Forklift gantry structure health diagnosis system based on neural network prediction

Abstract

The invention discloses a forklift portal structure health diagnosis system based on neural network prediction. The forklift portal structure health diagnosis system comprises a forklift portal oil cylinder pressure signal detection module, a forklift portal lifting height signal detection module, a forklift portal acceleration signal detection module, a health diagnosis module and an early warning module. A neural network prediction model is installed in the health diagnosis module, and the health diagnosis module calculates real-time structural stress data of the forklift portal frame through the neural network prediction model according to forklift portal frame oil cylinder pressure data, forklift portal frame lifting height data and forklift portal frame acceleration data which are obtained in real time; the health diagnosis module evaluates the health state of the current forklift mast according to the real-time structural stress data and judges whether to send an early warning signal to the early warning module or not according to the evaluation result. By establishing an accurate stress prediction model, non-contact structural stress prediction is achieved, health state evaluation is conducted on the forklift portal based on a real-time prediction result, and early warning can be conducted in time at the early stage of damage to the portal structure.

Description

Technical Field

[0001] The present invention relates to the technical field of forklift structural health monitoring, and in particular to a forklift mast structure health diagnosis system based on neural network prediction. Background Art

[0002] As an indispensable tool in modern logistics and warehousing operations, the mast structure of a forklift is a key component that bears the main load. However, due to the influence of long-term use and complex working conditions, problems such as fatigue damage and stress concentration are likely to occur in the forklift mast structure, leading to structural failure and further affecting the normal use and operation safety of the forklift. Traditional methods for forklift mast structure health monitoring mainly rely on regular manual inspections and off-line tests. This method is not only time-consuming and laborious but also difficult to capture the stress changes and damage conditions of the mast structure in real time. In addition, due to the limitations of stress testing technology, on-site operators cannot obtain the stress state of the mast structure in real time, making it difficult to detect potential safety hazards in a timely manner. Therefore, there is an urgent need for a method and system that can monitor the health status of the forklift mast structure in real time to give early warnings when approaching the safety threshold and improve operation safety. In the prior art, although there are some structural health monitoring methods based on sensors and data analysis, these methods often rely on complex hardware devices and a large amount of data processing, making it difficult to be widely promoted in practical applications. In addition, the existing methods also have certain limitations in terms of prediction accuracy and real-time performance and cannot meet the high requirements for structural health monitoring at the forklift operation site. In view of the above problems, the prior art urgently needs to be improved. Summary of the Invention

[0003] In view of this, the present invention provides a forklift mast structure health diagnosis system based on neural network prediction, which has the advantages of real-time monitoring of the stress state of the mast structure and timely warning.

[0004] To achieve the above object, the present invention provides the following technical solutions: A forklift mast structure health diagnosis system based on neural network prediction, comprising: a forklift mast cylinder pressure signal detection module, a forklift mast lift height signal detection module, a forklift mast acceleration signal detection module, a health diagnosis module, and a warning module.

[0005] Among them, the forklift mast cylinder pressure signal detection module, the forklift mast lifting height signal detection module, the forklift mast acceleration signal detection module, and the warning module are all connected to the health diagnosis module by signals; a neural network prediction model is installed in the health diagnosis module. The health diagnosis module calculates the real-time structural stress data of the forklift mast through the neural network prediction model based on the real-time obtained forklift mast cylinder pressure data, forklift mast lifting height data, and forklift mast acceleration data; the health diagnosis module evaluates the health status of the current forklift mast according to the real-time structural stress data and determines whether to send a warning signal to the warning module according to the evaluation result.

[0006] Preferably, the establishment of the neural network model includes: designing multiple groups of specific test conditions for forklifts; for each group of specific test conditions, the forklift mast cylinder pressure data, forklift mast lifting height data, forklift mast acceleration data, and forklift mast structural stress data are obtained in real time respectively; based on the data collected from all specific test conditions, a neural network prediction model between the forklift mast cylinder pressure, forklift mast lifting height, forklift mast acceleration, and forklift mast structural stress is established by using the neural network algorithm.

[0007] Preferably, the acquisition of the forklift mast structural stress data includes: performing finite element calculation and analysis on the forklift mast of each group of specific test conditions to calculate the positions where the forklift mast is subjected to greater stress under the current specific conditions, pasting strain gauges at the positions where the forklift mast is subjected to greater stress to obtain the real-time strain data of the forklift mast under the current specific conditions, and converting the real-time strain data into real-time structural stress data.

