Laboratory risk assessment method, apparatus, and computer device

Abstract

The application relates to the technical field of safety monitoring, in particular to a laboratory risk assessment method and device and computer equipment. The laboratory comprises at least one control node, each control node is associated with at least one target device, the method comprises the following steps: obtaining the second risk value of each control node based on the first risk value of the target device associated with each control node; determining the coupling coefficient between the control nodes according to the physical distance and shared resources between the control nodes; updating the second risk value of each control node based on the coupling coefficient to obtain the third risk value of each control node; and obtaining the global risk value of the laboratory based on the third risk value of each control node. The method can improve the evaluation accuracy.

Description

Technical Field

[0001] This application relates to the field of safety monitoring technology, and in particular to a laboratory risk assessment method, apparatus, and computer equipment. Background Technology

[0002] With the development of science and technology, more and more equipment is being applied to production and daily life, making risk assessment of various equipment a crucial task. For example, laboratories in universities and research institutions generally face the challenges of large equipment inventories, slow updates, and long service lives for some equipment. Outdated equipment often lacks built-in safety sensors and data communication capabilities, making it difficult to integrate with modern intelligent management systems. Simultaneously, insufficient operational experience among laboratory personnel increases the risk of safety accidents caused by misoperation.

[0003] Currently, most existing safety management systems rely on collecting environmental data and using network models for overall laboratory safety monitoring, which suffers from low prediction accuracy. For example, the existing technology CN118094461A collects laboratory environmental data and laboratory video, and uses a YOLO model for anomaly detection, but this also suffers from low prediction accuracy. Summary of the Invention

[0004] Therefore, it is necessary to provide a laboratory risk assessment method, apparatus, and computer equipment that can improve the accuracy of assessments in response to the aforementioned technical problems.

[0005] A risk assessment method for a laboratory, the laboratory comprising at least one control node, each control node being associated with at least one target device, the method comprising:

[0006] S1. Based on the first risk value of the target device associated with each control node, obtain the second risk value of each control node;

[0007] S2. Determine the coupling coefficient between each control node based on the physical distance and shared resources between each control node;

[0008] S3. Based on the coupling coefficient, update the second risk value of each control node to obtain the third risk value of each control node;

[0009] S4. Based on the third risk value of each control node, the global risk value of the laboratory is obtained.

[0010] In one embodiment, step S2 includes:

[0011] The physical distance and shared resources between each control node are obtained, and the shared resources include one or more of a shared power circuit and a shared gas supply circuit.

[0012] Determine the first coupling coefficient corresponding to the physical distance and the second coupling coefficient corresponding to the shared resource;

[0013] The coupling coefficients between each control node are obtained by adding the first coupling coefficient and the second coupling coefficient.

[0014] In one embodiment, step S3 includes:

[0015] The formula for calculating the third risk value is as follows: ;

[0016] Among them, V3 i To control the third risk value of node i, V2 i e is the second risk value for controlling node i. ij is the coupling coefficient between control node i and control node j, where j is not equal to i, and n is the number of control nodes.

[0017] In one embodiment, step S1 includes:

[0018] S11. Obtain multiple scoring results for each target device in the control node, and calculate the first weight of each target device in the control node based on the scoring results using the analytic hierarchy process (AHP).

[0019] S12. Based on the time series data of the target device, output the first risk value of the target device using the lightweight network model deployed in the control node;

[0020] S13. Based on the first weight and the first risk value of each target device in the control node, a weighted sum is performed to obtain the second risk value of the control node.

[0021] In this application, a second risk value for each control node is obtained based on the first risk value of the target equipment associated with each control node. A coupling coefficient between the control nodes is determined according to the physical distance and shared resources between them. Based on the coupling coefficient, the second risk value of each control node is updated to obtain a third risk value. Finally, based on the third risk value of each control node, the global risk value of the laboratory is obtained. This scheme achieves accurate risk quantification from local equipment to the global level through a hierarchical aggregation and coupling propagation mechanism, significantly improving the comprehensiveness, realism, and early warning capability of safety risk assessment, thereby enhancing the accuracy of risk assessment.

[0022] In one embodiment, step S2 includes:

[0023] The physical distance and shared resources between each control node are obtained, and the shared resources include one or more of a shared power circuit and a shared gas supply circuit.

[0024] Determine the first coupling coefficient corresponding to the physical distance and the second coupling coefficient corresponding to the shared resource;

[0025] The coupling coefficients between each control node are obtained by adding the first coupling coefficient and the second coupling coefficient.

[0026] In this application, by obtaining the physical distance and shared resources between each control node, a first coupling coefficient corresponding to the physical distance and a second coupling coefficient corresponding to the shared resources are determined. Based on the sum of the first coupling coefficient and the second coupling coefficient, the coupling coefficient between each control node is obtained, which can more accurately assess the coupling strength between control nodes.

[0027] In one embodiment, step S3 includes:

[0028] The formula for calculating the third risk value is as follows: ;

[0029] Among them, V3 i To control the third risk value of node i, V2 i e is the second risk value for controlling node i. ij is the coupling coefficient between control node i and control node j, where j is not equal to i, and n is the number of control nodes.

[0030] In this application, by using Calculating a third risk value can effectively capture the "contagion" and "cumulative effect" of risks, thereby achieving a fusion assessment of "local risks" and "coupled risks".

[0031] In one embodiment, step S1 includes:

[0032] S11. Obtain multiple scoring results for each target device in the control node, and calculate the first weight of each target device in the control node based on the scoring results using the analytic hierarchy process (AHP).

[0033] S12. Based on the time series data of the target device, output the first risk value of the target device using the lightweight network model deployed in the control node;

[0034] S13. Based on the first weight and the first risk value of each target device in the control node, a weighted sum is performed to obtain the second risk value of the control node.

[0035] In this application, by integrating the analytic hierarchy process (AHP) with a lightweight time series model, a dynamic, weighted, and interpretable comprehensive assessment of the risks of multiple devices in a control node is achieved, which significantly improves the accuracy and practicality of risk perception while ensuring computational efficiency.

