Water conservancy project management method and system

Abstract

The application discloses a kind of water affair water conservancy engineering management method and system, it is related to the technical field of engineering management, the method comprises: in response to water quality anomaly warning, the operation behavior of target user in front-end window is monitored, and session object is obtained;Pollution traceability index, parameter correlation index, analysis stagnation index and research and judgment convergence index are determined according to session object;Pollution traceability index, parameter correlation index, analysis stagnation index and research and judgment convergence index are compared with the threshold value corresponding to the index in turn;The index corresponding to the threshold value that the index is first exceeded is determined as target index;The push of front-end window is determined according to target index.This application quantifies the behavior characteristics of target user in each stage by index, to accurately determine the cognitive stage where target user is located, determines the push of front-end window according to the cognitive stage where target user is located, to provide appropriate intelligent assistance to target user, so as to improve the analysis and processing efficiency of water quality anomaly.

Description

Technical Field

[0001] This invention relates to the technical field of engineering management, and in particular to a method and system for managing water conservancy projects. Background Technology

[0002] Operation and maintenance of water conservancy projects is an important part of ensuring the normal operation and continuous performance of water facilities. Key aspects of water project management and operation and maintenance include equipment maintenance and repair, water quality monitoring and treatment, pipeline management and maintenance, emergency response and disaster management, energy management and energy conservation and emission reduction, etc.

[0003] When monitoring water quality, if maintenance personnel identify an anomaly, they need to compare a large amount of data to analyze the anomaly. The retrieval process during the analysis requires a lot of effort from the maintenance personnel, and when analyzing based on experience and knowledge, they need to check a large amount of data to verify their inferences. As a result, the analysis and handling of water quality anomalies in the management of water conservancy projects is relatively slow.

[0004] Therefore, improving the efficiency of analyzing and handling abnormal water quality situations during the management of water conservancy projects has become an urgent technical problem to be solved. Summary of the Invention

[0005] The technical problem solved by this invention is that the analysis and handling of abnormal water quality situations are relatively slow in the process of water conservancy project management.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a water conservancy project management method, comprising: In response to water quality anomaly warnings, monitor the target user's actions in the front-end window and obtain the session object; Based on the conversation object, determine the pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and judgment convergence indicators; The pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and judgment convergence indicators are compared with the thresholds corresponding to the indicators in turn; The indicator that first exceeds the threshold of the indicator is identified as the target indicator; The push of the front-end window is determined based on the target metrics.

[0007] Preferably, in response to a water quality anomaly warning, the system monitors the target user's actions in the front-end window to obtain a session object, including: In response to a water quality anomaly warning, listen for front-end events and obtain the original event object; Parse the operation object and operation type from the original event object to obtain multiple behavior atomic objects; Determine the timestamp of each behavior atomic object; Adjacent behavior atomic objects whose timestamp difference is greater than a preset threshold are identified as divisible behavior atomic objects. Divide the divisible behavior atomic object to obtain the session object.

[0008] Preferably, pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and analysis convergence indicators are determined based on the session object, including: The exploration parameters are determined based on the session object. These parameters are used to measure the consistency between the target user's exploration direction in the spatial path and the expected direction of the upstream flow. The exploration length is determined based on the session object, and the exploration length is used to characterize the span of the target user's exploration path; The exploration density is determined based on the conversation object. The exploration density is used to measure the orderliness of the target user's exploration in the spatial path. Pollution source tracing indicators are determined based on exploration parameters, exploration length, and exploration density. Among them, pollution source tracing indicators are positively correlated with exploration parameters, negatively correlated with exploration length, and positively correlated with exploration density. Based on the session object, determine the parameter correlation indicators, analyze the stagnation indicators, and judge the convergence indicators; The push notifications for the front-end window are determined based on the target metrics, including: If the pollution source tracing index is greater than the preset pollution source tracing threshold, the corresponding station in the water quality anomaly warning will be identified as an abnormal station. The front-end window pushes links to compare multiple parameters of multiple sites, including abnormal sites and preset sites before the abnormal sites.

