Safety risk assessment method for large recreation facility

Abstract

The invention discloses a safety risk assessment method for a large recreation facility, and relates to the field of safety risk assessment, and the method comprises the steps: deploying monitoring sensors at Q key positions of a roller coaster, and building a state monitoring network; performing state risk assessment timeliness analysis according to the operation scene data of the roller coaster in the preset time window, and determining a risk assessment time limit constraint; constructing a state monitoring scheme optimization space based on fusion of the Q key positions and the Q associated monitoring index sets; taking the risk assessment time limit constraint as a limiting condition, carrying out the optimization search of the state monitoring scheme, and obtaining an adaptive state monitoring scheme; and in a preset time window, state monitoring and operation risk assessment are carried out on the roller coaster according to the adaptive state monitoring scheme. According to the safety risk assessment method for the large recreation facility, the safety risk assessment effect of the large recreation facility is improved, and the problems of inaccurate safety risk assessment and poor effect are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of safety risk assessment, in particular to a safety risk assessment method for large amusement facilities. BACKGROUND

[0002] Safety assessment of amusement facilities identifies and analyzes the risks of amusement equipment such as roller coasters through safety monitoring devices and other means, and takes appropriate measures to reduce risks and improve safety.

[0003] However, the existing safety risk assessment method for large amusement facilities fails to monitor the running state and timeliness, resulting in inaccurate safety risk assessment.

[0004] Therefore, there is an urgent need for a method that can balance the timeliness and accuracy of risk assessment and optimize the safety risk assessment of large amusement facilities. SUMMARY

[0005] The present application provides a safety risk assessment method for large amusement facilities, which solves the problem of inaccurate safety risk assessment and poor timeliness in the prior art. In traditional safety risk assessment of large amusement facilities, the real-time running state and component fluctuation of the amusement equipment are not considered, which makes it difficult to guarantee the timeliness and accuracy of the assessment. By dynamically optimizing the state monitoring network, the time limit constraint of risk assessment is adjusted, and under the premise of risk assessment time limit constraint, the accuracy and reliability of risk assessment are realized, and efficient and accurate control of safety risk is realized.

[0006] In view of the above problems, the present application provides a safety risk assessment method for large amusement facilities, comprising: Deploying monitoring sensors at Q key positions of the roller coaster to build a state monitoring network, wherein the Q key positions include Q associated monitoring indicator sets; Performing state risk assessment timeliness analysis according to the running scene data of the roller coaster within a preset time window to determine the risk assessment time limit constraint; Based on the Q key positions and the Q associated monitoring indicator sets, a state monitoring scheme optimization space is constructed; Taking the risk assessment time limit constraint as the limiting condition and the risk assessment accuracy and timeliness as the dual optimization target, the state monitoring scheme optimization space is optimized to obtain an adaptive state monitoring scheme; Within the preset time window, the state monitoring network is called according to the adaptive state monitoring scheme to monitor the state of the roller coaster and perform running risk assessment.

[0007] One or more technical solutions provided in the present application have at least the following technical effects or advantages: The application provides a safety risk assessment method for large amusement facilities. The method comprises the following steps: constructing a state monitoring network, performing state risk assessment timeliness analysis, fusing and constructing a state monitoring scheme optimization space, obtaining an adaptive state monitoring scheme, and finally performing state monitoring and operation risk assessment on the obtained adaptive monitoring scheme to obtain a final safety risk assessment result. By constructing the state monitoring network, the universality of the data source and the reliability of the data are ensured. By fusing and constructing the state monitoring scheme optimization space and obtaining the adaptive state monitoring scheme, prediction is performed. The monitoring scheme is widely obtained through machine algorithm, and the adaptive monitoring scheme with the maximum matching degree of the first scheme is obtained through calculation, so that the reliability of the risk assessment is ensured, and the effect of the risk assessment is improved. Compared with the previous real-time state monitoring and safety risk timeliness assessment analysis, the safety risk assessment method for large amusement facilities solves the problems of inaccurate and untimely safety risk assessment of large amusement facilities. The problems of poor timeliness and low accuracy of safety risk assessment are avoided, the effect of safety risk assessment is improved, the accuracy and reliability of risk assessment are realized under the premise of risk assessment time limit constraint, and efficient and accurate control of safety risk is realized. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0009] Figure 1 is a flowchart of the safety risk assessment method for large amusement facilities based on the application.

[0010] Figure 2 is a flowchart of obtaining an adaptive state monitoring scheme in the safety risk assessment method for large amusement facilities based on the application. DETAILED DESCRIPTION

[0011] The application provides a safety risk assessment method for large amusement facilities. The method comprises the following steps: constructing a state monitoring network, performing state risk assessment timeliness analysis, fusing and constructing a state monitoring scheme optimization space, obtaining an adaptive state monitoring scheme, and finally performing state monitoring and operation risk assessment on the obtained adaptive monitoring scheme to obtain a final safety risk assessment result. By constructing the state monitoring network, the universality of the data source and the reliability of the data are ensured. By fusing and constructing the state monitoring scheme optimization space and obtaining the adaptive state monitoring scheme, prediction is performed. The monitoring scheme is widely obtained through machine algorithm, and the adaptive monitoring scheme with the maximum matching degree of the first scheme is obtained through calculation, so that the reliability of the risk assessment is ensured, and the effect of the risk assessment is improved. Compared with the previous real-time state monitoring and safety risk timeliness assessment analysis, the safety risk assessment method for large amusement facilities solves the problems of inaccurate and untimely safety risk assessment of large amusement facilities. The problems of poor timeliness and low accuracy of safety risk assessment are avoided, the effect of safety risk assessment is improved, the accuracy and reliability of risk assessment are realized under the premise of risk assessment time limit constraint, and efficient and accurate control of safety risk is realized.

[0012] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0013] It should be noted that the terms "comprising" and "having" are intended to cover the inclusion of not exclusive, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0014] The present application will be described in detail below with reference to the drawings.