[0036] In one embodiment, step S12 includes:

[0037] Collect time-series data of the target device within a preset time period;

[0038] If the time series data has a corresponding constraint threshold, then the time series data is standardized using the constraint threshold;

[0039] If the time series data does not have a corresponding constraint threshold, the mean and standard deviation are calculated based on the historical time series data, and the time series data is standardized based on the mean and standard deviation.

[0040] Target features are extracted from the standardized time-series data, and based on the target features, a first risk value of the target device is output using a lightweight network model deployed in the control node. The target features include one or more of abnormal change rate, continuous deviation ratio, and rated deviation.

[0041] In this application, the accuracy, generalization ability and interpretability of target equipment risk assessment are significantly improved by adaptive standardization and domain-oriented feature extraction.

[0042] In one embodiment, if the time series data does not have a corresponding constraint threshold, then the mean and standard deviation are calculated based on historical time series data, and the time series data is standardized based on the mean and standard deviation, including:

[0043] Statistical analysis of historical time-series data of a sliding time window, wherein the sliding time window ends at the current time;

[0044] Calculate the mean and standard deviation of the historical time series data, and construct a safety reference interval based on the mean and standard deviation;

[0045] Based on the maximum and minimum values, the mean and the standard deviation in the safety reference interval, the time series data is standardized to obtain standardized time series data.

[0046] The standardization formula for the time series data is as follows:

[0047] ;

[0048] X represents time-series data, X norm For the standardized time series data, μ is the mean, σ is the variance, μ+2σ is the maximum value, and μ-2σ is the minimum value.

[0049] In this application, a sliding time window is used to dynamically construct and standardize a safety reference interval, achieving adaptive normalization and anomaly enhancement of time-series data. This significantly improves the accuracy, robustness, and deployment feasibility of edge-side risk assessment, while also possessing good interpretability and engineering practicality. Furthermore, dynamically updating the safety reference interval can adapt to changes in the operating environment, forming a complete safety governance closed loop.

[0050] In one embodiment, extracting target features from the standardized time-series data includes:

[0051] Calculate the fluctuation amplitude of the time series data within a unit time period, and determine the abnormal change rate based on the fluctuation amplitude and a first time length, where the first time length is the duration of the unit time period;

[0052] The duration of time series data exceeding the constraint threshold is counted, and the percentage of continuous deviation is obtained based on the quotient of the duration divided by the second time length, where the second time length is the duration of the preset time period.

[0053] The rated deviation is determined based on the difference between the time series data and the preset standard threshold.

[0054] In this application, by constructing a three-in-one feature system of abnormal change rate, continuous deviation ratio and rated deviation, a multi-angle, highly interpretable and low-overhead quantitative characterization of the abnormal operation of target equipment is achieved, which significantly improves the comprehensiveness, accuracy and engineering practicality of risk identification and lays a solid foundation for subsequent intelligent assessment and response.

[0055] In one embodiment, step S4 is followed by:

[0056] Obtain the preset risk classification criteria;

[0057] Based on the risk grading criteria, the global risk level corresponding to the global risk value is determined;

[0058] Based on the global risk level, the corresponding response strategy is executed.

[0059] In this application, a preset risk grading standard is obtained, and based on this standard, the global risk level corresponding to the global risk value is determined. This allows for the execution of corresponding response strategies based on the global risk level. Furthermore, executing corresponding response strategies based on the global risk level enables automatic handling of high-risk situations and multi-terminal alarms, while providing lightweight monitoring for low-risk situations, thus balancing security and resource efficiency.

[0060] A laboratory risk assessment device, the device comprising:

[0061] The second risk value determination module is used to obtain the second risk value of each control node based on the first risk value of the target device associated with each control node.

[0062] The coupling coefficient determination module is used to determine the coupling coefficient between each of the control nodes based on the physical distance and shared resources between them.

[0063] The third risk value determination module is used to update the second risk value of each control node based on the coupling coefficient to obtain the third risk value of each control node.

[0064] The global risk value determination module is used to obtain the global risk value of the laboratory based on the third risk value of each of the control nodes.

[0065] A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method described above.

[0066] The beneficial effects of the aforementioned laboratory risk assessment method, apparatus, and computer equipment are as follows: Based on the first risk value of the target equipment associated with each control node, a second risk value is obtained for each control node; based on the physical distance and shared resources between the control nodes, a coupling coefficient is determined; based on the coupling coefficient, the second risk value of each control node is updated to obtain a third risk value; and based on the third risk value of each control node, the global risk value of the laboratory is obtained. This scheme, through hierarchical aggregation and coupling propagation mechanisms, achieves accurate risk quantification from local equipment to the global level, significantly improving the comprehensiveness, realism, and early warning capability of safety risk assessment, thereby enhancing the accuracy of risk assessment. Attached Figure Description

[0067] Figure 1 This is a diagram illustrating the application environment of a laboratory risk assessment method in one embodiment.

[0068] Figure 2 This is a flowchart illustrating a laboratory risk assessment method in one embodiment;

[0069] Figure 3 This is a schematic diagram of the control node hardware in one embodiment;

[0070] Figure 4 This is a schematic diagram of one side of the sensor in one embodiment;

[0071] Figure 5 This is a schematic diagram of the other side of the sensor in one embodiment;

[0072] Figure 6 This is a schematic diagram of multiple sensors and control nodes stacked in one embodiment;

[0073] Figure 7 This is a schematic diagram of multiple sensors and control nodes stacked together in one embodiment;

[0074] Figure 8 This is a schematic diagram of a laboratory risk assessment system in one embodiment;

[0075] Figure 9 This is a structural block diagram of a laboratory risk assessment device in one embodiment;

[0076] Figure 10 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0077] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0078] The laboratory risk assessment method provided in this application embodiment can be applied to, for example... Figure 1 In the environment shown, the laboratory includes at least one control node, each control node being associated with at least one target device. The server obtains a second risk value for each control node based on a first risk value of the target device associated with each control node. The server determines a coupling coefficient between each control node based on the physical distance and shared resources between them. Based on the coupling coefficient, the server updates the second risk value of each control node to obtain a third risk value. Finally, the server obtains the global risk value of the laboratory based on the third risk value of each control node.