[0009] Preferably, the exploration parameters are determined based on the session object, including: Extract multiple sites from the session object to obtain a site sequence formed according to chronological order; A movement vector is formed based on the movement direction corresponding to two adjacent stations in the station sequence; Determine the average vector among multiple movement vectors; The starting and ending points of the water flow are determined from the station sequence based on the actual direction of the water flow. The vector formed by pointing from the end point of the water flow to the beginning point of the water flow is determined as the reverse flow vector; The cosine similarity between the average vector and the reverse flow vector is determined as the exploration parameter.

[0010] Preferably, determining the probe density based on the session object includes: The multiple actual stations through which the water flows from the end point to the beginning point of the water flow are determined as the reverse water flow sequence, which is the opposite of the actual water flow direction. Extract multiple character-type subsequences between the upstream flow sequence and the station sequence. The character-type subsequence is a sequence formed by multiple elements that exist simultaneously in both the upstream flow sequence and the station sequence and have the same sequential order. Mutually exclusive character subsequences among multiple character subsequences are defined as subsequence pairs, where the number of elements in each subsequence of the two character subsequence pairs is greater than 1; The subsequence pair with the largest total number of elements among multiple subsequence pairs is identified as the target subsequence pair; The search density is determined by the ratio of the sum of the number of elements in the target subsequence pairs to the number of elements in the site sequence.

[0011] Preferably, determining the probe length based on the session object includes: The search length is determined by the sum of the magnitudes of multiple movement vectors in the station sequence. Before identifying the target subsequence pair as the one with the largest total number of elements among multiple subsequence pairs, the method also includes: If a character subsequence does not form a subsequence pair, then an empty sequence is configured to form a subsequence pair with the character subsequence.

[0012] Preferably, the parameter correlation indicators, stagnation indicators, and convergence indicators are determined based on the session object, including: Extract the operation objects corresponding to the water quality influencing factors of the session object to obtain the sequence of influencing objects; Determine whether there are multiple elements alternating in the sequence of affected objects; If multiple elements appear alternately, then the multiple alternating elements are identified as related elements; Determine the first occurrence number of each relevant element in the sequence of affected objects; Determine the second number of times that multiple related elements alternate in the sequence of affected objects; The sum of multiple first numbers is obtained, and the ratio between the second number and the sum of multiple first numbers is determined as the parameter correlation index. Based on the conversation partner, determine the stagnation indicators and convergence indicators for analysis; Determining the push of the front-end window based on target metrics also includes: If the pollution source tracing index is less than or equal to the preset pollution source tracing threshold, then determine whether the parameter association index is greater than the preset parameter association threshold. If the parameter correlation index is greater than the preset parameter correlation threshold, an overlay comparison chart of multiple related elements is generated, and the correlation coefficient between the multiple related elements is determined. A jump link is pushed to the front-end window, displaying an overlay comparison chart and correlation coefficient.

[0013] Preferably, the analysis stagnation index and the convergence index are determined based on the conversation object, including: Extract operation objects of type search from the session object to obtain multiple search objects; Determine the search entropy of the session object based on the information entropy of multiple search objects; Atomic objects of the same operation type that are consecutive in time within a session object are identified as consecutive operation objects; Determine the standard deviation of the time interval between multiple consecutive operation objects based on the timestamp of the behavior atomic object; Operational quality is determined by standard deviation, and operational quality is negatively correlated with standard deviation. Extract all manipulated water quality parameters from the session object to obtain a parameter set; The information entropy of the parameter set is determined as the parameter entropy; The focus rate is obtained by taking the sum of the number of operations of the N most frequently operated parameters in the parameter set according to the session object, and the ratio of the sum to the total number of operations of the parameters in the parameter set. The hypothetical convergence is determined based on parameter entropy and focusing rate. The hypothetical convergence is negatively correlated with parameter entropy and positively correlated with focusing rate. The analysis stagnation index is determined based on search entropy, operation quality, parameter correlation index, and hypothesis convergence. The analysis stagnation index is positively correlated with search entropy, negatively correlated with operation quality, negatively correlated with parameter correlation index, and negatively correlated with hypothesis convergence. Determine the convergence index based on the conversation partner; Determining the push of the front-end window based on target metrics also includes: If the parameter correlation index is less than or equal to the preset parameter correlation threshold, then determine whether the analysis stagnation index is greater than the preset analysis stagnation threshold. If the analysis stagnation index is greater than the preset analysis stagnation threshold, then obtain the external events and abnormal parameters when the water quality anomaly warning occurs; Push jump links for external events and exception parameters to the front-end window.