[0008] Preferably, the conversion formula for converting strain data into structural stress data is: ; where: is the structural stress, is the material elastic modulus, is the measured strain data.

[0009] Preferably, a static strength stress safety threshold of the forklift mast is preset in the health diagnosis module. When the static strength stress safety factor of the forklift mast is greater than or equal to the static strength stress safety threshold of the forklift mast, the health diagnosis module sends a static strength stress safety warning signal to the warning module for strength warning.

[0010] Preferably, the calculation formula for the static strength stress safety factor of the forklift mast is: ; where: is the static strength stress safety factor, is the yield strength of the forklift mast material, is the structural stress.

[0011] Preferably, a forklift mast fatigue cumulative damage warning value is preset in the health diagnosis module. When the fatigue damage cumulative value of the forklift mast during the forklift operation time approaches the forklift mast fatigue cumulative damage warning value, the health diagnosis module sends a fatigue life warning signal to the warning module for fatigue life warning.

[0012] Preferably, the calculation formula for the fatigue damage cumulative value of the forklift mast is: ; where: D is the fatigue damage cumulative value, is the number of stress cycles passed under the i-th stress level, is the number of stress cycles to failure under the i-th stress level, is a certain stress amplitude level causing damage.

[0013] The beneficial effects of the present invention are as follows: Compared with the prior art, a forklift mast structure health diagnosis system based on neural network prediction disclosed in the present application realizes non-contact structure stress prediction and solves the technical problem that stress cannot be directly measured at the customer site. By establishing an accurate stress prediction model, the complexity of system deployment is reduced while ensuring the monitoring accuracy. Based on the real-time prediction results, the health status of the forklift mast is evaluated, and early warning can be given in time when damage occurs to the mast structure, effectively preventing the occurrence of mast fracture accidents and improving the safety of forklift operations.

[0014] Additional aspects and advantages of the present invention will be given in part in the following description, become apparent in part from the following description, or be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is the technical flow chart of the forklift mast health diagnosis system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] The embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.

[0017] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality" means two or more unless otherwise specifically defined.

[0018] In the prior art, the forklift mast, as a core load-bearing component, generally has a risk of structural failure. Traditional health monitoring relies on strain gauges to directly measure structural stress, but there are technical bottlenecks such as difficult sensor installation and poor long-term stability in complex operating environments. When operating at the customer site, the mast is subjected to dynamic loads, resulting in complex stress distributions. The direct measurement method is difficult to cover all working conditions, and the measuring device is vulnerable to oil contamination and vibration interference, leading to data distortion and lagging in the assessment of the structural health state.

[0019] To solve the above problems, the inventor found that the oil cylinder pressure, lift height, and acceleration signals can indirectly reflect the structural mechanical state. Through analysis, it was found that the oil cylinder pressure is positively correlated with the load moment, the lift height affects the structural flexural deformation, and the acceleration reflects the dynamic impact load. Based on this, a prediction model integrating multi-source signals was proposed, and a mapping relationship between non-directly measured parameters and structural stress was established using a neural network, breaking through the technical limitations of traditional stress measurement methods.

[0020] Therefore, this application proposes a structural health diagnosis system for forklift masts based on neural network prediction, including: a forklift mast oil cylinder pressure signal detection module, a forklift mast lift height signal detection module, a forklift mast acceleration signal detection module, a health diagnosis module, and a warning module. The forklift mast oil cylinder pressure signal detection module, the forklift mast lift height signal detection module, the forklift mast acceleration signal detection module, and the warning module are all connected to the health diagnosis module by signals; a neural network prediction model is installed in the health diagnosis module. The health diagnosis module calculates the real-time structural stress data of the forklift mast through the neural network prediction model based on the real-time obtained forklift mast oil cylinder pressure data, forklift mast lift height data, and forklift mast acceleration data; the health diagnosis module evaluates the health state of the current forklift mast according to the real-time structural stress data and determines whether to send a warning signal to the warning module according to the evaluation result.