[0015] Embodiment one, as shown in the present application provides a safety risk assessment method of a large amusement facility, comprising: Figure 1 S10: deploying monitoring sensors at Q key positions of the roller coaster to build a state monitoring network, wherein the Q key positions include Q associated monitoring indicator sets.

[0016] In the embodiments of the present application, sensors are deployed at several key positions of the roller coaster, each position includes a plurality of monitoring indicators, the key positions and possible monitoring indicators are combined to build a state monitoring network for monitoring the running state thereof.

[0017] The step S10 in the method provided by the embodiments of the present application comprises: determining Q high-frequency risk positions of the roller coaster as Q key positions based on historical running records of the same type of roller coaster, wherein the key positions at least include track bends, track joints, wheels, bearings, traction systems, braking systems, pneumatic systems and electric control systems; determining Q associated monitoring indicator sets according to the structure attribute information of the Q key positions, determining that Q monitoring sensor groups are deployed at the Q key positions based on the Q associated monitoring indicator sets, and generating a state monitoring network.

[0018] Specifically, according to the historical running records of the same type of roller coaster, Q high-frequency risk positions of the roller coaster are determined as Q key positions. The historical state records of the same type of roller coaster refer to the historical state records of the roller coaster of the same type in terms of model and service life, and the historical state records include the number of daily uses and the daily use time. At the same time, the high-frequency risk positions of the roller coaster are taken as the key positions, which ensures the effectiveness of the monitoring and enables the effective monitoring indicators to be quickly obtained. The monitoring indicators can be different according to different key positions. ​

[0019] For example, the possible monitoring indicators of the key positions can be: the dislocation value of the track, the structural fatigue of the track, the weld crack at the track joint; the wear condition, vibration condition and overheating of the wheel; the wear condition, vibration condition and overheating of the bearing; the fracture risk and tensile fatigue condition of the traction system; whether the brake of the braking system is invalid, whether it is overheated, the leakage condition of the pneumatic system; whether the pressure is insufficient and whether the electric control system is short-circuited, overloaded or signal failure.

[0020] Further, the Q sets of associated monitoring indicators are determined according to the structural attribute information of the Q key positions. Different associated monitoring indicators are determined according to different structural attributes of the key positions, and the Q sets of associated monitoring indicators are determined according to the associated monitoring indicators corresponding to the key positions.

[0021] The Q sets of monitoring sensor groups are determined to be deployed at the Q key positions based on the Q sets of associated monitoring indicators, and a state monitoring network is generated. According to the Q sets of associated monitoring indicators, the Q sets of monitoring sensor groups are deployed at the corresponding key positions. The state monitoring network is composed of the key positions, the sets of associated monitoring indicators corresponding to the key positions and the associated sensor groups. In the later stage, the monitoring sensor groups can be used to collect data at the key positions.

[0022] For example, the Q key positions can include a track curve, a track joint, a wheel, a bearing, a traction system, a braking system, a pneumatic system and an electric control system; and the Q sets of associated monitoring indicators can include the dislocation value of the track, the structural fatigue of the track, the weld crack at the track joint; the wear condition, vibration condition and overheating of the wheel; the wear condition, vibration condition and overheating of the bearing, the fracture risk and tensile fatigue condition of the traction system; whether the brake of the braking system is invalid, whether it is overheated; the leakage condition of the pneumatic system, whether the pressure is insufficient; whether the electric control system is short-circuited, overloaded or signal failure.

[0023] In the embodiments of the present application, monitoring sensors are deployed at key positions of the roller coaster, and a state monitoring network is built based thereon to ensure the effectiveness and reliability of the monitoring data, thereby providing solid data support for the system.

[0024] S20: Perform state risk assessment timeliness analysis according to the running scene data of the roller coaster in a preset time window to determine a risk assessment time limit constraint.

[0025] In the embodiments of the present application, the state risk assessment timeliness analysis is performed according to the running scene data of the roller coaster in a preset time window, such as the load of the passengers, and the state risk sensitivity coefficient output by the risk assessment expert system, and the risk assessment time limit constraint can be determined. According to the risk assessment time limit constraint, the risk can be avoided to ensure that the risk of the equipment is within a controllable range within a certain time limit.

[0026] The step S20 in the method provided by the embodiment of the application comprises: obtaining running scene data of the roller coaster within a preset time window, wherein the running scene data comprises passenger load, maintenance interval length, component residual life, regional wind speed sequence and regional temperature sequence; calculating adjacent temperature difference based on the regional temperature sequence to obtain a regional temperature difference sequence, and performing mean value calculation on the regional wind speed sequence and the regional temperature difference sequence respectively to obtain a regional wind speed mean value and a regional temperature difference mean value; performing state risk sensitivity analysis according to the passenger load, the maintenance interval length, the component residual life, the regional wind speed mean value and the regional temperature difference mean value by using a risk assessment expert system, and outputting a state risk sensitivity coefficient; setting a risk assessment time limit constraint based on the state risk sensitivity coefficient.

[0027] Specifically, first, the running scene data of the roller coaster within a preset time window is obtained. The preset time window is a set time interval, which can be set artificially. According to the risk assessment time limit constraint, the length of risk prediction can be set, such as 10 minutes, etc. Within the preset time window, a monitoring sensor group is used to monitor the monitoring indicators at the key positions to obtain the data of the running scene. The running scene data comprises passenger load, maintenance interval length, component residual life, regional wind speed sequence and regional temperature sequence.

[0028] For example, the preset time window is, for example, 12 hours in the future, 24 hours in the future, etc., and the risk assessment time limit constraint is, for example, 10 minutes, etc. The passenger load is, for example, 20 people, the maintenance interval length is, for example, 7 days, and the component residual life is, for example, 5 years.

