Method and system for evaluating operation quality of dual-mode network based on context analysis

Abstract

The invention discloses a context analysis-based dual-mode network operation quality evaluation method and system, and relates to the technical field of electric power communication networks. The method comprises the following steps: collecting multi-dimensional operation data of an HPLC + HRF dual-mode network, and extracting context information including a network environment, an equipment state and service characteristics; constructing a communication quality-network stability-service transmission efficiency three-dimensional evaluation index system, performing dynamic weight distribution in combination with context features, and calculating a comprehensive operation quality score of the dual-mode network; and adaptively adjusting an evaluation threshold based on a context scene to realize abnormal state recognition and tracing. According to the method, the problem of poor evaluation accuracy caused by neglecting context feature relevance and single index evaluation in an existing evaluation method is solved, the real operation state of the dual-mode network in a complex power scene can be reflected more accurately, and support is provided for network optimization and troubleshooting.

Description

Technical Field

[0001] This invention relates to the field of dual-mode communication network technology in power user electricity consumption information collection systems, specifically to a method and system for evaluating the operational quality of dual-mode networks based on context analysis. Background Technology

[0002] In power user electricity consumption information collection systems, the HPLC+HRF dual-mode communication network has become the mainstream communication solution due to its combination of wide coverage of power line carrier and anti-interference capabilities of radio frequency. The quality of network operation directly determines the accuracy and real-time performance of electricity consumption information collection; therefore, its operational quality needs to be effectively evaluated.

[0003] Existing dual-mode network operation quality assessment methods rely on relatively singular evaluation metrics, mostly based solely on communication success rate or signal-to-noise ratio, without considering the interrelationships between these metrics, leading to biased assessment results. Multiple networks, high-priority service loads, and proxy device anomalies all affect the assessment results, resulting in poor accuracy. Using fixed thresholds can easily lead to misjudgments in different contexts such as peak service periods and equipment failure recovery periods. Current technologies can only simply identify abnormal network states and cannot combine contextual analysis results to pinpoint problems, resulting in weak anomaly tracing capabilities.

[0004] Therefore, there is an urgent need to construct an evaluation method that integrates context analysis to accurately map the real operating state of dual-mode networks and locate the source of anomalies. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a method for evaluating the operational quality of dual-mode networks based on context analysis. By monitoring packets to obtain network data and contextual characteristics, the method completes the evaluation of the operational quality of dual-mode networks and the tracing of anomalies. The method includes the following steps: S1: Collect multi-dimensional data from each node in the dual-mode network. The multi-dimensional data includes: By monitoring HPLC SOF frames and Select Acknowledgment frames, HPLC channel data is obtained, including SNR (Signal-to-Noise Ratio), statistical uplink and downlink communication success rates, physical block check failure rates, and phase line identification results. HRF channel data, including RSSI (Resonance Signal-to-Signal Index), number of wireless channel collisions, and discovery list reception rate, is obtained by parsing HRF beacon frames and wireless discovery list messages. Device status data, including node TEI (Terminal Equipment Identity), role, number of restarts, and online status, is obtained by monitoring heartbeat detection messages and agent change request messages. Service transmission data, including service type, link identifier, and number of message retransmissions, is obtained by parsing service message MAC frames. Network topology data, including Network Identifier (NID), neighbor network information, NID collision status, agent change frequency, and number of route repairs, is also obtained.

[0006] S2: Analyze the collected data and extract contextual features. The contextual features include: Network environment context, specifically including multi-network coexistence status, HPLC phase line status, and HRF wireless channel load; device status context, specifically including node role changes, device offline / online status, cumulative reset count, and proxy link quality; service context, specifically including service priority and message transmission type.