[0079] In one embodiment, such as Figure 2 As shown, a laboratory risk assessment method is provided, which can be applied to... Figure 1 Taking the server in the example, the laboratory includes at least one control node, and each control node is associated with at least one target device, including the following steps:

[0080] S1. Based on the first risk value of the target device associated with each control node, obtain the second risk value of each control node;

[0081] Each control node is equipped with a lightweight network model, which is used to output the first risk value of the target device.

[0082] The second risk value of the control node can be obtained by summing the first risk values ​​of each target device, or by weighted summing of the first risk values ​​of each target device. Further, taking any target device as the device to be evaluated, a score is given to the device to be evaluated based on its importance relative to other target devices, resulting in multiple score results for each target device. Based on the multiple score results for each target device, the first weight of each target device in the control node is calculated using the analytic hierarchy process (AHP). Finally, the second risk value of the control node is obtained by weighted summing of the first weights and first risk values ​​of each target device in the control node.

[0083] The control nodes of the laboratory can be divided according to function or importance. For example, when divided according to function, the laboratory can be divided into three control nodes: a core experimental area, an auxiliary area, and a storage area. When divided according to importance, the laboratory can be divided into a first control node, a second control node, and a third control node, where the first control node is more important than the second control node, and the second control node is less important than the third control node.

[0084] S2. Determine the coupling coefficient between each control node based on the physical distance and shared resources between each control node;

[0085] Physical distance refers to the spatial interval between two control nodes in three-dimensional space. Shared resources refer to resources used jointly by the two control nodes. Shared resources include, but are not limited to, shared power circuits and shared gas supply circuits. For example, if control node 1 and control node 2 only share the same power circuit, then the power circuit is a shared resource between control node 1 and control node 2.

[0086] Both physical distance and shared resources correspond to a coupling coefficient. The coupling coefficient between each control node can be obtained by adding the coupling coefficients corresponding to physical distance and shared resources respectively.

[0087] S3. Based on the coupling coefficient, update the second risk value of each control node to obtain the third risk value of each control node;

[0088] The coupling coefficient refers to the coupling coefficient between two control nodes, which is mainly used to characterize the degree of association and coupling between the two control nodes.

[0089] In some embodiments, the formula for determining the third risk value includes: V3 i To control the third risk value of node i, V2 i e is the second risk value for controlling node i. ij Let be the coupling coefficient between control node i and control node j, where j is not equal to i, and n be the number of control nodes. This is a coupling strength adjustment factor used to control the degree of impact of external risk propagation on the region.

[0090] S4. Based on the third risk value of each control node, the global risk value of the laboratory is obtained.

[0091] The laboratory's global risk value can be obtained by summing the third risk values ​​of each control node, or by weighted summing of the third risk values ​​of each control node. Further, taking any control node as the node to be evaluated, a score is assigned to the node to be evaluated based on its importance relative to other control nodes, resulting in multiple score results for each control node. Based on these multiple score results, the analytic hierarchy process (AHP) is used to calculate the second weight of each control node. The second weight and the third risk value of each control node are then weighted and summed to obtain the laboratory's global risk value. Global risk value V global The calculation formula is W2 j It is the second weight of the control node j, V3 j It is the third risk value of control node j, and n is the number of control nodes.

[0092] The aforementioned laboratory risk assessment method derives a second risk value for each control node based on the first risk value of the target equipment associated with each control node. Then, based on the physical distance and shared resources between control nodes, a coupling coefficient is determined. This coupling coefficient is used to update the second risk value of each control node, resulting in a third risk value. Finally, based on the third risk value of each control node, the overall risk value of the laboratory is obtained. This scheme, through hierarchical aggregation and coupling propagation mechanisms, achieves precise risk quantification from local equipment to the global level, significantly improving the comprehensiveness, accuracy, and early warning capabilities of safety risk assessment, thus enhancing the accuracy of risk assessment.

[0093] In one embodiment, step S2 includes:

[0094] Obtain the physical distance and shared resources between each control node. Shared resources include one or more of the following: shared power circuit and shared gas supply circuit.

[0095] Determine the first coupling coefficient corresponding to physical distance and the second coupling coefficient corresponding to shared resources;

[0096] The coupling coefficients between each control node are obtained by adding the first coupling coefficient and the second coupling coefficient.

[0097] Specifically, obtaining the physical distance and shared resources between control nodes refers to obtaining the physical distance and shared resources between two control nodes. For example, if a laboratory includes control node 1, control node 2, and control node 3, then the physical distance and shared resources between control node 1 and control node 2, control node 1 and control node 3, and control node 2 and control node 3 must be obtained.

[0098] The first coupling coefficient is positively correlated with the physical distance. The first coupling coefficient corresponding to the physical distance can be obtained from a physical distance mapping library, which includes various candidate physical distances and the first coupling coefficient associated with each candidate physical distance. Specifically, the physical distance is matched with various candidate physical distances, and the candidate physical distance that matches the physical distance is determined as the matching distance. The first coupling coefficient associated with the matching distance is then used as the first coupling coefficient corresponding to the physical distance.

[0099] The second coupling coefficient corresponding to the shared resource can be obtained through a shared resource mapping library, which includes various candidate shared resources and the second coupling coefficient associated with each candidate shared resource. Specifically, the shared resource is matched with various candidate shared resources, and the candidate shared resource that matches the shared resource is determined as the matched resource. The second coupling coefficient associated with the matched resource is then used as the second coupling coefficient corresponding to the shared resource.

[0100] The number of shared resources between control nodes may be zero, multiple, or one. Furthermore, when the number of shared resources between two control nodes is zero, the second coupling coefficient is also zero.