[0029] Further, the regional temperature difference sequence is obtained by calculating the adjacent temperature difference based on the regional temperature sequence. The regional temperature sequence represents the sequence of the monitored temperatures at the key positions in the monitoring area. The regional temperature difference sequence is obtained by the adjacent temperature difference of the regional temperature sequence, i.e. the difference between the adjacent regional temperatures. For a regional temperature sequence containing N temperature values, N-1 adjacent temperature differences can be calculated to form the regional temperature difference sequence. The regional temperature difference mean value is (N-1 regional temperature difference sum) / (N-1). For example, the regional temperature sequence is A, B, C, D. The adjacent temperature differences are A-B, B-C, C-D. The regional wind speed sequence represents the sequence of the monitored wind speeds at the key positions in the monitoring area. The regional wind speed mean value and the regional temperature difference mean value are obtained by performing mean value calculation on the regional wind speed sequence and the regional temperature difference sequence respectively.

[0030] For example, the region temperature sequence is 20℃, 30℃, 25℃, and 40℃. The difference between 20℃ and 30℃, 30℃ and 25℃, and 25℃ and 40℃ is calculated respectively, and the region temperature difference sequence is 10℃, 5℃, and 15℃. The region temperature difference average is (10+5+15) / 3=10℃.

[0031] Since the temperature difference change can cause material expansion and contraction and fatigue aggravation, the track expands at high temperature in summer, thereby causing the risk of misalignment. Therefore, it is necessary to monitor the region temperature sequence from the operation scene, perform risk assessment, monitor the temperature difference in the roller coaster operation process, and help perform safety risk assessment.

[0032] Further, the state risk sensitivity analysis is performed according to the passenger load, maintenance interval length, component remaining life, region wind speed average, and region temperature difference average by using the risk assessment expert system, and a state risk sensitivity coefficient is output. The expert system is a computer program system in the field of artificial intelligence, which can simulate the decision-making ability of human experts in a specific field. By combining the data or knowledge base and the reasoning engine, the logical reasoning and analysis of the experience of field experts are realized.

[0033] The data such as the passenger load, maintenance interval length, component remaining life, region wind speed average, and region temperature difference average are input into the knowledge base of the expert system by using the risk assessment expert system, the state risk sensitivity analysis is performed by using the reasoning engine, and the state risk sensitivity coefficient is obtained. The greater the passenger load, the longer the maintenance interval length, the shorter the component remaining life, the greater the region wind speed average, and the greater the region temperature difference average, the greater the risk probability and the state risk sensitivity coefficient; otherwise, the smaller the state risk sensitivity coefficient and the greater the risk.

[0034] Further, based on the state risk sensitivity coefficient, the risk assessment time limit constraint is set. Referring to the state risk sensitivity coefficient output by the risk assessment expert system, the risk assessment time limit constraint is set. The greater the state risk sensitivity coefficient, the higher the risk, and the shorter the risk assessment time limit; the smaller the state risk sensitivity coefficient, the lower the risk, and the longer the risk assessment time limit.

[0035] For example, the state risk sensitivity coefficient is 0.5, the risk assessment time limit is set to 10min, the state risk sensitivity coefficient is 0.7, and the risk assessment time limit is set to 5min.

[0036] In step S20 of the method provided in the embodiments of the application, the risk assessment time limit constraint is set based on the state risk sensitivity coefficient, including: The parameter volatility analysis is performed on the region wind speed sequence and the region temperature difference sequence respectively, and the wind speed fluctuation coefficient and the temperature difference fluctuation coefficient are obtained, wherein the fluctuation coefficient is the ratio of the standard deviation to the parameter average in the parameter sequence. The ratio of the wind speed fluctuation coefficient to the preset standard wind speed fluctuation coefficient is set as the wind speed compensation coefficient. The ratio of the temperature difference fluctuation coefficient to the preset standard temperature difference fluctuation coefficient is set as the temperature difference compensation coefficient. The environmental compensation coefficient is obtained by weighted fusion of wind speed compensation coefficient and temperature difference compensation coefficient, and the product of environmental compensation coefficient and constant K plus 1 is used as risk-sensitive compensation coefficient, where K is 0.1; The state risk sensitivity coefficient is optimized and corrected based on the risk sensitivity compensation coefficient to obtain the optimized state risk sensitivity coefficient. The ratio of the optimized state risk sensitivity coefficient to the preset standard state risk sensitivity coefficient is multiplied by the preset standard risk assessment duration to obtain the adaptive risk assessment duration, which serves as a constraint on the risk assessment time limit.

[0037] Specifically, parameter fluctuation analysis was performed on the regional wind speed series and the regional temperature difference series to obtain the wind speed fluctuation coefficient and the temperature difference fluctuation coefficient. The fluctuation coefficient is the ratio of the standard deviation of the parameters to the mean of the parameters in the series. The standard deviation of the parameters is the square root of the arithmetic mean of the squared deviations of the parameters from the mean.

[0038] The larger the volatility coefficient, the greater the dispersion of the data, and the higher the risk may be; conversely, the smaller the volatility coefficient, the lower the risk may be.

[0039] For example, the regional wind speed sequence is 1 m / s, 3 m / s, and 1.5 m / s, with a mean of 2 m / s. The standard deviation of the parameter is √[(1-2)²+(3-2)²+(1.5-2)² / (3-1)]=1.06, and the fluctuation coefficient is 1.06 / 2=0.53. A large fluctuation coefficient indicates a high degree of dispersion in the regional wind speed sequence, suggesting a potentially high risk. The regional temperature difference sequence is 10℃, 5℃, and 15℃, with a mean of 10℃. The standard deviation of the parameter is √[(10-10)²+(10-5)²+(10-15)² / (3-1)]=5, and the fluctuation coefficient is 5 / 10=0.5. A large fluctuation coefficient indicates a high degree of dispersion in the regional temperature difference sequence, suggesting a potentially high risk.

