formed, so that the system can not only optimize the current decision with future data, but also dynamically balance the relationship between construction and ecology. Based on the GIS ecological red line dynamic matching algorithm, the high-precision identification of the operation prohibited area boundary is realized, and the interference rate of the key ecological area is reduced, so that the protection effect of biological diversity is improved.

[0142] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and necessary general hardware platforms, and of course, they can also be realized by hardware. Based on such understanding, the above technical solutions or the essential part of the prior art can be embodied in the form of a software product, which can be stored in a computer readable storage medium such as ROM / RAM, magnetic disk, optical disk, etc., and includes a plurality of instructions for making a computer device (which can be a personal computer, server, or network device, etc.) execute the method of each embodiment or some part of the embodiment.

[0143] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; and although the present application has been described in detail with reference to the foregoing embodiments, it should be appreciated by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent management system for engineering machinery based on the Internet of Things, characterized in that, include: The sensing and acquisition module is used to collect equipment status parameters and environmental parameters of the construction machinery in real time; The spatiotemporal data processing module is used to process, store, and analyze the device status parameters and environmental parameters, and output standard parameter data. The intelligent decision-making module is used to generate intermediate-level decision-making instructions based on the standard parameter data; The intelligent decision-making module includes a rule engine unit, an optimization unit, and a decision generation unit. The rule engine unit is used to evaluate the standard parameter data based on predefined business rules and generate preliminary decision instructions. The optimization unit is used to perform multi-objective optimization on the preliminary decision instructions and generate advanced decision instructions. The decision generation unit is used to confirm, resolve conflicts, and encapsulate the advanced decision instructions and output the intermediate decision instructions. The steps of the optimization unit in performing multi-objective optimization include: The preliminary decision instructions are parsed into a decision variable vector, and a multi-objective function is constructed based on the decision variable vector; Based on the multi-objective function, the initial instruction set is optimized using a multi-objective genetic algorithm, and solutions with sensitivity higher than the threshold are eliminated to obtain the preferred instruction set; The optimization unit obtains the preferred instruction set through a multi-objective genetic algorithm, including the following steps: Initialize the optimization parameters of the multi-objective genetic algorithm, and construct an initial solution set using the decision variable vector as input; A multi-objective genetic algorithm is executed to iteratively optimize the initial solution set and generate a Pareto front candidate solution set; Sensitivity analysis is performed on the Pareto front candidate solution set to calculate the sensitivity index for each solution and generate a sensitivity score. Compare the sensitivity score with a preset sensitivity threshold, remove solutions whose sensitivity scores are higher than the preset sensitivity threshold, and obtain the preferred instruction set; The decision variables of the preferred instruction set are encoded, and the advanced decision instructions are output. The decision execution module is used to send the intermediate decision instructions to the target engineering machinery control system; The ecological collaboration module is used to interact with external ecological monitoring systems, execute ecological compensation strategies, update the intermediate decision instructions, and output the final decision instructions.

2. The intelligent management system for construction machinery based on the Internet of Things according to claim 1, characterized in that, The spatiotemporal data processing module includes a data processing unit, a storage unit, and an analysis unit; the data processing unit is used to preprocess the device status parameters and environmental parameters to obtain preprocessed fused parameter data. The storage unit is used to store the preprocessed fusion parameter data to obtain fusion parameter storage data; The analysis unit is used to perform spatiotemporal analysis on the fusion parameter storage data to obtain the standard parameter data.

3. The intelligent management system for engineering machinery based on the Internet of Things according to claim 2, characterized in that, The steps for spatiotemporal analysis performed by the analysis unit include: The key parameter data is obtained by retrieving the stored data of the fusion parameters; Perform time series analysis and spatial analysis on the key parameter data, and output the time dimension analysis results and spatial dimension analysis results; The time dimension analysis results are correlated and fused with the spatial dimension analysis results to generate multidimensional analysis results; The multidimensional analysis results are normalized and aggregated to output the standard parameter data.

4. The intelligent management system for engineering machinery based on the Internet of Things according to claim 1, characterized in that, The steps for the rule engine unit to generate preliminary decision instructions include: Load the trigger conditions and action mapping table of the predefined business rule base; Based on the business rule base, the standard parameter data is matched with the triggering conditions of the rules; Based on the triggering conditions, an initial action instruction set is generated; Based on the initial action instruction set, a preliminary decision instruction is generated.

5. The intelligent management system for engineering machinery based on the Internet of Things according to claim 1, characterized in that, The decision execution module includes an instruction formatting unit and an instruction issuing unit; the instruction formatting unit is used to convert the intermediate decision instructions into a compatible format of the engineering machinery control system and output formatted instructions; The instruction issuing unit is used to issue the formatting instruction to the target engineering machinery control system, and to process response feedback and monitor the execution status.

6. The intelligent management system for engineering machinery based on the Internet of Things according to claim 1, characterized in that, The ecological collaboration module includes a data interaction unit, an ecological compensation unit, and a decision correction unit. The data interaction unit is used to interact with the external ecological monitoring system, send status data of the engineering machinery, and receive external ecological data. The ecological compensation unit is used to combine the status data and the external ecological data to generate an ecological compensation instruction. The decision correction unit is used to modify the intermediate decision instruction according to the ecological compensation instruction to obtain the final decision instruction.

7. The intelligent management system for engineering machinery based on the Internet of Things according to claim 6, characterized in that, The steps by which the strategy update unit modifies the intermediate-level decision instruction include: Based on the constraints in the ecological compensation instruction, the control parameters in the intermediate decision instruction are dynamically modified; When the external ecological data exceeds a preset threshold, the forced policy replacement mechanism is activated to generate a replacement decision instruction; Establish a priority mapping rule between the ecological compensation command and the intermediate decision command, and output the final decision command based on the result of the rule conflict resolution.

Images