[0007] S3: Construct a three-dimensional evaluation index system for communication quality, network stability, and service transmission efficiency, including: Based on the preprocessed data, communication quality indicators are calculated, including HPLC communication quality, HRF communication quality, and dual-mode handover success rate; network stability indicators are calculated, including node online rate, agent change success rate, route repair response time, and multi-network coordination success rate; and service transmission efficiency indicators are calculated, including high-priority service message latency, message retransmission rate, and message loss rate. The calculation methods for these indicators are as follows: (1) Communication quality indicators: HPLC communication quality value Q HPLC =R up ×0.4+R down ×0.4+SNR×0.2, where R up R represents the HPLC ascent success rate. down This indicates the HPLC downlink success rate, and SNR represents the normalized SNR value of the HPLC channel. HRF communication quality value Q HRF =R list ×0.5+RSSI×0.3+R avoid ×0.2, where R list R represents the HRF discovery list reception rate, RSSI represents the normalized RSSI value of HRF, and R avoid Indicates the HRF channel collision avoidance rate; Dual-mode switching success rate In the formula N s-switch N represents the number of times the dual-mode handover was successfully completed. t-switch This indicates the total number of times the dual-mode switching was performed.

[0008] (2) Network stability indicators: Node online rate In the formula N active N represents the number of active nodes within a heartbeat cycle. total Indicates the total number of nodes in the network; Success rate of agent change In the formula N s-proxy N represents the number of nodes that successfully completed the proxy change. t-proxyIndicates the number of nodes that initiated the proxy change; Router repair response time In the formula, T req-i T represents the time when the i-th routing request message is sent. reply-i This indicates the time when the i-th route reply message is received, and n represents the total number of route repairs. Multi-network coordination success rate In the formula N s-conflict N represents the number of times NID collisions / channel collisions were successfully resolved. t-conflict This indicates the total number of conflicts.

[0009] (3) Service transmission efficiency indicators: High-priority service message latency In the formula, T delay-i represents the transmission delay of the i-th high-priority service message, and m represents the total number of high-priority service messages; Message retransmission rate In the formula N retrans N represents the number of retransmitted messages. t-traffic Indicates the total number of business messages; Message loss rate In the formula N loss N represents the number of lost service messages. send This indicates the total number of service messages sent.

[0010] S4: Adjust the weight allocation of evaluation indicators based on contextual characteristics, and calculate the overall operational quality score, including: The default weights of each evaluation indicator are preset, as well as the mapping rules between context features and weight compensation values; The weight compensation value is obtained based on the contextual features of the network to which it belongs, and the weights of each evaluation index are calculated in combination with the default weights. The values ​​of each evaluation indicator are normalized, and the communication quality score, network stability score, and service transmission efficiency score are calculated using a weighted summation formula. The average of the three is taken as the comprehensive operational quality score.

[0011] Communication Quality Score: (Q HPLC ×ω1+Q HRF ×ω2+Q switch ×ω3) Network stability score: (R) online ×ω4+R proxy ×ω5+T route ×ω6+R coord ×ω7) Service transmission efficiency score: (T) high ×ω8+R retrans ×ω9+R loss ×ω10 ) In the formula, ω1+ω2+ω3=1, ω4+ω5+ω6+ω7=1, ω8+ω9+ω 10 =1 The overall operational quality score is the average of the communication quality score, network stability score, and service transmission efficiency score.

[0012] Accumulate historical network quality scores and corresponding network operation data, and periodically optimize the default weights of each evaluation indicator and the mapping rules between context features and weight compensation values ​​through statistical analysis or machine learning methods.

[0013] S5: Based on the comprehensive operational quality score and the dynamic threshold corresponding to the context scenario, determine the network operating status and output the quality assessment result and anomaly tracing information. Specific steps include: Based on historical operational data, an evaluation threshold library is established for different context scenarios. The dynamic threshold for the current scenario is determined by combining the current context features with the evaluation threshold library. If the comprehensive score is greater than or equal to the dynamic threshold, the network is considered to be operating normally. If the comprehensive score is less than the dynamic threshold, it is considered to be abnormal. When it is considered abnormal, the root cause of the abnormality needs to be located by combining the context features and the sample library of the transformer area. Finally, the comprehensive evaluation results, the scores of each dimension index, the abnormality type and the source information are output.