[0014] Preferably, the convergence index is determined based on the conversation object, including: Extract the operation type from the session object; Determine if a preset end type exists in the operation type; If a preset end type exists in the operation type, the session object will be divided into an early session object and a late session object based on the timestamp. Classify the objects of operation according to the type of operation; Based on the categorized operation objects, determine the early information entropy corresponding to each operation type in the early session objects and the late information entropy corresponding to each operation type in the later session objects; The ratio between the sum of multiple previous information entropies and the sum of multiple later information entropies is determined as the convergence criterion. Determining the push of the front-end window based on target metrics also includes: If the analysis stagnation index is less than or equal to the preset analysis stagnation threshold, then determine whether the analysis convergence index is greater than the preset analysis convergence threshold. If the convergence index is judged to be greater than the preset convergence threshold, a draft report will be generated based on multiple session objects. A link to the report draft is pushed to the front-end window.

[0015] On the other hand, this application also provides a water resources and water conservancy project management system for use in any of the above methods.

[0016] The beneficial effects of this invention are as follows: In response to water quality anomaly warnings, the system monitors the target user's actions in the front-end window to obtain a session object. Based on the session object, it determines pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and judgment convergence indicators. These indicators are then compared sequentially with their corresponding thresholds. The indicator that first exceeds the threshold is identified as the target indicator. This follows the classic workflow of water experts, from anomaly detection to pollution source tracing, correlation verification between pollution parameters, and attribution decision-making. By quantifying the target user's behavioral characteristics at each stage, the system accurately determines the target user's cognitive stage and determines the front-end window push accordingly, providing appropriate intelligent assistance and improving the efficiency of analyzing and handling water quality anomalies in water conservancy project management. Attached Figure Description

[0017] Figure 1 This is a basic flowchart of a water resources and water conservancy project management method and system provided in one embodiment of the present invention. Detailed Implementation

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0019] Example 1, referring to Figure 1 As an embodiment of the present invention, a water resources and water conservancy project management method is provided. This method operates within a water resources and water conservancy project management system and includes: S110, in response to a water quality anomaly warning, monitors the target user's actions in the front-end window and obtains the session object.

[0020] S120, determine pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and judgment convergence indicators based on the session object.

[0021] S130 compares the pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and judgment convergence indicators with the thresholds corresponding to those indicators in sequence.

[0022] S140, the indicator that first exceeds the threshold of the indicator is identified as the target indicator.

[0023] S150 determines the push of the front-end window based on the target indicators.

[0024] The object of operation refers to the specific data object that the user interaction is directed to. The object of operation is the most direct reflection of the user's focus of thought.

[0025] Operation type refers to the abstract category of the action a user performs on an object. Operation type is the cornerstone for understanding user intent (e.g., selection is viewing, comparison is verifying relationships).

[0026] Operational objectives refer to the macro-level goals that are hoped to be achieved behind an operation. Operational objectives are key to judging analytical strategies (such as tracing the source and drilling down).

[0027] The spatiotemporal context parameter timestamp refers to the precise moment when the operation occurs. The spatiotemporal context parameter timestamp is the cornerstone of all timing calculations and is used to divide sessions, calculate duration and rhythm. In practice, the spatiotemporal context parameter timestamp is used as an attribute of the operation object.

[0028] Operation duration refers to the time a user spends on a particular view or object. Operation duration is a direct indicator of attention and depth of thought.

[0029] Spatial context refers to the spatial hierarchy and location of the entity corresponding to the currently operated object. Users can switch spatial hierarchy by zooming in and out of the map, and switch spatial location by dragging the mouse and selecting objects.