[0021] Among them, the oil cylinder pressure signal detection module is used to capture the acting force of the hydraulic system. The lift height signal detection module is used to measure the displacement of the fork. The acceleration signal detection module is used to obtain the characteristics of dynamic loads, which can be realized by a three-axis acceleration sensor and installed at key positions such as the fork carriage. The health diagnosis module runs a neural network algorithm through an embedded processor, inputs multi-source signals into the trained neural network prediction model, and outputs the predicted value of the forklift mast structural stress. The warning module is used to give warnings to the staff.

[0022] In a specific embodiment, when the system is running, an oil cylinder pressure sensor is used to monitor the pressure changes of the forklift oil cylinders in real time, such as the lifting cylinder, tilt cylinder, etc., a lifting height sensor is used to monitor the displacement trajectory of the fork to obtain the lifting height data of the current forklift mast, and an acceleration sensor is used to collect the acceleration signal of the fork. The health diagnosis module normalizes the three types of signals and inputs them into the neural network prediction model, which outputs the real-time structural stress prediction value of the forklift mast through non-linear operations. The health diagnosis module will evaluate the health status of the current forklift mast according to the real-time structural stress prediction value, and determine whether to send a warning signal to the warning module according to the evaluation result. For example, in the static strength stress safety assessment of the forklift mast, when the ratio of the real-time structural stress to the material yield strength is lower than 1.5 times the safety margin, the health diagnosis module will judge that the static strength of the current forklift mast is insufficient and send a static strength stress safety warning signal to the warning module to make the warning module give a warning.

[0023] Compared with the prior art, the traditional method needs to paste strain gauges on the surface of the mast for direct stress measurement, but it is limited by the fixed installation position of the sensor and the harsh working environment, resulting in measurement failure. This solution predicts the stress distribution through indirect parameters, eliminating the installation limitation of physical sensors under complex working conditions. In the prior art, the stress estimation method based on finite element simulation has a large amount of calculation and cannot respond in real time, while this solution can achieve rapid prediction through the trained neural network model, meeting the real-time monitoring requirements.

[0024] Through the above technical solution, this application realizes non-contact structural stress prediction, solving the technical problem that stress cannot be directly measured at the customer site. By establishing an accurate stress prediction model, the complexity of system deployment is reduced while ensuring the monitoring accuracy. Based on the real-time prediction results, the health status of the forklift mast is evaluated, which can give an early warning in the initial stage of the damage of the mast structure, effectively preventing the occurrence of mast fracture accidents and improving the safety of forklift operations.

[0025] In some embodiments, the establishment of the neural network model includes: designing multiple groups of specific test conditions for the forklift; each group of specific test conditions respectively obtains the oil cylinder pressure data of the forklift mast, the lifting height data of the forklift mast, the acceleration data of the forklift mast and the structural stress data of the forklift mast in real time; based on the data collected from all specific test conditions, a neural network prediction model between the oil cylinder pressure of the forklift mast, the lifting height of the forklift mast, the acceleration of the forklift mast and the structural stress of the forklift mast is established by using the neural network algorithm.

[0026] Among them, multiple groups of specific forklift test conditions refer to the operating scenarios formed by setting different combinations of load, lifting height, and ground slope. For example, the load range can be set from 0 to 5 tons, the lifting height range from 0 to 3 meters, and the ground slope range from 0 to 15 degrees. By covering the typical operating parameter intervals, it is ensured that the training data contains variable combinations in the actual operation of the forklift. This design can provide the input-output mapping relationship reflecting the real working conditions for the neural network model.

[0027] Among them, real-time data acquisition means synchronously collecting cylinder pressure, lifting height, acceleration, and structural stress data under each group of test conditions. For example, sensors with a sampling frequency of 100 Hz can be used to continuously record each parameter to form a multi-dimensional data set corresponding to the time series. This synchronous acquisition method ensures the strict temporal correspondence between the input features and the output targets, providing effective samples for neural network training.

[0028] Through the above technical solutions, this application realizes the real-time monitoring ability of the forklift mast structure stress. By training the neural network model with test data covering a variety of operating parameters, the prediction results can adapt to changes in different loads, heights, and ground conditions. This model uses cylinder pressure to reflect the load state, lifting height to characterize the mast attitude, and acceleration to capture dynamic impacts, comprehensively using multi-source data to improve the prediction accuracy of structural stress, providing reliable inputs for subsequent health status assessment.