[0014] Preferably, the evaluation threshold library records the context scenario, the specific scenario with multi-level subdivision, and the threshold coefficient corresponding to the specific scenario; the dynamic threshold is equal to the product of the threshold coefficients obtained from different context scenarios; the context scenario includes network environment, service load, proportion of high-priority services, and transformer area environment; Preferably, the transformer area sample library records at least the main scenarios, operating indicators, problem analysis and handling methods of the abnormal transformer areas; in the process of tracing the source of the anomaly, the main scenarios are first matched with context features to determine the similar transformer area sample set, and then one or more target transformer area samples are selected according to the indicator status of the current transformer area and the indicator status of the similar transformer area sample set. Finally, the problem analysis and handling methods of the target transformer area samples are statistically analyzed to give 1-3 handling opinions. The present invention also provides a dual-mode network operation quality evaluation system based on context analysis, including a data acquisition module, a context feature extraction module, an evaluation index calculation module, a context weighting module, an anomaly analysis module, a result output module, and a storage module.

[0015] The data acquisition module includes an HPLC data acquisition unit, an HRF data acquisition unit, and a device and service data acquisition unit; it acquires multi-dimensional operational data by listening to HPLC channel frames, parsing HRF beacon frames, and reading management and service messages; it can interact with the MAC sublayer and network management sublayer of the dual-mode network to obtain routing data and device status data in real time. The context feature extraction module preprocesses the collected data and extracts network environment context, device status context, and service context features based on preset rules. The evaluation index calculation module calculates evaluation indexes for communication quality, network stability, and service transmission efficiency based on the preprocessed data and according to the formula in step S3, and normalizes the index values. The context weighting module stores preset context feature-weight mapping rules; based on the features output by the context feature extraction module, it dynamically adjusts the weights of each evaluation indicator and calculates the comprehensive operational quality score using a weighted summation formula. The anomaly analysis module stores a context scenario-dynamic threshold library; it determines the operating status based on the comparison between the comprehensive score and the dynamic threshold; if an anomaly is detected, it locates the root cause of the anomaly by combining the correlation between context features and evaluation indicators. The result output module outputs the comprehensive evaluation results, scores of each dimension indicator, anomaly types and source information in a visual interface, and generates an evaluation report, which includes historical trend analysis. The storage module stores the collected raw data, contextual features, evaluation index values, comprehensive scores, historical evaluation reports, weight mapping rules, and dynamic threshold library. It uses a distributed database to store massive amounts of historical data and supports data backtracking and trend analysis.

[0016] This invention improves the accuracy of dual-mode network operation quality assessment by constructing a context feature system; it dynamically adjusts weights and thresholds based on the unique scenarios of dual-mode networks, avoiding the limitations of single indicators and fixed thresholds, and better reflecting actual operating conditions; by associating abnormal states with context features, it can quickly locate the root cause of anomalies, reducing troubleshooting time; and it can update weight rules and threshold libraries according to actual needs, providing support for operation and maintenance work. Attached Figure Description

[0017] Figure 1 This is a flowchart of the steps of the method of the present invention.

[0018] Figure 2 This is a block diagram of the system of the present invention.

[0019] Figure 3 This is a visual example of network operation quality assessment in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and a complete embodiment. It should be understood that the embodiments described herein are for illustrative purposes only and do not constitute a limitation on the scope of protection of this invention.

[0021] An HPLC+HRF dual-mode electricity information collection network for a city's power distribution area comprises one Control Center (CCO), ten Power Control Centers (PCOs), and 200 Stations (STAs). Three adjacent power distribution areas exist within this dual-mode network. Common services include low-priority standard meter reading and high-priority real-time cost control. This power distribution area employs a context-based analysis-based dual-mode network operation quality assessment method and system described in this invention to evaluate the operational quality of the dual-mode network.