[0030] Spatial path refers to the sequence of stations explored by a user in geographic space. It intuitively reflects the user's spatial reasoning logic and pollution source tracing hypothesis. In practice, the stations in geographic space are used as attributes of the operation object. The process of determining the spatial path is as follows: read the attributes of the operation object, arrange the stations corresponding to the operation object according to the order of the operation object in the session window, and obtain the access sequence of the stations, that is, the spatial path.

[0031] The sequence parameter operation order describes the order in which a series of operation atoms occur. The sequence parameter operation order is used to characterize the user's internal analysis logic and is the key to distinguishing between ordered exploration and disordered blind points.

[0032] To provide a more detailed explanation of the above process, preferably, S110 includes sub-steps S111 to S115: S111 responds to a water quality anomaly warning by listening to front-end events and obtaining the original event object.

[0033] Listen for events triggered by the user in the foreground window (i.e., foreground events). Specifically, bind event listeners to UI components (buttons, maps, sliders) to monitor user actions. The listener output is a raw, low-level browser event object. For example, `{ type: 'click', target: $('#ammonia-button'), clientX:123, clientY: 456, timestamp: 1620000000000}`, `{ type: 'dragend', target: $('#time-slider'), value: ['09:00', '11:00'], timestamp: 1620000001000}`.

[0034] S112, extract the operation object and operation type from the original event object to obtain multiple behavior atomic objects.

[0035] By writing a parsing function, the raw events are mapped to predefined behavior atomic objects. The output of the parsing function is a standardized behavior atomic object. Specifically, the parsing function includes parsing the operation object and the operation type. For example, parsing `event.target` to determine the operation object; for example, if the clicked button ID is `ammonia-button`, it is mapped to the parameter ID: "ammonia". For the operation type, for example, based on `event.type` and the context, the operation type is determined; a `dragend` event on the timeline component is mapped to the operation type: "scaling".

[0036] Finally, the parsing function outputs the mapping results, yielding multiple behavior atomic objects, which are the behavior atomic flows: {action: 'Select', entity: { type: 'Parameter', id: 'ammonia'}, timestamp:1620000000000}, {action: 'Zoom', entity: { type: 'TimeRange', value: ['09:00', '11:00']}, timestamp: 1620000001000}.

[0037] S113, determine the timestamp of each behavior atomic object.

[0038] The timestamp, also known as the timestamp of the spatiotemporal context parameter mentioned above.

[0039] S114, Identify adjacent behavior atomic objects whose timestamp difference is greater than a preset threshold as divisible behavior atomic objects.

[0040] S115, divide the divisible behavior atomic object to obtain the session object.

[0041] The session division includes: receiving input behavior atomic objects; the system maintains a current session object and a last operation timestamp; when a new behavior atomic arrives, it checks the difference Δt between its timestamp and the last operation time; if Δt is greater than a preset threshold (60s), the current session is marked as complete and proceeds to the subsequent process, and a new session object is created, with the new atomic added to it; if Δt is less than or equal to the preset threshold (60s), the new behavior atomic is added to the current session, and the last operation timestamp is updated to the current atomic timestamp; finally, an independent session object is output.

[0042] Before proceeding to the next step, the system also performs time-series analysis, statistical aggregation analysis, spatial analysis, and semantic analysis based on all the atomic behavioral objects in a session object.

[0043] Temporal analysis is used to extract temporal features from a session object, including operation intensity, operation rhythm, and focus.

[0044] Statistical aggregation analysis is used to extract aggregate features from a session object, including frequency and co-occurrence of associations.

[0045] Spatial analysis is used to extract spatial features from a session object, including spatial path vectors and spatial hierarchy changes.

[0046] Semantic analysis is used to extract search features from a session object. These search features include semantic relevance between operation objects whose operation type is search.

[0047] Operation intensity = (Total number of specific operation types within the session) / (Total session duration). For example, to calculate the intensity of the "Compare" operation, if the user performs Compare 5 times within the session and the session duration is 120 seconds, then the comparison intensity = 5 / 120 ≈ 0.042 times / second.