[0101] The coupling coefficient between each control node is obtained by adding the first coupling coefficient and the second coupling coefficient. Specifically, it refers to obtaining the coupling coefficient between two control nodes by adding the first coupling coefficient and the second coupling coefficient between them. For example, the coupling coefficient between control node A and control node B can be obtained by adding the first coupling coefficient and the second coupling coefficient between them.

[0102] In this embodiment, by obtaining the physical distance and shared resources between each control node, a first coupling coefficient corresponding to the physical distance and a second coupling coefficient corresponding to the shared resources are determined. Based on the sum of the first coupling coefficient and the second coupling coefficient, the coupling coefficient between each control node is obtained, which can more accurately assess the coupling strength between control nodes.

[0103] In one embodiment, step S4 includes:

[0104] The formula for calculating the third risk value is: ;

[0105] Among them, V3 i To control the third risk value of node i, V2 i e is the second risk value for controlling node i. ij is the coupling coefficient between control node i and control node j, where j is not equal to i, and n is the number of control nodes.

[0106] In this embodiment, by using Calculating a third risk value can effectively capture the "contagion" and "cumulative effect" of risks, thereby achieving a fusion assessment of "local risks" and "coupled risks".

[0107] In one embodiment, step S1 includes:

[0108] S11. Obtain multiple scoring results for each target device in the control node, and calculate the first weight of each target device in the control node based on the scoring results using the analytic hierarchy process (AHP).

[0109] Among them, multiple rating results of the target device represent the importance of the target device relative to other target devices.

[0110] Based on the scoring results, the formula for calculating the first weight of each target device in the control node using the analytic hierarchy process is as follows: , , Let be the geometric mean of the target device i. The score is given by target device i relative to target device j, where n is the number of target devices in this control node. Let i be the first weight of the target device i.

[0111] In a specific application, based on the importance of devices 1, 2 and 3 in a certain control node, a 1-9 scale (1 = equally important, 3 = slightly important, 5 = significantly important, 7 = strongly important, 9 = extremely important) is used to score them. The scoring results of devices 1, 2 and 3 are shown in Table 1.

[0112] Table 1

[0113]

[0114] According to the analytic hierarchy process (AHP), the geometric mean of device 1 is: The geometric mean of device 2 is The geometric mean of device 3 is The geometric mean of each device is summed to obtain... Based on the summation result, the first weight of device 1 is obtained. The first weight of device 2 The first weight of device 3 .

[0115] S12. Based on the time series data of the target device, output the first risk value of the target device using the lightweight network model deployed in the control node;

[0116] The timing data of the target device includes the target device's own data, the data of the environment in which the target device is located, and the sequence data of operations performed on the target device. The timing data of the target device itself can be collected by sensors deployed in the target device, while the timing data of the environment in which the target device is located can be collected by sensors deployed in the environment or in the control node hardware.

[0117] Specifically, sensors are installed on older target equipment to collect physical quantities such as electrical signals (current, voltage, power), gas concentration, vibration, and sound. These sensors collect the target equipment's own electrical signals, gas concentration, vibration, and sound in real time. The sensors support power supply from batteries, solar panels, or external power sources, adapting to various power supply methods. For new equipment, the control node hardware is directly connected to the new equipment's data interface to acquire timing data. For older equipment without communication interfaces, sensors are fixed to the data acquisition point of the target equipment using methods such as adhesive or screws, thus achieving rapid sensor deployment without modifying the target equipment itself.

[0118] Specifically, sensors for temperature, humidity, smoke, flame, and harmful gases are integrated into the control node associated with the target device to monitor environmental parameters in real time. At the same time, an expansion interface is reserved to facilitate the installation of special gas sensors in the control node.

[0119] Specifically, based on the timestamps and duration of sensor data, three core time-series data categories are extracted: "operation duration, operation interval, and operation sequence." The operation duration is determined by the time span of continuous sensor fluctuations; the operation interval is determined by the timestamp difference between two similar operation data points from the sensor; and the operation sequence is determined by the order of timestamps of sudden changes in data from multiple device sensors.

[0120] The control node is responsible for aggregating time-series data from each associated target device. It possesses edge computing capabilities, deploys a lightweight network model, and supports real-time analysis of single-device / local area risks, reducing cloud pressure and transmission latency. The control node is equipped with a tri-color LED (Light Emitting Diode) audible and visual alarm (red / yellow / green) and a data storage module for storing time-series data. The control node supports data interaction with cloud platforms, large screens, and mobile terminals. The control node hardware can be externally powered and features a tri-color LED on top. The top also includes magnets, positive and negative contacts, and a positioning groove. It incorporates built-in environmental sensors (such as air pressure and humidity), Bluetooth and wireless network connectivity hardware, and edge computing hardware. The control node hardware includes... Figure 3 As shown. The sensor can be powered by an external power source and also has a built-in battery, allowing it to operate independently of an external power supply. The sensor incorporates internal circuitry and sensing elements (such as a thermistor, detection circuit, and power management circuit). One side of the sensor features a magnetic block, positive and negative contacts, and a shaped positioning protrusion; the opposite side features a magnet, positive and negative contacts, and a shaped positioning recess. The sensor also has screw holes for easy screw mounting on equipment. The sensor can interact with control nodes or other sensors, and its positioning is achieved through the shaped protrusions and recesses of the interaction module, ensuring that the positive and negative contacts of the sensors make contact and transfer electrical energy. Multiple sensors can be stacked. The positive and negative contacts on both sides of a single sensor are connected, creating a parallel power supply when multiple sensors are stacked, simultaneously powering multiple sensors. A schematic diagram of the sensor is shown below. Figure 4 and Figure 5 As shown in the diagram, a stacked configuration of multiple sensors and control nodes is illustrated. Figure 6 As shown in the diagram, multiple sensors and control nodes are stacked together. Figure 7 As shown.

[0121] The control node itself has environmental awareness capabilities and can continuously monitor the environment. When equipment monitoring is required, the sensor is removed and placed at the device location, powered by an external power supply or battery. When equipment monitoring is not required, the sensor can be stacked and stored at the control node for charging.