[0022] like Figure 2 As shown, a dual-mode network operation quality assessment system based on context analysis includes a data acquisition module, a context feature extraction module, an evaluation index calculation module, a context weighting module, an anomaly analysis module, a result output module, and a storage module. In this embodiment, an embedded chip is used to fabricate the data acquisition module, supporting HPLC and HRF channel monitoring; a 1TB SD card is used as the storage module, supporting massive data storage; the software integrates the context feature extraction module, evaluation index calculation module, context weighting module, anomaly analysis module, and result output module, wherein the anomaly analysis module is developed using Python to support real-time data processing, and the result output module is developed using a Web front-end.

[0023] like Figure 1 As shown, the evaluation steps for the operational quality of a dual-mode network are as follows: S1: Collect multi-dimensional data from each node in the dual-mode network.

[0024] Specific data collection results include: HPLC data acquisition results obtained by monitoring SOF frames and selection confirmation frames of the HPLC channel: HPLC uplink success rate 92%, downlink success rate 88%, average SNR=18dB, 30 unknown nodes on the phase line (15%); HRF data acquisition results obtained by parsing HRF proxy beacon and wireless discovery list messages: average HRF RSSI=-65dBm, discovery list reception rate 85%, channel collisions 7 times / hour; node online rate obtained by reading heartbeat detection messages: 98%; proxy change request messages obtained: 80%; high-priority service ratio obtained: 60%; average latency: 150ms; message retransmission rate: 5%; and 3 neighboring networks and 1 NID collision obtained through network collision reporting messages.

[0025] S2: Analyze the collected data and extract contextual features.

[0026] The context features extracted in this embodiment are as follows: Network environment context features include multi-network coexistence, NID conflict, HRF channel conflict 7 times / hour, and HPLC phase line unknown ratio of 15%; Device status context features include agent node hard reset 1 time / day and node online rate of 98%; Service context features include high priority service ratio of 60% and medium service traffic.

[0027] S3: Construct a three-dimensional evaluation index system for communication quality, network stability, and service transmission efficiency.

[0028] Referring to the formula in the invention description section, the following calculation was performed: Communication quality indicators: HPLC communication quality value (0.92×0.4+0.88×0.4+0.72×0.2)=0.86, HRF communication quality value (0.85×0.5+0.6×0.3+0.4×0.2)=0.69, channel collision avoidance rate=0.4, dual-mode handover success rate=95%; Network stability metrics: Node online rate = 98%, Agent change success rate = 80%, Route repair response time = 80ms, Normalized value = 0.8, Multi-network coordination success rate = 100%; Service transmission efficiency indicators: high priority service message latency = 150ms, normalized value 0.7, message retransmission rate = 5%, normalized value 0.95, service message loss rate = 2%, normalized value 0.98.

[0029] S4: Adjust the weights of each evaluation indicator and calculate the overall operational quality score.

[0030] Since high-priority services account for 60%, the message delay weight for high-priority services is increased from 0.2 to 0.3. The HRF channel collision rate is 7 times / hour, so the HRF communication quality value weight is increased from 0.3 to 0.4. Other indicator weights remain at their default values. The specific weight allocation is as follows: ω1=0.3, ω2=0.4, ω3=0.3, ω4=0.2, ω5=0.15, ω6=0.25, ω7=0.4, ω8=0.3, ω9=0.4, ω... 10 =0.3; The overall operational quality score is calculated as follows: Communication quality score (0.86×0.3+0.69×0.4+0.95×0.3)=0.81, Network stability score (0.98×0.2+0.8×0.15+0.8×0.25+1×0.4)=0.93, Service transmission efficiency score (0.7×0.3+0.95×0.4+0.98×0.3)=0.89, Overall score 0.81+0.93+0.89=2.63, and the average value is 0.88.