[0040] Furthermore, the ratio of the wind speed fluctuation coefficient to the preset standard wind speed fluctuation coefficient is set as the wind speed compensation coefficient. Similarly, the ratio of the temperature difference fluctuation coefficient to the preset standard temperature difference fluctuation coefficient is set as the temperature difference compensation coefficient. The wind speed compensation coefficient and the temperature difference compensation coefficient represent the degree of deviation between the regional wind speed sequence and the regional temperature difference sequence. The larger the wind speed compensation coefficient and the temperature difference compensation coefficient, the greater the deviation between the regional wind speed sequence and the regional temperature difference sequence, and the greater the degree of compensation and correction required.

[0041] The preset standard wind speed fluctuation coefficient and preset standard temperature difference fluctuation coefficient are preset values ​​that can be adaptively adjusted by those skilled in the art to ensure that the preset values ​​match the needs of safety risk assessment. The preset standard wind speed fluctuation value can be set as, for example, a custom safety standard value or the average of historical wind speed fluctuation coefficients over a historical time range.

[0042] For example, the preset standard wind speed fluctuation coefficient and the preset standard temperature difference fluctuation coefficient are set to 0.25.

[0043] The wind speed compensation coefficient is 0.53 / 0.25=2.12, and the temperature difference compensation coefficient is 0.5 / 0.25=2.

[0044] Furthermore, an environmental compensation coefficient is obtained by weighted fusion of the wind speed compensation coefficient and the temperature difference compensation coefficient. The product of the environmental compensation coefficient and a constant K is used as the risk-sensitive compensation coefficient, where K is 0.1. The wind speed compensation coefficient and the temperature difference compensation coefficient are then weighted and fused, such as by calculating a weighted average. The larger the environmental compensation coefficient, the larger the risk-sensitive compensation coefficient, and the greater the risk value of the monitored location.

[0045] For example, the wind speed compensation coefficient and the temperature difference compensation coefficient are calculated using a weighted average. The weights are typically determined by the correlation between the parameter and the state risk, or the degree of influence of the parameter on the state risk; the greater the influence, the greater the weight. For instance, since the wind speed compensation coefficient and the temperature difference compensation coefficient have the same influence on the risk sensitivity compensation coefficient, their weights are 50% and 50%, respectively. The environmental compensation coefficient is 2.12 × 50% + 2 × 50% = 2.06. The risk sensitivity compensation coefficient is 2.06 × K, where K = 0.1, meaning the risk sensitivity compensation coefficient is 2.06 × 0.1 = 0.206.

[0046] Furthermore, the state risk sensitivity coefficient is optimized and corrected based on the risk sensitivity compensation coefficient to obtain the optimized state risk sensitivity coefficient. Multiplying the risk sensitivity compensation coefficient by the state risk sensitivity coefficient yields the optimized state risk sensitivity coefficient, which can be used as the correction result. The larger the optimized state risk sensitivity coefficient, the greater the risk, and the more optimization is needed to ensure the safety of the amusement facility.

[0047] For example, the risk sensitivity compensation coefficient is 0.206, the state risk sensitivity coefficient is 0.5, and the optimized state risk sensitivity coefficient is 0.206 × 0.5 = 0.1.

[0048] Finally, the ratio of the optimized state risk sensitivity coefficient to the preset standard state risk sensitivity coefficient is multiplied by the preset standard risk assessment duration to obtain the adapted risk assessment duration, which serves as the risk assessment time constraint. The preset standard risk assessment duration is a pre-set risk assessment duration.

[0049] There is a corresponding relationship between the state risk sensitivity coefficient and the risk assessment duration; their ratio is a constant. To obtain the appropriate adaptive risk assessment duration, the ratio of the optimized state risk sensitivity coefficient to the preset standard state risk sensitivity coefficient is multiplied by the preset standard risk assessment duration. The calculation formula is: Adaptive Risk Assessment Duration = (Optimized State Risk Sensitivity Coefficient / Preset Standard State Risk Sensitivity Coefficient) × Preset Standard Risk Assessment Duration.

[0050] The obtained adaptive risk assessment duration will be used as the risk assessment time constraint. That is, the risk assessment must be completed within the adaptive risk assessment duration, thus ensuring timely risk assessment of high-risk facilities to guarantee their safety. Furthermore, the higher the risk of the facility, the shorter the risk assessment time constraint.

[0051] For example, the optimized state risk sensitivity coefficient is 0.1, the preset standard state risk sensitivity coefficient is 0.3, and the preset standard risk assessment time is 7 minutes, resulting in an adaptation risk assessment time of (0.1 / 0.3)×7=21 minutes.

[0052] In this embodiment, by collecting roller coaster operation scenario data within a preset time window, the mean values ​​of regional wind speed and temperature difference are calculated based on regional wind speed and temperature difference sequences. Using a risk assessment expert system, a state risk sensitivity coefficient is output. Parameter fluctuation analysis is performed on the regional wind speed and temperature difference sequences to obtain wind speed fluctuation coefficients and temperature difference fluctuation coefficients. The ratios of these coefficients to preset standard wind speed and temperature difference fluctuation coefficients are calculated and set as wind speed compensation coefficients and temperature difference compensation coefficients, respectively. A weighted fusion is performed to obtain an environmental compensation coefficient, which is then used as the risk sensitivity compensation coefficient. The state risk sensitivity coefficient is optimized and corrected based on the risk sensitivity compensation coefficient to obtain an optimized state risk sensitivity coefficient. The optimized state risk sensitivity coefficient is used to calculate the appropriate risk assessment duration, which serves as the risk assessment time constraint. By imposing time constraints on risk assessments, the timeliness of risk assessments is ensured. At the same time, through precise analysis and calculation, an appropriate risk assessment duration corresponding to the risk sensitivity coefficient of the optimized state is obtained, which guarantees the accuracy of risk assessments and reduces the problem of inaccurate safety risk assessments caused by outdated safety monitoring equipment.