[0122] Sensors and control nodes are connected via Bluetooth Low Energy (BLE) Mesh networking. BLE Mesh supports star and mesh topologies. BLE Mesh has low power consumption, making it suitable for long-term deployment. The single-hop communication distance of BLE Mesh is 10-30m, which can cover the entire laboratory.

[0123] The control nodes possess data processing, communication, and alarm functions. They support data encryption, enable data sharing and collaborative analysis among nodes, and communicate via Wi-Fi Mesh to form a distributed sensing network covering multiple laboratories. Each control node is equipped with an audible and visual alarm; the alarm light uses tri-color LEDs, and the alarm volume is adjustable from 60-90dB. The control nodes can push various data to cloud platforms, large screens, or mobile terminals, and also have local caching capabilities, allowing data to be saved during network interruptions and automatically re-uploaded upon network recovery.

[0124] S13. Based on the first weight and first risk value of each target device in the control node, perform a weighted summation to obtain the second risk value of the control node.

[0125] The first risk value is the result obtained by combining the risk values ​​of various time-series data from the target device. Specifically, when there are multiple types of time-series data collected from the target device, the lightweight network model deployed in the control node outputs the risk value corresponding to each type of time-series data. The risk values ​​corresponding to various time-series data are weighted and summed to obtain the first risk value of the target device. When there is only one type of time-series data, the risk value corresponding to that time-series data is determined as the first risk value of the target device.

[0126] The formula for calculating the second risk value of the control node is as follows: , The second risk value for the control node. Let the first risk value be that of target device i. is the first weight of target device i, and N is the number of target devices in the control node.

[0127] In some embodiments, a preset second risk grading standard is obtained, and based on the second risk grading standard, a second risk level corresponding to a second risk value is determined. An alarm is then triggered based on the second risk level. The second risk level is the risk level of the control node. The second risk grading standard includes various candidate second risk levels and reference intervals for second risk values ​​corresponding to the candidate second risk levels. The process of determining the second risk level corresponding to the second risk value includes: determining the second target interval where the second risk value is located from the reference intervals for second risk values ​​corresponding to the candidate second risk levels, and determining the candidate second risk level associated with the second target interval as the second risk level corresponding to the second risk value. For example, the second candidate risk levels include three types: low risk, medium risk, and high risk. The reference interval for the second risk value of low risk is [0, 0.3], the reference interval for the second risk value of medium risk is (0.3, 0.6], and the reference interval for the second risk value of high risk is (0.6, 1]. If the second risk value is 0.5, then the second risk level corresponding to the second risk value is medium risk. Furthermore, the control node also outputs the second risk value and the second risk level and reports them.

[0128] In this embodiment, by integrating the analytic hierarchy process (AHP) with a lightweight time series model, a dynamic, weighted, and interpretable comprehensive assessment of the risks of multiple devices in the control node is achieved, which significantly improves the accuracy and practicality of risk perception while ensuring computational efficiency.

[0129] In one embodiment, step S12 includes:

[0130] Collect time-series data of the target device within a preset time period;

[0131] If the time series data has a corresponding constraint threshold, then the constraint threshold is used to standardize the time series data;

[0132] If the time series data does not have a corresponding constraint threshold, the mean and standard deviation are calculated based on the historical time series data, and the time series data is standardized based on the mean and standard deviation.

[0133] The target features are extracted from the standardized time-series data, and based on the target features, the first risk value of the target device is output using the lightweight network model deployed in the control node. The target features include one or more of the following: abnormal change rate, percentage of continuous deviation, and rated deviation.

[0134] The preset time period is a pre-defined start and end time. For example, if time-series data is collected within the past minute, then the past minute is the preset time period.

[0135] Whether time series data has a corresponding constraint threshold can be determined through querying. The constraint threshold can be preset by staff. If a constraint threshold for a certain type of time series data is preset, the corresponding constraint threshold can be found through querying, thus confirming that the time series data has a corresponding constraint threshold. If a constraint threshold for a certain type of time series data is not preset, the corresponding constraint threshold cannot be found through querying, thus confirming that the time series data does not have a corresponding constraint threshold.

[0136] The original time series data has "differences in dimensions, distribution, and thresholds", which cannot be directly used for horizontal comparison and data fusion. Therefore, constraint thresholds are used to standardize the time series data.

[0137] In some embodiments, the formula for standardizing time-series data using constraint thresholds is as follows: , It is the lower limit threshold in the constraint threshold. It is the upper limit threshold in the constraint threshold. It is the mean in the constraint threshold, where X is the time series data. norm This refers to standardized time-series data. The mean in the constraint threshold can be the mean of the historical data corresponding to this type of time-series data. Furthermore, when... or At that time, X norm Forced to be set to 1. For example, if the lower threshold of temperature is 10℃, the upper threshold is 60℃, and the average is 30℃, then 55℃ after standardization is (55-30) / (60-30)=0.83, which is directly related to the degree of risk.

[0138] In some embodiments, if the timing data is an operation sequence, then when the operation sequence is correct, X norm If the value is 0, X will be zero if the order of operations is incorrect. norm The value is 1. For example, if the reactor is required to start later than the vacuum pump, when the vibration sensor on the vacuum pump detects a jump in vibration from the "standby 0 value", the timestamp T of the jump is recorded. vacuum The current sensor in the reactor detects a jump in current from the "standby 0 value" and records the timestamp T of the jump. reactor By comparing the order of the two timestamps, the sequence of operations can be determined. If T vacuum Before T reactor This indicates that the vacuum pump starts first, followed by the reactor, which conforms to the preset sequence. At this time, X... norm It is 0.

[0139] Historical time-series data is data collected over a historical period. Historical time-series data and time-series data can be completely identical or partially identical.

[0140] When time series data does not have corresponding constraint thresholds, the formula for standardizing time series data based on the mean and standard deviation can be: X represents time-series data. norm The data are standardized time series data, where μ is the mean and σ is the variance.