[0031] S5: Quality assessment and anomaly tracing.

[0032] First, based on the current context features and the evaluation threshold library, the dynamic threshold for the current scene is determined; the table below shows an example of an evaluation threshold library.

[0033] In this embodiment, the matched context scenario is urban transformer area, multiple network coexistence, medium service load, and high-priority service ratio ≥ 50%. A specific scenario is matched against the evaluation threshold library, resulting in a dynamic threshold of (0.98 × 0.98 × 0.98 × 0.98) = 0.922. The current network operating status is determined based on the comprehensive score and the dynamic threshold. Since the comprehensive score is 0.88 < 0.922, it is judged as abnormal.

[0034] When locating the cause of an anomaly, it is necessary to match the target transformer area sample based on contextual features, and provide the preferred handling opinion based on the problem analysis and handling methods of the target transformer area sample. The table below is an example of a transformer area sample library, recording the contextual features, operating indicators, comprehensive score, problems encountered, and handling measures of the transformer area.

[0035] Serial Number Main scenarios Indicator Status Score Abnormal Problem Analysis Record S-001 Rural distribution areas + single network + high service load + high priority service ratio ≥ 70% High-load, high-priority services account for 70% of the workload; a small number of nodes experience a 5-second latency but maintain a 96% online rate. 0.75 In rural areas where nodes have long communication distances, adding relay nodes could be considered to improve the stability of the end nodes. S-002 Urban distribution area + multiple networks + high service load + high-priority service ratio <50% Multiple networks (4), 1 NID collision, 6 HRF channel collisions per hour 0.85 High HRF channel conflict can be addressed by optimizing the HRF channel configuration. S-003 Industrial distribution area + single network + medium service load + high priority service ratio <50% Mild electromagnetic interference (31-40dB), medium service load, and 6s service delay at some nodes. 0.72 Slight electromagnetic interference is affecting the PLC channel; the source of interference needs to be eliminated. S-004 Industrial distribution area + single network + low service load + high priority service ratio <50% Severe electromagnetic interference (≥41dB), a large number of nodes experiencing service delays exceeding 10s with an online rate of 90%. 0.53 Severe electromagnetic interference affects the PLC channel; the source of interference must be eliminated. In this embodiment, the target station sample matched by the station area is S-002. Compared with the target station sample, which has 6 HRF channel conflicts per hour, the HRF channel conflicts in this station area are slightly higher (7 times per hour). It is advisable to try optimizing the HRF channel configuration.

[0036] like Figure 3 As shown, the system outputs comprehensive evaluation results, scores for each dimension of indicators, anomaly information and suggestions, and the comprehensive score trend over the past 24 hours in a visual format, helping maintenance workers to quickly locate the root cause of anomalies and reduce troubleshooting time.

[0037] The specific implementation of this invention can be adjusted according to the actual power distribution area's business load, business type, and network environment; the weight mapping rules and dynamic threshold library can be optimized through machine learning, and a more accurate mapping model can be trained based on historical data to further improve the evaluation accuracy; the causes of anomalies and solutions can be gradually enriched and refined into the historical problem library as the distribution area sample library increases, thereby improving the accuracy of locating the causes and handling measures.