[0053] S30: Based on the fusion of Q key locations and Q related monitoring indicator sets, construct the optimization space for the status monitoring scheme.

[0054] The state monitoring scheme optimization space refers to the set of all possible key locations and all possible combinations of monitoring indicators. Each time, M key locations are randomly selected from Q key locations, and N related monitoring indicators are randomly selected from Q related monitoring indicators, until all possible schemes are obtained. All schemes are then combined into a set, resulting in the state monitoring scheme optimization space containing the set of all possible schemes.

[0055] For example, the optimization space of the condition monitoring scheme can contain a set of related monitoring indicators for key locations such as the track, track joints, and wheels. For the track, two monitoring indicators are selected: misalignment value and structural fatigue. For the track joints, one monitoring indicator is weld cracks. For the wheels, two monitoring indicators are selected: wear condition.

[0056] In this embodiment, by constructing an optimization space for state monitoring schemes that includes all schemes, data omissions are avoided, thereby ensuring the accuracy of security risk assessment.

[0057] S40: Using the risk assessment time limit as a constraint and the accuracy and timeliness of risk assessment as dual optimization objectives, optimize the state monitoring scheme within the state monitoring scheme optimization space to obtain a suitable state monitoring scheme.

[0058] Using the risk assessment timeframe as a constraint, while ensuring the accuracy and timeliness of the risk assessment, the obtained state monitoring schemes are optimized in the space matching to obtain the most suitable state monitoring scheme, which is then used as the final scheme for risk assessment.

[0059] like Figure 2 As shown, step S40 in the method provided in this application embodiment includes: Within the optimization space of the state monitoring scheme, several key locations are randomly selected as preset monitoring locations, and several associated monitoring indicators are randomly set at each preset monitoring location to construct the first state monitoring scheme. The pre-training evaluation duration prediction plugin is used to perform risk assessment duration analysis on the first state monitoring scheme and output the first predicted evaluation duration. If the duration of the first prediction assessment exceeds the time limit for risk assessment, then the first state monitoring scheme shall be abandoned. If the first prediction assessment duration is less than or equal to the risk assessment time limit constraint, then the assessment accuracy prediction plugin is used to predict the risk assessment accuracy based on the first state monitoring scheme, and the first prediction assessment accuracy is output. The first scheme matching degree is calculated based on the first prediction and evaluation duration and the first prediction and evaluation accuracy. The scheme matching degree is negatively correlated with the prediction and evaluation duration and positively correlated with the prediction and evaluation accuracy. Within the optimization space of the state monitoring scheme, continue to randomly select state monitoring schemes and calculate the scheme matching degree until the preset number of convergences is reached. Output the state monitoring scheme corresponding to the maximum scheme matching degree and set it as the adaptive state monitoring scheme.

[0060] Specifically, several key locations are randomly selected as preset monitoring locations within the optimization space of the state monitoring scheme, and several related monitoring indicators are randomly set at each preset monitoring location as the first state monitoring scheme. Data is acquired using a random selection method to pre-train the evaluation duration prediction plugin.

[0061] For example, four key locations are randomly selected for monitoring: bearings, traction system, braking system, and pneumatic system. The monitoring indicators are as follows: for bearings, two indicators are wear and vibration; for traction systems, two indicators are fracture risk and tensile fatigue; for braking systems, one indicator is whether the brakes have failed; and for pneumatic systems, two indicators are whether the pressure is insufficient and whether the electronic control system is short-circuited or overloaded.

[0062] Furthermore, a pre-trained evaluation duration prediction plugin is used to perform risk assessment time limit analysis on the first state monitoring scheme and output the first predicted evaluation duration. A randomly selected first state monitoring scheme is input into the evaluation duration prediction plugin, and the first predicted evaluation duration is output. For example, if four key locations are input as pre-monitoring locations and their corresponding associated monitoring indicators, the evaluation duration prediction plugin will predict the corresponding evaluation durations, such as 10 minutes, 15 minutes, and 9 minutes.

[0063] Furthermore, if the duration of the first prediction assessment exceeds the time limit constraint of the risk assessment, it indicates that the duration of the first prediction assessment exceeds the time limit constraint of the risk assessment, and the result is inaccurate. If used, it will cause safety problems, and the first state monitoring scheme should be abandoned.

[0064] For example, the first prediction assessment time is 15 minutes, but the risk assessment time limit is 10 minutes. Since 15 minutes > 10 minutes, the first state monitoring scheme should be discarded.

[0065] Furthermore, if the first prediction assessment duration is less than or equal to the risk assessment time limit constraint, the assessment accuracy prediction plugin is used to predict the accuracy of the risk assessment based on the first state monitoring scheme, and outputs the first prediction assessment accuracy. The accuracy of the first prediction assessment determines whether the first state monitoring scheme should be adopted; the higher the accuracy, the more suitable the first state monitoring scheme. The assessment accuracy prediction plugin and the assessment duration prediction plugin are built on the same principle. The assessment accuracy prediction plugin outputs the assessment accuracy, and the percentage of historically identical accurately predicted events can be statistically analyzed and set as sample supervision data. For example, if the first prediction assessment duration is 5 minutes, but the risk assessment time limit constraint is 10 minutes, since 5 minutes < 10 minutes, further risk assessment accuracy prediction should be performed. Through the prediction of the assessment accuracy prediction plugin, the first prediction assessment accuracy is found to be 80%.

[0066] Furthermore, the matching degree of the first scheme is calculated based on the first prediction and evaluation duration and the first prediction and evaluation accuracy. First, the weights of the first prediction and evaluation duration and the first prediction and evaluation accuracy are calculated, and then the matching degree of the first scheme is calculated using these weights. For example, the weights of the first prediction and evaluation duration and the first prediction and evaluation accuracy are 0.5 and 0.5, respectively.