[0141] The target features are one or more of the following: abnormal change rate, percentage of persistent deviation, and rated deviation. The abnormal change rate reflects the severity of temperature fluctuations per unit time and is used to identify short-term abnormal fluctuations. The percentage of persistent deviation refers to the proportion of time series data that continuously exceeds a constraint threshold, used to identify long-term abnormal states, or the ratio between the number of time series data exceeding the constraint threshold and the total number of time series data. The rated deviation is the difference between the time series data and a preset standard threshold.

[0142] The lightweight network model consists of the following layers:

[0143] Model input layer: receives target features;

[0144] Model hidden layers: Set 2-3 hidden layers, each with 16-32 neurons, using the ReLU activation function, responsible for extracting non-linear correlations between features, compressing feature dimensions, and strengthening core risk signals;

[0145] Model output layer: The Sigmoid activation function is used to force the first risk value to be mapped to the range of 0.0-1.0, which is the standardized risk value; the total number of model parameters is less than or equal to 1500, the inference speed is fast, and it can be deployed on the edge module of the device.

[0146] In some embodiments, the training process of a lightweight network model includes:

[0147] Step 1: Generate training samples. Process the regular time-series sensor data before enabling the risk assessment function into 3D feature samples. Automatically label each sample according to fixed rules and randomly assign a "virtual risk value," such as: Low risk: all feature risk values ​​are less than 0.3, and the virtual risk value is randomly selected from 0.0 to 0.3; Medium risk: the risk value of any feature is between 0.3 and 0.6, and the virtual risk value is randomly selected from 0.3 to 0.6; High risk: the risk value of any feature is greater than 0.6, and the virtual risk value is randomly selected from 0.6 to 1.0.

[0148] Step 2: Build a lightweight model. Experience replay pool: Set a maximum capacity, such as 1000 samples, and store them according to the "first-in, first-out" principle. If the capacity is exceeded, delete the oldest sample. During training, a fixed number of samples, such as 100, are randomly selected from the pool for training to ensure sample independence. Reward function: Reward and penalty rules (priority: high-risk missed detection penalty > medium-risk missed detection penalty > false positive penalty) to reduce the model's missed detection of high-risk cases. Main network and target network: The main network is responsible for updating weights in real time, receiving 3D feature input, and outputting risk values. The target network has the same structure as the main network, but its weights are not updated in real time; they are only copied from the main network periodically to avoid prediction oscillations caused by weight fluctuations during model training.

[0149] Step 3: Model training. Randomly sample from the experience replay pool, calculate the "predicted risk value" of the sample, introduce a reward and punishment mechanism, such as adding 1 point for accurate prediction and deducting 2 points for false positives / false negatives, adjust the main model parameters to make the "predicted value" close to the "target risk value", and thus obtain the trained lightweight network model.

[0150] In some embodiments, a preset first risk classification standard is obtained; based on the first risk classification standard, a first risk level corresponding to a first risk value is determined; and an alarm is triggered based on the first risk level. The first risk level is the risk level of the target device. The first risk classification standard includes various candidate first risk levels and reference intervals for first risk values ​​corresponding to the candidate first risk levels. The process of determining the first risk level corresponding to the first risk value includes: determining a first target interval where the first risk value is located from the reference intervals for first risk values ​​corresponding to the candidate first risk levels; and determining the candidate first risk levels associated with the first target interval as the first risk level corresponding to the first risk value.

[0151] In some embodiments, if the standardized time-series data exceeds a preset value, the standardized time-series data is adjusted to the preset value. For example, if the standardized temperature exceeds 1, the standardized temperature is adjusted to 1.

[0152] In some embodiments, a first risk value for the target device can be output using a lightweight network model deployed in the control node, based on standardized time-series data.

[0153] In this embodiment, adaptive standardization and domain-oriented feature extraction significantly improve the accuracy, generalization ability, and interpretability of target equipment risk assessment.

[0154] In one embodiment, if the time series data does not have a corresponding constraint threshold, the mean and standard deviation are calculated based on historical time series data, and the time series data is standardized based on the mean and standard deviation, including:

[0155] Statistical analysis of historical time-series data for a sliding time window, with the current time as the endpoint of the sliding time window;

[0156] Calculate the mean and standard deviation of historical time series data, and construct a safe reference interval based on the mean and standard deviation;

[0157] Based on the maximum and minimum values, mean and standard deviation in the safety reference interval, the time series data is standardized to obtain standardized time series data.

[0158] The standardization formula for time series data is as follows:

[0159] ;

[0160] X represents time-series data, X norm For the standardized time series data, μ is the mean, σ is the variance, μ+2σ is the maximum value, and μ-2σ is the minimum value.

[0161] The sliding time window has a configurable duration, which can be dynamically adjusted according to actual monitoring needs or data characteristics. For example, the sliding time window can be 3 months or 6 months.

[0162] When there are multiple types of time series data, the historical time series data for each type of time series data within the sliding time window are obtained separately, and the mean and standard deviation of each type of historical time series data are calculated separately. The duration of the sliding time window corresponding to each type of time series data can be the same or different.

[0163] In this embodiment, a safety reference interval is dynamically constructed and standardized through a sliding time window, achieving adaptive normalization and anomaly enhancement of time-series data. This significantly improves the accuracy, robustness, and deployment feasibility of edge-side risk assessment, while also possessing good interpretability and engineering practicality. Furthermore, dynamically updating the safety reference interval can adapt to changes in the operating environment, forming a complete safety governance closed loop.

[0164] In one embodiment, extracting target features from standardized time-series data includes:

[0165] Calculate the fluctuation amplitude of time series data within a unit time period, and determine the abnormal change rate based on the fluctuation amplitude and the first time length, where the first time length is the length of a unit time period;

[0166] The duration of statistical time-series data exceeding the constraint threshold is used to obtain the percentage of continuous deviation based on the quotient of the duration divided by the second time length, where the second time length is the length of a preset time period.