Claims

1. A method for evaluating the quality of operation of a dual-mode network based on context analysis, characterized in that, The method comprises the following steps: S1: Collecting multi-dimensional data of each node in the dual-mode network, wherein the multi-dimensional data comprises HPLC channel data, HRF channel data, device state data, service transmission data and network topology data; S2: Preprocessing the collected multi-dimensional data, and extracting context features, wherein the context features comprise network environment context, device state context and service context; S3: Constructing a three-dimensional evaluation index system of communication quality-network stability-service transmission efficiency, wherein the three-dimensional evaluation index comprises communication quality index, network stability index and service transmission efficiency index; The communication quality index comprises HPLC communication quality value, HRF communication quality value and dual-mode switching success rate; The network stability index comprises node online rate, proxy change success rate, route repair response time and multi-network coordination success rate; The service transmission efficiency index comprises high-priority service message delay, message retransmission rate and message loss rate; S4: Adjusting the evaluation index weight according to the context features, and respectively calculating the communication quality score, the network stability score and the service transmission efficiency score by using a weighted summation formula, and taking the average value of the three as the comprehensive operation quality score; S5: Judging the network operation state based on the comprehensive operation quality score and the dynamic threshold corresponding to the context scene, and outputting the quality evaluation result and the abnormality tracing information, specifically comprising: Establishing an evaluation threshold library under different context scenes based on historical operation data; Determining the dynamic threshold of the current scene according to the current context features and the evaluation threshold library; If the comprehensive score is greater than or equal to the dynamic threshold, it is determined that the network operation is normal; if the comprehensive score is less than the dynamic threshold, it is determined that the network operation is abnormal; When it is determined that the network operation is abnormal, the abnormal root cause needs to be located in combination with the context features and the transformer area sample library; Finally, the comprehensive evaluation result, the dimension index score, the abnormal type and the tracing information are outputted.

2. The method of claim 1, wherein, In step S1, the HPLC channel data comprises the signal-to-noise ratio (SNR) of the HPLC channel, the uplink and downlink communication success rate, the phase line identification result and the physical block verification failure rate; the HRF channel data comprises the received signal strength indicator (RSSI) of the HRF channel, the wireless channel conflict frequency and the wireless discovery list receiving rate; the device state data comprises the node TEI, the proxy role, the restart frequency and the active frequency in the heartbeat detection period; the service transmission data comprises the service type, the link identifier and the message retransmission frequency; and the network topology data comprises the network identifier (NID), the neighbor network information, the NID conflict state, the proxy change frequency and the route repair frequency. In step S2, the network environment context comprises the multi-network coexistence state, the HPLC phase line state and the HRF wireless channel load; the device state context comprises the node role change, the device offline / online state, the reset accumulation frequency and the proxy link quality; and the service context comprises the service priority and the message transmission type.

3. The method of claim 1, wherein, In step S3, the calculation method of the evaluation index is as follows:

4. The method of claim 1, wherein, In step S4, the adjustment of the evaluation index weight according to the context features comprises: HPLC communication quality value Q HPLC = R up × 0.4 + R down × 0.4 + SNR x 0.2, where R up represents the HPLC uplink success rate, R down represents the HPLC downlink success rate, and SNR represents the HPLC channel SNR signal-to-noise ratio normalized value; HRF communication quality value Q HRF = R list × 0.5 + RSSI x 0.3 + R avoid x 0.2, where R list represents the HRF discovery list reception rate, RSSI represents the RSSI normalized value of the HRF, R avoid represents the HRF channel collision avoidance rate; Dual-mode handover success rate where N s-switch represents the number of times a dual-mode handover is successfully completed, N t-switch represents the total number of dual-mode handovers; Node online rate , where N active represents the number of active nodes in a heartbeat period, N total represents the total number of nodes in the network; proxy change success rate , where N s-proxy represents the number of nodes that successfully complete the proxy change, N t-proxy represents the number of nodes that initiate the proxy change; Route repair response time where T req-i represents the time of the i-th route request message sending, T reply-i represents the time of the i-th route reply message receiving, and n represents the total number of route repairs. Multi-network coordination success rate , where N s-conflict represents the number of times that NID conflict / channel conflict is successfully resolved, N t-conflict represents the total number of conflicts; High-priority service packet delay where T delay-i denotes the transmission delay of the i-th high-priority service packet, and m denotes the total number of high-priority service packets. Packet retransmission rate where N retrans represents the number of retransmitted packets, N t-traffic represents the total number of service packets; packet loss rate where N loss denotes the number of lost service packets, N send denotes the total number of transmitted service packets.