[0048] The operation rhythm is calculated by setting the standard deviation of the time intervals between all consecutive operations within a session. For example, an interval of [1.2s, 0.8s, 1.1s, 0.9s] with a small standard deviation indicates a stable and rapid rhythm, suggesting a strong intent. An interval of [0.5s, 10.2s, 0.3s, 30.1s] with a large standard deviation indicates a chaotic rhythm, suggesting the user may be hesitant or distracted.

[0049] Attention span is calculated by summing the total time a user spends on each manipulated object (such as a parameter). For example, if a user spends time on the ammonia nitrogen parameter three times (2s, 10s, 5s), the total attention span is 17 seconds.

[0050] Frequency / repetition is counted by grouping all behavior atomic objects in the session by (operation type, operation object). For example, (Compare, (ammonia nitrogen, dissolved oxygen)) appeared 4 times, and (Select, ammonia nitrogen) appeared 7 times.

[0051] Co-occurrence is calculated using a modified Jaccard index to determine the strength of co-occurrence of two operands (e.g., parameters A and B) within the same session window. The formula is: P_Link(A, B) = (Count_Session(A∩ B)) / (Count_Session(A ∪ B)), where A ∩ B represents the number of times the parameter pair formed by parameters A and B co-occurs within a session, and A ∪ B represents the number of times either A or B occurs. This formula calculates the degree of correlation between parameters A and B in the current session. Parameter co-occurrence characterizes the user's confidence in the relationship between two parameters; a high value indicates that the user is more likely to analyze A and B together, meaning the user strongly suspects a relationship between A and B.

[0052] The spatial path vector is a vector that converts the order in which a user operates on a site into a single vector. For example, if the operation order is [Site_A, Site_B, Site_C], the resulting path vector is [A, B, C].

[0053] The spatial hierarchy change is the difference between the final zoom level and the initial zoom level. A positive value indicates zooming in (focusing on details), while a negative value indicates zooming out (viewing the whole picture).

[0054] Semantic relevance is the information entropy of the objects corresponding to all search behaviors in a session, using the formula H(X) = -Σ p(x_i) * log2(p(x_i)), where p(x_i) is the frequency of a keyword appearing in all search terms in this session. For example, a user searched for [algae, dissolved oxygen, algae, water temperature]. There were 4 searches and 3 terms. p(algae) = 2 / 4 = 0.5; p(dissolved oxygen) = 1 / 4 = 0.25; p(water temperature) = 1 / 4 = 0.25; H(X) = - [0.5 * log2(0.5) + 0.25 * log2(0.25) + 0.25 * log2(0.25)] ≈ 1.5. A higher entropy value indicates a more dispersed search and a more divergent user's thinking.

[0055] Anomaly awareness indicators are used to reflect the user's level of awareness of abnormal situations. Anomaly awareness indicators include pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and judgment convergence indicators.

[0056] Water is fluid, and pollutants spread with the flow, making pollution source tracing the primary task in water quality anomaly analysis. If upstream pollution is suspected, the user is in the initial awareness stage (anomaly confirmation stage). Specific behaviors during this stage include the user tracing upstream from the downstream anomaly point to track the pollution source, and the user will purposefully and quickly access upstream data stations for comparison.

[0057] Based on this, the process for determining pollution source tracing indicators includes S210~S240: S210, determine the exploration parameters based on the session object. The exploration parameters are used to measure the consistency between the target user's exploration direction in the spatial path and the expected reverse flow direction.

[0058] Preferably, S210 includes sub-steps S211 to S216: S211, extract multiple sites from the session object to obtain a site sequence formed according to the chronological order.

[0059] The site sequence is a sequence of multiple sites along a spatial path corresponding to a user's operation, ordered from first to last according to their timestamps.

[0060] S212, form a movement vector based on the movement directions of two adjacent stations in the station sequence.

[0061] For example, if the station sequence is [a,b,d,c], where a~c are all stations, then there are movement vectors ab, bd, and dc. The direction of movement vector ab is the direction from station a to station b on the front-end map. The direction of movement vector ab is the actual straight-line distance between station a and station b, that is, the distance between the corresponding coordinates of station a and the corresponding coordinates of station b.