[0067] Furthermore, within the optimization space of the state monitoring scheme, random selection of state monitoring schemes continues, and scheme matching degree is calculated until the preset number of convergences is reached, and the output is stable. This yields the evaluation duration prediction plugin and the evaluation accuracy prediction plugin, which can set the state monitoring scheme corresponding to the output maximum scheme matching degree as the adaptive state monitoring scheme.

[0068] For example, the four key locations are randomly selected and the matching degree of the scheme is calculated. After reaching the preset number of convergences, the four selected key locations are input as pre-monitoring locations and corresponding associated monitoring indicators to obtain the adapted state monitoring scheme.

[0069] Step S40 of the method provided in this application embodiment, which involves pre-training and evaluating the duration prediction plugin, includes: Based on the historical status monitoring records of roller coasters, a sample status monitoring scheme set was collected, and the historical assessment duration under different sample status monitoring schemes was set as the sample assessment duration, resulting in a sample assessment duration set. The historical assessment duration is the sum of the data collection duration, data preprocessing duration, and risk assessment duration. Using a set of sample state monitoring schemes as input and a set of sample evaluation durations as supervision, a BP neural network is trained until convergence to generate an evaluation duration prediction plugin.

[0070] Specifically, based on the historical status monitoring records of the roller coaster, a sample status monitoring scheme set is collected. The historical evaluation duration under different sample status monitoring schemes is set as the sample evaluation duration, resulting in a sample evaluation duration set. The historical evaluation duration is the sum of data acquisition duration, data preprocessing duration, and risk assessment duration. For example, historical evaluation durations of roller coasters under different loads can be collected as samples, such as historical evaluation durations for holidays (high load) and weekdays (low load).

[0071] For example, the data acquisition time is 5 minutes, the data preprocessing time is 5 minutes, the risk assessment time is 7 minutes, and the historical assessment time is 17 minutes.

[0072] Furthermore, using a set of sample state monitoring schemes as input and a set of sample evaluation durations as supervision, a BP neural network is trained until convergence to generate an evaluation duration prediction plugin.

[0073] For example, the sample state monitoring scheme set is used as input, the sample evaluation duration set is used as supervision, the Adam optimizer is used (initial learning rate 1e-1), the mean squared error function (MSE) is selected, the parameters are updated through forward propagation and back propagation, and the performance is evaluated with the validation set after each training round to avoid overfitting. When the MSE loss of the training set decreases by less than 1e-6 for 5 consecutive rounds and the MSE loss of the validation set stabilizes below 0.01, the model is considered to have converged, training is stopped and the final network parameters are saved, resulting in the evaluation duration prediction plugin.

[0074] Step S40 of the method provided in this application embodiment, which calculates the first scheme matching degree based on the first prediction evaluation time and the first prediction evaluation accuracy, includes: The difference between the preset standard risk assessment time and the adapted risk assessment time is taken as the assessment time difference, and the ratio of the assessment time difference to the preset standard assessment time difference is set as the efficiency weight compensation coefficient. After compensating the initial evaluation efficiency weight with an efficiency weight compensation coefficient, the adaptive evaluation efficiency weight is obtained. The initial evaluation efficiency weight is 0.5, and the adaptive evaluation efficiency weight is greater than or equal to 0.25 and less than or equal to 0.75. The fitting evaluation accuracy weight is obtained by subtracting the fitting evaluation efficiency weight from 1. After dimensionless processing of the first prediction evaluation duration and the first prediction evaluation accuracy, the matching degree of the first scheme is obtained based on the matching evaluation efficiency weight and the matching evaluation accuracy weight.

[0075] Specifically, the difference between the preset standard risk assessment time and the adapted risk assessment time is taken as the assessment time difference, and the ratio of the assessment time difference to the preset standard assessment time difference is set as the efficiency weight compensation coefficient. The preset standard assessment time is a preset value that can be adaptively adjusted by those skilled in the art to ensure that the preset value matches the safety risk assessment requirements. The preset standard assessment time difference is the preset standard assessment time minus the risk assessment time. The efficiency weight compensation coefficient is positively correlated with the assessment time difference and negatively correlated with the predicted standard assessment time difference. The larger the efficiency weight compensation coefficient, the larger the assessment time difference, the shorter the risk assessment time limit, and in this case, the greater the weight of assessment efficiency should be.

[0076] For example, the preset standard risk assessment time is 7 minutes, and the actual risk assessment time is 4 minutes. The difference in assessment time is 7 - 4 = 3 minutes. If the preset standard assessment time difference is 4 minutes, the efficiency weight compensation coefficient is 3 / 4 = 0.75.

[0077] Furthermore, after compensating the initial evaluation efficiency weight with an efficiency weight compensation coefficient, an adapted evaluation efficiency weight is obtained. The initial evaluation efficiency weight is 0.5, and the adapted evaluation efficiency weight is greater than or equal to 0.25 and less than or equal to 0.75. The adapted evaluation efficiency weight is the product of the initial evaluation efficiency weight and the efficiency weight compensation coefficient. The larger the efficiency weight compensation coefficient, the greater the difference in evaluation time, and thus the shorter the time limit; therefore, the weight proportion of evaluation efficiency should be larger.

[0078] For example, the efficiency weight for adaptation evaluation is 0.5 × 0.75 = 0.375.

[0079] Furthermore, the fitting evaluation efficiency weight is obtained by subtracting the fitting evaluation efficiency weight from 1. The greater the difference in evaluation time and the shorter the time limit, the smaller the weight of the evaluation accuracy.

[0080] The accuracy weight of the adaptation assessment is 1-0.375=0.625.