[0167] The rated deviation is determined based on the difference between the time series data and the preset standard threshold.

[0168] In this context, the fluctuation range of time-series data within a unit of time is the difference between the maximum and minimum time-series data. For example, if the temperature rises from the standardized value of 0.2 to 0.8 within one minute, where 0.2 is the minimum temperature within one minute and 0.8 is the maximum temperature within one minute, then the rate of anomalous change is 0.6 / minute. The number of anomalous change rates is the quotient of the second time length divided by the first time length.

[0169] The preset standard threshold is a manually set threshold and is an important safety indicator. For example, if the manually set standard threshold is 1, the rated deviation = 1.0 - time series data. If the time series data is 0.9 and the rated deviation is 0.1, it means "very close to the danger limit". The number of rated deviations is consistent with the number of such time series data.

[0170] Constraint thresholds include lower and upper limits. Time-series data exceeding a constraint threshold means the time-series data is greater than the upper limit or less than the lower limit. Every operation has a reasonable safe baseline range; prolonged deviation from the baseline indicates an anomaly, and the constraint threshold is the safe baseline range.

[0171] In this embodiment, by constructing a three-in-one feature system of abnormal change rate, continuous deviation ratio and rated deviation, a multi-angle, highly interpretable and low-overhead quantitative characterization of the target equipment's operational anomalies is achieved, which significantly improves the comprehensiveness, accuracy and engineering practicality of risk identification and lays a solid foundation for subsequent intelligent assessment and response.

[0172] In one embodiment, step S4 is followed by:

[0173] Obtain the preset risk classification criteria;

[0174] Based on the risk grading criteria, the global risk level corresponding to the global risk value is determined;

[0175] Based on the overall risk level, execute the corresponding response strategy.

[0176] The risk grading standard includes various candidate risk levels and corresponding risk value reference ranges. For example, the candidate risk levels include low risk, medium risk, and high risk. The risk value reference range for low risk is [0, 0.3], for medium risk it is (0.3, 0.6], and for high risk it is (0.6, 1).

[0177] The process of determining the global risk level includes: determining the target range where the global risk value is located from the risk value reference range corresponding to the candidate risk level, and determining the candidate risk level associated with the target range as the global risk level corresponding to the global risk value.

[0178] Response strategies include, but are not limited to, sampling frequency of time-series data, risk alerts, and early warning push notifications. For example, candidate risk levels include low, medium, and high risk. Low risk indicates no safety hazards or potential risks; medium risk indicates clear safety hazards; and high risk indicates serious safety risks. If the risk level is low, the response strategy is: maintain the initial sampling frequency of 5 seconds per sampling, provide a risk alert via green audio-visual aids (light constantly on, volume ≤60dB), display the risk area and abnormal parameters on the screen (font color blue), and log the information (including timestamp, sensor data, and risk level). If the risk level is medium, the response strategy is: increase the sampling frequency to 2 seconds per sampling, provide a risk alert via yellow audio-visual aids (light flashing at 1Hz, volume ≥75dB), display risk details on the screen in red font, and push early warning information to the administrator's mobile application. The pushed early warning information includes, but is not limited to, the risk area, the cause of the anomaly, and handling suggestions, and supports remote issuance of investigation instructions by administrators. If the risk level is high, the response strategy is as follows: increase the sampling frequency to 1 second / time to capture the details of sudden risks, automatically cut off the power, gas and liquid supply to relevant equipment, and support precise cutting off by area to avoid affecting non-risk areas. Risk warnings are given through red sound and light (light flashing at 2Hz, volume ≥85dB). Alarm information is pushed synchronously to the mobile application and teaching tablet. Alarm information includes but is not limited to real-time video footage and risk level. The laboratory emergency broadcast can automatically play evacuation prompts and link with the access control system to open emergency passages.

[0179] In this embodiment, by acquiring a preset risk grading standard, the global risk level corresponding to the global risk value is determined based on the risk grading standard. Therefore, corresponding response strategies can be executed based on the global risk level. Furthermore, executing corresponding response strategies based on the global risk level enables automatic handling of high-risk situations and multi-terminal alarms, while providing lightweight monitoring for low-risk situations. This approach balances security and resource efficiency.

[0180] In some embodiments, the laboratory risk assessment method further includes displaying scoring results, time-series data, equipment risk heatmaps, control node risk heatmaps, and regional heatmaps, while also supporting scaling heatmaps, filtering time-series data, and traversing time-series data back along a time axis. Specifically, the equipment risk heatmaps are generated based on the first risk value of each target equipment, the control node heatmaps are generated based on the second / third risk values ​​of each control node, and the regional heatmaps are generated based on the global risk values ​​of multiple laboratories.

[0181] In some embodiments, the laboratory risk assessment method further includes automatically labeling abnormal time-series data, recording alarm events and handling results, and supporting the retrieval of time-series data by time, device, and region; it can automatically generate a risk assessment report containing timestamps, time-series data, and global risk values. The risk assessment report can be associated with video recordings of the corresponding time period, facilitating accident tracing and teaching review.

[0182] In some embodiments, the laboratory risk assessment method also includes supporting the use of time series data for continuous optimization of reinforcement learning models, improving prediction accuracy, and enabling system self-evolution.

[0183] In a specific application scenario, laboratory risk assessment methods can be applied to Figure 8 In the laboratory risk assessment system shown, specifically, the perception layer collects sensor values ​​from each target device and the values ​​from sensors deployed in the environment. The collected sensor values ​​are transmitted to the control node via BLE Mesh. In the data processing layer, the lightweight network model deployed in the control node performs risk value prediction to obtain the first risk value of the target device, the second risk value of the control node, the global risk value of the laboratory, and the corresponding risk level. In the application layer, a graded response is performed according to the risk level, and various data, risk levels, and risk values ​​are displayed. It also supports zooming in on heatmaps, filtering time-series data, and tracing back time-series data via a time axis.