[0029] Furthermore, the acquisition of forklift mast structure stress data includes: performing finite element calculation and analysis on the forklift mast under each group of specific test conditions to calculate the positions where the forklift mast is subjected to greater stress under the current specific conditions, attaching strain gauges at the positions where the forklift mast is subjected to greater stress to obtain the real-time strain data of the forklift mast under the current specific conditions, and converting the real-time strain data into real-time structural stress data.

[0030] Among them, the positions where the stress is greater refer to the regions screened based on the finite element calculation results where the equivalent stress exceeds a set proportion of the material yield strength. For example, the region where the stress exceeds 50% of the material yield strength is defined as the high-stress area. Arranging strain gauges in such regions can effectively capture the deformation data of key positions. The conversion formula for converting strain data into structural stress data is: ; where: is the structural stress, is the material elastic modulus, is the measured strain data.

[0031] Specifically, during the actual testing process, first, the stress distribution of the mast under different test conditions is predicted through finite element calculation and analysis, and the high-stress areas are screened out as key measurement points. Subsequently, strain gauges are fixed on the surface of the high-stress areas to collect the strain signals at the corresponding positions in real time. The collected strain signal data is subjected to stress conversion calculation. Through the synergistic effect of finite element calculation and physical measurement, the accurate identification of the stress concentration area and high-precision data conversion are achieved, providing reliable stress data input for the neural network prediction model and avoiding the risk of misjudging the health status caused by measurement position deviation or data distortion.

[0032] In some embodiments, a static strength stress safety threshold of the forklift mast is preset in the health diagnosis module. When the static strength stress safety factor of the forklift mast is greater than or equal to the static strength stress safety threshold of the forklift mast, the health diagnosis module sends a static strength stress safety warning signal to the warning module for strength warning.

[0033] Among them, the static strength stress safety threshold refers to a critical reference value preset for judging whether the mast structure is in a safe state. The static strength stress safety factor refers to the ratio of the material yield strength to the real-time structural stress, and can be specifically calculated by the formula: Calculated, where: is the static strength stress safety factor, is the yield strength of the forklift mast material, is the structural stress. Through this coefficient, the safety margin of the mast structure under static load can be quantitatively evaluated.

[0034] Specifically, the health diagnosis module collects the cylinder pressure, lifting height and acceleration data in real time, uses the neural network model to predict the real-time stress of the mast structure, and calculates the corresponding safety factor. When the safety factor is lower than the preset threshold, for example, when n is greater than 1.5, the system immediately triggers a warning signal. This solves the problem that the forklift mast cannot be warned in time when the static strength stress is close to the safety threshold, realizes the dynamic evaluation and real-time response to the structural safety state, and can effectively prevent the failure or accident of the forklift mast structure.

[0035] In some embodiments, a fatigue cumulative damage warning value of the forklift mast is preset in the health diagnosis module. When the fatigue damage cumulative value of the forklift mast during the forklift operation time is close to the fatigue cumulative damage warning value of the forklift mast, the health diagnosis module sends a fatigue life warning signal to the warning module for fatigue life warning.

[0036] Among them, the fatigue cumulative damage warning value refers to the damage critical threshold preset according to the fatigue characteristics of the material. The fatigue damage cumulative value is calculated from the real-time stress data. Specifically, the rain flow counting method is used to extract the stress cycle characteristics, and the linear cumulative damage theory is combined for hourly accumulation. For example, the stress spectrum is decomposed into different stress amplitude levels, and the sum of the ratios of the cycle numbers of each level to the corresponding failure cycle numbers is statistically calculated. The calculation formula for the fatigue damage cumulative value of the forklift mast is: ; where: D is the fatigue damage cumulative value, is the number of stress cycles passed under the i-th stress level, is the number of stress cycles at failure under the i-th stress level, is a certain stress amplitude level caused damage.

[0037] Specifically, during operation, the health diagnosis module continuously collects the stress data of the mast structure, and decomposes the stress-time history into discrete stress cycle events through the rain flow counting method. Each cycle event corresponds to the number of failure cycles in the material S-N curve according to its stress amplitude, and the damage ratio caused by this cycle is calculated. The damage ratios within each hour are accumulated and stored in the database, and are compared with the preset fatigue cumulative damage warning value in real time. For example, the fatigue cumulative damage warning value is set to 0.7. When the cumulative damage value reaches the set ratio of the threshold range, the health diagnosis module automatically generates a warning signal and transmits it to the warning module to trigger the alarm device.