[0081] Furthermore, the first prediction evaluation duration and the first prediction evaluation accuracy are dimensionlessly processed. Dimensionless processing is a data preprocessing technique designed to eliminate the influence of dimensions between different features, making the data comparable. Normalization scales the values ​​of a column of numerical features in the training set to between 0 and 1. Dimensionless processing is performed using software tools such as SPSSAU with the Min-Max normalization method. Enter the data processing module and select "Generate Variables." Then select the two variables to be normalized, or multiple indicators can be selected in batches. Next, select the normalization method and perform dimensionless processing using the formula: (X-Min) / (Max-Min). This method compresses the data to the range of 0 and 1, where the minimum value corresponds to 0 and the maximum value corresponds to 1. After confirming and executing the processing, the normalization operation is complete, generating new variables.

[0082] After dimensionless processing, the matching degree of the first scheme is obtained based on the efficiency weight and accuracy weight of the adaptation evaluation. The matching degree of the first scheme = (accuracy weight of adaptation evaluation × normalized accuracy) + (efficiency weight of adaptation evaluation × (1 - normalized time)). The larger the matching degree of the first scheme, the more accurate the prediction, and the more accurate the corresponding state monitoring scheme.

[0083] For example, if the accuracy of the first prediction assessment is 80% and the duration of the first prediction assessment is 5 minutes, inputting 80% and 5 minutes into SPSSAU yields results such as 0.8 and 0.17. The matching degree of the first scheme is (0.625×0.8)+(0.375×(1-0.17))=0.81. At this point, the matching degree of the first scheme is the highest, and its corresponding state monitoring scheme is taken as the adapted state monitoring scheme.

[0084] In this embodiment, a first state monitoring scheme is constructed. By training an evaluation duration prediction plugin and an evaluation accuracy prediction plugin, a first predicted evaluation duration and a first predicted evaluation accuracy are output. A first scheme matching degree is calculated based on the first predicted evaluation duration and the first predicted evaluation accuracy. After obtaining the first scheme matching degree, state monitoring schemes are randomly selected within the state monitoring scheme optimization space, and scheme matching degrees are calculated again to obtain an adapted state monitoring scheme. Dimensionless processing is used during prediction, and the obtained data is normalized, improving data comparability and analytical accuracy. Simultaneously, by comparing the obtained first scheme matching degrees, an accurate adapted state monitoring scheme is obtained, improving the accuracy of security risk assessment.

[0085] S50: Within a preset time window, the system invokes the status monitoring network to monitor the roller coaster's status and assess operational risks according to the adapted status monitoring scheme. Risk assessment is performed based on the monitored data, for example, through machine learning using neural networks. A neural network is an algorithm that simulates the structure of a human brain's neuronal network, enabling complex pattern recognition and prediction through multi-layered neuronal connections. In risk assessment, neural networks can be used to identify potential risk factors and predict risk events.

[0086] Ultimately, by obtaining further condition monitoring and operational risk assessment, the accuracy and reliability of monitoring can be improved, thereby enhancing the effectiveness of safety risk assessment.

[0087] In summary, the embodiments of this application, through the specific implementation methods described above, achieve the following technical effects: Compared to existing technologies, this application first deploys monitoring sensors at Q key locations on the roller coaster to establish a status monitoring network, ensuring the validity and reliability of the status monitoring data. Secondly, it performs a timeliness analysis of status risk assessment based on the roller coaster's operational scenario data within a preset time window, determining the time constraint for risk assessment. Under the premise of ensuring safety, it limits the assessment duration, mitigating safety issues caused by excessively long risk assessment times and ensuring the safety of risk assessment. Furthermore, it constructs a status monitoring scheme optimization space based on the fusion of the Q key locations and Q related monitoring indicator sets. By combining schemes through this optimization space, the accuracy of monitoring is improved.

[0088] Finally, this application randomly selects several key locations as preset monitoring locations within the state monitoring scheme optimization space, and randomly sets several associated monitoring indicators at each preset monitoring location to construct a first state monitoring scheme. By training a BP neural network, an evaluation duration prediction plugin is generated. This plugin is used to perform risk assessment time limit analysis on the first state monitoring scheme, outputting the first predicted evaluation duration. The first predicted evaluation duration is used to predict the accuracy of risk assessment, obtaining the first predicted evaluation accuracy rate. The matching degree of the first scheme is calculated based on the first predicted evaluation duration and the first predicted evaluation accuracy rate. After obtaining the first scheme matching degree, the state monitoring scheme is randomly selected again within the state monitoring scheme optimization space, and the matching degree is calculated to obtain an adapted state monitoring scheme. Dimensionless processing is applied to the first predicted evaluation duration and the first predicted evaluation accuracy rate, and the obtained data is normalized, improving the comparability and analytical accuracy of the data. Interference from different directions and units in the first predicted evaluation accuracy rate and the first predicted evaluation duration is eliminated. Dimensionless normalization processing avoids interference with safety risk assessment and improves the accuracy of risk assessment. Meanwhile, the obtained adaptive status monitoring scheme is subjected to final status monitoring and operational risk assessment, and the monitoring process is given final feedback to ensure the accuracy and reliability of the safety risk assessment and solve the problems of untimely and inaccurate safety risk assessment.

[0089] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0090] The above are merely preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0091] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method of safety risk assessment of a large-scale amusement facility, characterized by, The method comprises: deploying a monitoring sensor to build a state monitoring network at Q key positions of the roller coaster, wherein the Q key positions include a Q set of associated monitoring indicators; performing state risk assessment timeliness analysis according to the running scene data of the roller coaster within a preset time window to determine a risk assessment time limit constraint; based on the Q key positions and the Q set of associated monitoring indicators, a state monitoring scheme optimization space is constructed; with the risk assessment time limit constraint as a limiting condition, and the risk assessment accuracy and the assessment timeliness as dual optimization targets, an optimized search is performed in the state monitoring scheme optimization space to obtain an adaptive state monitoring scheme; within the preset time window, the roller coaster is monitored and the running risk is assessed according to the adaptive state monitoring scheme.