[0184] Furthermore, Figure 8 The laboratory risk assessment system shown also includes:

[0185] Power cut-off device, adapted to the power supply circuit of the laboratory or the remote control air switch of a single target device;

[0186] Gas shut-off device, a remote-controlled solenoid valve adapted to the gas supply circuit of a laboratory or the gas pipeline of a single target device;

[0187] Liquid shut-off device, a remote-controlled solenoid valve adapted to the liquid supply circuit of a laboratory or the liquid pipeline of a single target device;

[0188] Emergency broadcasting equipment, covering the laboratory's wired and wireless broadcasting systems;

[0189] Access control linkage device, electric access control locks for passageways in the laboratory.

[0190] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0191] Based on the same inventive concept, this application also provides a laboratory risk assessment device for implementing the laboratory risk assessment method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations of one or more laboratory risk assessment device embodiments provided below can be found in the limitations of the laboratory risk assessment method described above, and will not be repeated here.

[0192] In one embodiment, such as Figure 9 As shown, a laboratory risk assessment device is provided, comprising:

[0193] The second risk value determination module is used to obtain the second risk value of each control node based on the first risk value of the target device associated with each control node.

[0194] The coupling coefficient determination module is used to determine the coupling coefficient between each control node based on the physical distance and shared resources between them.

[0195] The third risk value determination module is used to update the second risk value of each control node based on the coupling coefficient, so as to obtain the third risk value of each control node.

[0196] The global risk value determination module is used to obtain the global risk value of the laboratory based on the third risk value of each control node.

[0197] The various modules in the aforementioned laboratory risk assessment device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can invoke and execute the corresponding operations of each module.

[0198] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores a first risk value, a second risk value, physical distance, shared resources, coupling coefficients, a third risk value, and a global risk value. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a laboratory risk assessment method.

[0199] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0200] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0201] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0202] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0203] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0204] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0205] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method of risk assessment of a laboratory, characterized in that, The laboratory comprises at least one control node, each of the control nodes is respectively associated with at least one target device, and the method comprises: S1, obtaining a second risk value of each control node based on a first risk value of the target device associated with each control node; S2, determining a coupling coefficient between each control node according to a physical distance and a shared resource between the control nodes; S3, updating the second risk value of each control node based on the coupling coefficient to obtain a third risk value of each control node; S4, performing weighted summation on the third risk value of each control node to obtain a global risk value of the laboratory; Step S1 comprises: S11, obtaining a plurality of scoring results of each target device in the control node, and calculating a first weight of each target device in the control node using an analytic hierarchy process based on the scoring results; S12, outputting a first risk value of the target device using a lightweight network model deployed in the control node based on time series data of the target device; S13, performing weighted summation on the first weight and the first risk value of each target device in the control node to obtain a second risk value of the control node; The calculation formula of the third risk value is: ; wherein, V3 i is the third risk value of the control node i, V2 i is the second risk value of the control node i, e ij is the coupling coefficient between the control node i and the control node j, j is not equal to i, and n is the number of control nodes.

2. The method of claim 1, wherein, Step S2 comprises: Obtaining a physical distance and a shared resource between each control node, the shared resource comprising one or more of a shared power loop and a shared air supply loop; Determining a first coupling coefficient corresponding to the physical distance and a second coupling coefficient corresponding to the shared resource; Adding the first coupling coefficient and the second coupling coefficient to obtain a coupling coefficient between each control node.

3. The method of claim 1, wherein, Step S12 comprises: Collecting time series data of the target device within a preset time period; If the time series data has a corresponding constraint threshold, normalizing the time series data using the constraint threshold; If the time series data does not have a corresponding constraint threshold, calculating a mean and a standard deviation based on historical time series data, and normalizing the time series data based on the mean and the standard deviation; Extracting target features from the normalized time series data, and outputting a first risk value of the target device using a lightweight network model deployed in the control node based on the target features, the target features comprising one or more of an abnormal change rate, a continuous deviation proportion, and a rated deviation.

4. The method of claim 3, wherein, If the time series data does not have a corresponding constraint threshold, calculating a mean and a standard deviation based on historical time series data, and normalizing the time series data based on the mean and the standard deviation, comprises: Statistically analyzing historical time series data in a sliding time window, the sliding time window ending at the current time; Calculating a mean and a standard deviation of the historical time series data, and constructing a safety reference interval based on the mean and the standard deviation; Based on the maximum and minimum values in the safety reference interval, the mean and the standard deviation, normalizing the time series data to obtain normalized time series data; The normalization formula of the time series data is: ； X is the time series data, X norm X is the time series data, X norm X is the time series data, X norm X is the time series data, X norm X is the time series data, X norm X is the time series data, X norm X is the time series data, X norm 5. The method of claim 3, wherein, Extracting target features from the normalized time series data comprises: The fluctuation amplitude of the time series data in a unit time is calculated, and an abnormal change rate is determined based on the fluctuation amplitude and a first time length, the first time length being a time length of the unit time; The duration of the time series data exceeding the constraint threshold is counted, and a persistent deviation proportion is obtained based on a quotient of the duration divided by a second time length, the second time length being a time length of the preset time period; The rated deviation is determined based on a difference between the time series data and a preset standard threshold.

6. The method of claim 1, wherein, After step S4, the method further includes: obtaining a preset risk grading standard; determining a global risk level corresponding to the global risk value based on the risk grading standard; executing a corresponding response strategy based on the global risk level.

7. A laboratory risk assessment apparatus for performing the method of any one of claims 1 to 6, characterized in that, The apparatus includes: a second risk value determination module configured to obtain a second risk value of each control node based on a first risk value of a target device associated with each control node; a coupling coefficient determination module configured to determine a coupling coefficient between each control node according to a physical distance and a shared resource between the control nodes; a third risk value determination module configured to update the second risk value of each control node based on the coupling coefficient to obtain a third risk value of each control node; a global risk value determination module configured to obtain a global risk value of the laboratory based on the third risk value of each control node.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method of any one of claims 1 to 6.

Images