[0038] Through the above technical solution, the present application effectively solves the problem of structural failure of the forklift mast caused by long-term fatigue damage accumulation. By quantifying the damage accumulation process and setting a dynamic warning threshold, a warning can be actively issued before the remaining life of the structure reaches the critical state, avoiding sudden fatigue failure. This solution realizes the active monitoring of the progressive damage of the metal structure, improves the prediction ability of the fatigue failure risk during the forklift operation process, and thus reduces the probability of safety accidents.

[0039] The other components and operations of the forklift mast structure health diagnosis system based on neural network prediction according to the embodiments of the present invention are known to those of ordinary skill in the art and will not be described in detail here.

[0040] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0041] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A forklift mast structure health diagnosis system based on neural network prediction, characterized in that Including: A forklift mast cylinder pressure signal detection module, a forklift mast lifting height signal detection module, a forklift mast acceleration signal detection module, a health diagnosis module, and a warning module; The forklift mast cylinder pressure signal detection module, the forklift mast lifting height signal detection module, the forklift mast acceleration signal detection module, and the warning module are all signal-connected to the health diagnosis module; A neural network prediction model is installed in the health diagnosis module. The health diagnosis module calculates the real-time structural stress data of the forklift mast through the neural network prediction model based on the real-time obtained forklift mast cylinder pressure data, forklift mast lifting height data, and forklift mast acceleration data; The health diagnosis module evaluates the health status of the current forklift mast according to the real-time structural stress data and determines whether to send a warning signal to the warning module according to the evaluation result.

2. The forklift mast structure health diagnosis system based on neural network prediction according to claim 1, wherein The establishment of the neural network model includes: Designing multiple groups of specific test conditions for forklifts; For each group of specific test conditions, the forklift mast cylinder pressure data, forklift mast lifting height data, forklift mast acceleration data, and forklift mast structural stress data are respectively obtained in real time; Based on the data collected from all specific test conditions, a neural network prediction model between the forklift mast cylinder pressure, forklift mast lifting height, forklift mast acceleration, and forklift mast structural stress is established using the neural network algorithm.

3. The forklift mast structure health diagnosis system based on neural network prediction according to claim 2, wherein The acquisition of the forklift mast structural stress data includes: performing finite element calculation and analysis on the forklift mast of each group of specific test conditions to calculate the position where the forklift mast is subjected to greater stress under the current specific condition, pasting strain gauges at the positions where the forklift mast is subjected to greater stress to obtain the real-time strain data of the forklift mast under the current specific condition, and converting the real-time strain data into real-time structural stress data.

4. The forklift mast structure health diagnosis system based on neural network prediction according to claim 3, characterized in that, The conversion formula for converting strain data into structural stress data is as follows: ; Wherein: is the structural stress, is the elastic modulus of the material, is the measured strain data.

5. The forklift mast structure health diagnosis system based on neural network prediction according to claim 1, characterized in that A static strength stress safety threshold of the forklift mast is preset in the health diagnosis module. When the static strength stress safety factor of the forklift mast is greater than or equal to the static strength stress safety threshold of the forklift mast, the health diagnosis module sends a static strength stress safety warning signal to the warning module for strength warning.

6. The forklift mast structure health diagnosis system based on neural network prediction according to claim 5, characterized in that The calculation formula for the safety factor of the stress of the forklift mast static strength is as follows: ; In the formula: is the static strength stress safety factor, is the yield strength of the forklift mast material, is the structural stress.

7. The forklift mast structure health diagnosis system based on neural network prediction according to claim 1, characterized in that, A fatigue cumulative damage warning value of the forklift mast is preset in the health diagnosis module. When the cumulative fatigue damage value of the forklift mast during the operation time of the forklift approaches the fatigue cumulative damage warning value of the forklift mast, the health diagnosis module sends a fatigue life warning signal to the warning module for fatigue life warning.

8. The forklift mast structure health diagnosis system based on neural network prediction according to claim 7, characterized in that The calculation formula for the cumulative value of the fatigue damage of the forklift mast is as follows: ; Where: D is the cumulative value of fatigue damage, is the number of stress cycles experienced at the i-th stress level, is the number of stress cycles to failure at the i-th stress level, is a certain stress amplitude level causing damage.

Images