2. A method of safety risk assessment of an amusement ride according to claim 1, characterized in that, The state monitoring network is built, comprising: determining Q high-frequency risk parts of the roller coaster as the Q key positions based on the historical running records of similar roller coasters, wherein the key positions at least include track bends, track joints, wheels, bearings, traction systems, braking systems, pneumatic systems and electrical control systems; determining the Q set of associated monitoring indicators according to the structural attribute information of the Q key positions, and determining that the Q monitoring sensor groups are deployed at the Q key positions based on the Q set of associated monitoring indicators to generate the state monitoring network.

3. A method of safety risk assessment of an amusement ride according to claim 1, characterized in that, The state risk assessment timeliness analysis is performed according to the running scene data of the roller coaster within a preset time window to determine a risk assessment time limit constraint, comprising: obtaining the running scene data of the roller coaster within a preset time window, wherein the running scene data includes the load of the passengers, the maintenance interval time, the remaining life of the components, the regional wind speed sequence and the regional temperature sequence; calculating the adjacent temperature difference based on the regional temperature sequence to obtain the regional temperature difference sequence, and calculating the mean value of the regional wind speed sequence and the regional temperature difference sequence to obtain the regional wind speed mean value and the regional temperature difference mean value; performing state risk sensitivity analysis according to the load of the passengers, the maintenance interval time, the remaining life of the components, the regional wind speed mean value and the regional temperature difference mean value by using a risk assessment expert system, and outputting a state risk sensitivity coefficient; setting the risk assessment time limit constraint based on the state risk sensitivity coefficient.

4. A method of safety risk assessment of an amusement ride according to claim 3, characterized in that, Setting the risk assessment time limit constraint based on the state risk sensitivity coefficient, comprising: performing parameter volatility analysis on the regional wind speed sequence and the regional temperature difference sequence respectively to obtain a wind speed fluctuation coefficient and a temperature difference fluctuation coefficient, wherein the fluctuation coefficient is the ratio of the parameter standard deviation to the parameter mean value in the parameter sequence; setting the ratio of the wind speed fluctuation coefficient to the preset standard wind speed fluctuation coefficient as the wind speed compensation coefficient; setting the ratio of the temperature difference fluctuation coefficient to the preset standard temperature difference fluctuation coefficient as the temperature difference compensation coefficient; weighting and fusing the wind speed compensation coefficient and the temperature difference compensation coefficient to obtain an environmental compensation coefficient, and adding 1 to the product of the environmental compensation coefficient and a constant K as a risk sensitivity compensation coefficient, wherein K is 0.1; optimizing and correcting the state risk sensitivity coefficient according to the risk sensitivity compensation coefficient to obtain an optimized state risk sensitivity coefficient; multiplying the ratio of the optimized state risk sensitivity coefficient to the preset standard state risk sensitivity coefficient by a preset standard risk assessment time to obtain an adaptive risk assessment time, which is used as the risk assessment time limit constraint.

5. A method of safety risk assessment of an amusement ride according to claim 4, characterized in that, The adaptive state monitoring scheme is obtained by performing optimization search of the state monitoring scheme in a state monitoring scheme optimization space, including: a plurality of key positions are randomly selected as preset monitoring positions in the state monitoring scheme optimization space, and a plurality of associated monitoring indexes are randomly set at each preset monitoring position to construct a first state monitoring scheme; a pre-trained evaluation time length prediction plug-in is used to analyze the risk evaluation time limit of the first state monitoring scheme, and output a first predicted evaluation time length; if the first predicted evaluation time length is greater than the risk evaluation time limit constraint, the first state monitoring scheme is discarded; if the first predicted evaluation time length is less than or equal to the risk evaluation time limit constraint, an evaluation accuracy prediction plug-in is used to perform risk evaluation accuracy prediction according to the first state monitoring scheme, and output a first predicted evaluation accuracy; a first scheme matching degree is calculated according to the first predicted evaluation time length and the first predicted evaluation accuracy, wherein the scheme matching degree is negatively correlated with the predicted evaluation time length and positively correlated with the predicted evaluation accuracy; the random selection of the state monitoring scheme is continued in the state monitoring scheme optimization space, and the scheme matching degree is calculated until a preset convergence number is reached, and the state monitoring scheme corresponding to the maximum scheme matching degree is output as the adaptive state monitoring scheme.

6. A method of safety risk assessment of an amusement ride according to claim 5, characterized in that, The pre-trained evaluation time length prediction plug-in includes: Based on the historical state monitoring records of the roller coaster, a sample state monitoring scheme set is collected, and the historical evaluation time length under different sample state monitoring schemes is obtained as a sample evaluation time length, and a sample evaluation time length set is obtained, wherein the historical evaluation time length is the sum of the data collection time length, the data preprocessing time length and the risk evaluation time length; the sample state monitoring scheme set is used as input, and the sample evaluation time length set is used as supervision, and a BP neural network is trained to convergence to generate an evaluation time length prediction plug-in.

7. A method of safety risk assessment of an amusement ride according to claim 5, characterized in that, The first scheme matching degree is calculated according to the first predicted evaluation time length and the first predicted evaluation accuracy, including: the difference between the preset standard risk evaluation time length and the adaptive risk evaluation time length is taken as the evaluation time length difference, and the ratio of the evaluation time length difference to the preset standard evaluation time length difference is taken as the efficiency weight compensation coefficient; after the initial evaluation efficiency weight is compensated by using the efficiency weight compensation coefficient, the adaptive evaluation efficiency weight is obtained, wherein the initial evaluation efficiency weight is 0.5, and the adaptive evaluation efficiency weight is greater than or equal to 0.25 and less than or equal to 0.75; the adaptive evaluation accuracy weight is obtained by subtracting the adaptive evaluation efficiency weight from 1; after the first predicted evaluation time length and the first predicted evaluation accuracy are dimensionless processed, the first scheme matching degree is evaluated based on the adaptive evaluation efficiency weight and the adaptive evaluation accuracy weight.

Images