5. The method of claim 1, wherein, Predefining the default weight of each evaluation index and the mapping rule of the context features and the weight compensation value. ​ The weight compensation value is obtained based on the context characteristics of the network, and the weight of each evaluation index is calculated in combination with a default weight; The evaluation index values are normalized, and the communication quality score, the network stability score, and the service transmission efficiency score are calculated by using a weighted summation formula, and the average of the three scores is taken as the comprehensive operation quality score; The communication quality score is (Q HPLC x ω1 + Q HRF x ω2 + Q switch x ω3); The network stability score is (R online x ω4 + R proxy x ω5 + T route x ω6 + R coord x ω7); The service transmission efficiency score is (T high x ω8 + R retrans x ω9 + R loss x ω 10 ); where ω1+ω2+ω3=1, ω4+ω5+ω6+ω7=1, ω8+ω9+ω 10 =1; The comprehensive operation quality score is the average of the communication quality score, the network stability score, and the service transmission efficiency score; The historical network quality scores and corresponding network operation data are accumulated, and the default weights of the evaluation indexes and the mapping rules of the context characteristics and the weight compensation values are periodically optimized by statistical analysis or machine learning methods.

6. The method of claim 1, wherein, In step S5: The evaluation threshold library records the context scenarios, the specific scenarios of multi-level subdivision, and the threshold coefficients corresponding to the specific scenarios; The dynamic threshold is equal to the product of the threshold coefficients obtained in different context scenarios; The context scenarios include network environment, service load, proportion of high-priority services, and transformer area environment; The transformer area sample library at least records the main scenarios, operation indexes, problem analysis, and processing methods of abnormal transformer areas; in the abnormal source tracing process, the main scenarios are first matched by using the context characteristics, the similar transformer area sample set is determined, one or more target transformer area samples are selected according to the index conditions of the current transformer area and the index conditions of the similar transformer area sample set, and finally the problem analysis and processing methods of the target transformer area samples are counted to give 1-3 processing suggestions.

7. A dual-mode network operation quality evaluation system based on context analysis, characterized by, The system comprises a data acquisition module, a context characteristic extraction module, an evaluation index calculation module, a context weighting module, an abnormality analysis module, a result output module, and a storage module; The data acquisition module comprises an HPLC data acquisition unit, an HRF data acquisition unit, and a device and service data acquisition unit; multi-dimensional operation data are acquired by listening to HPLC channel frames, analyzing HRF beacon frames, and reading management and service messages; the MAC sublayer and the network management sublayer of the dual-mode network can be interacted with to acquire route data and device state data in real time; The context characteristic extraction module pre-processes the acquired data, and extracts network environment context, device state context, and service context characteristics based on preset rules; The evaluation index calculation module calculates the evaluation indexes of the communication quality, the network stability, and the service transmission efficiency dimensions based on the pre-processed data according to the formula in step S3, and normalizes the index values; The context weighting module stores the preset context characteristic-weight mapping rules; the weights of the evaluation indexes are dynamically adjusted according to the characteristics output by the context characteristic extraction module, and the comprehensive operation quality score is calculated by using a weighted summation formula; The abnormality analysis module stores the context scenario-dynamic threshold library; the running state is determined based on the comparison between the comprehensive score and the dynamic threshold; If it is abnormal, the abnormal root cause is located in combination with the correlation between the context characteristics and the evaluation indexes; The result output module outputs the comprehensive evaluation result, the index scores of each dimension, the abnormal type, and the source tracing information in a visual interface, generates an evaluation report, and includes historical trend analysis. The storage module stores the collected original data, context features, evaluation index values, comprehensive scores, historical evaluation reports, weight mapping rules and dynamic threshold libraries, and adopts a distributed database to store massive historical data, and supports data backtracking and trend analysis.

Images