[0062] S213, determine the average vector among multiple movement vectors.

[0063] The average vector is also known as the mean of a vector.

[0064] S214, determine the starting point and ending point of the water flow from the station sequence based on the actual water flow direction.

[0065] For example, if the station sequence is [a,b,d,c], and the water flow direction is dcba and the counter-current direction is abcd, then the counter-current sequence is [a,b,c,d], with the starting point of the water flow being d and the ending point being a.

[0066] S215, the vector formed by pointing from the end point of the water flow to the beginning point of the water flow is determined as the reverse water flow vector.

[0067] For example, if the station sequence is [a,b,d,c] and the water flow direction is dcba, then the reverse water flow vector is vector ad.

[0068] S216, the cosine similarity between the average vector and the reverse flow vector is determined as the exploration parameter.

[0069] The exploration parameter is used to measure the consistency between the user's exploration direction in the spatial path and the expected reverse flow direction. If the user wants to track the pollution source, it will trigger the action of repeatedly exploring against the flow direction. That is, starting from the abnormal site, click the mouse and drag to the upstream site of the abnormal site. Correspondingly, at this time, the parameter value of the exploration parameter will be very close to 1.

[0070] S220, determine the exploration length based on the session object. The exploration length is used to characterize the span of the target user's exploration path.

[0071] Preferably, S220 includes sub-step S221, which determines the search length by the sum of the magnitudes of the multiple movement vectors of the station sequence.

[0072] The exploration length is the sum of the magnitudes of multiple movement vectors in the station sequence. The exploration length is used to characterize the span of the user's exploration path. The longer the exploration length, the more it represents the uncertainty of the user's exploration.

[0073] S230, determine the exploration density based on the session object. The exploration density is used to measure the orderliness of the target user's exploration in the spatial path.

[0074] Preferably, S230 includes sub-steps S231 to S235: S231, the multiple actual stations through which the water flows from the end point to the beginning point of the water flow are determined as the reverse water flow sequence, in the opposite direction of the actual water flow.

[0075] For example, if the station sequence is [a,b,d,c], and the water flow direction is dcba and the counter-current direction is abcd, then the counter-current sequence is [a,b,c,d], with the starting point of the water flow being d and the ending point being a.

[0076] S232, extract multiple character-type subsequences between the upstream flow sequence and the station sequence, wherein the character-type subsequence is a sequence formed by multiple elements that exist simultaneously in the upstream flow sequence and the station sequence and have the same sequential order.

[0077] S233, mutually exclusive character subsequences among multiple character subsequences are determined as subsequence pairs. The number of elements in either subsequence of the two character subsequences of the subsequence pair is greater than 1. If the character subsequences do not form a subsequence pair, an empty sequence is configured to form a subsequence pair with the character subsequence.

[0078] S234, the sub-sequence pair with the largest total number of elements among multiple sub-sequence pairs is determined as the target sub-sequence pair.

[0079] S235, the search density is determined by the ratio between the sum of the number of elements in the target subsequence pairs and the number of elements in the site sequence.

[0080] Exploration density measures the orderliness of a user's exploration along a spatial path. The more orderly the exploration, the more likely the user will click on upstream stations in the opposite direction of the water flow to view detailed information. Therefore, the more orderly the exploration, the greater the exploration density, and the more certain it is that the user is in a source tracing state.

[0081] For example, the character subsequences between the station sequence [a,b,d,c] and the countercurrent sequence [a,b,c,d] are [a], [b], [c], [d], [a,b], [a,b,c]. [a,b,c] and [d] are mutually exclusive. However, [d] has only one element, so [a,b,c] forms a subsequence pair with the empty sequence. [a,b] is mutually exclusive with both [c] and [d]. However, [c] and [d] both have only one element, so [a,b] forms a subsequence pair with the empty sequence. In subsequence pair 1 ([a,b,c] forms with the empty sequence) and subsequence pair 2 ([a,b] forms with the empty sequence), subsequence pair 1 contains 3 elements, and subsequence pair 2 contains 2 elements. Therefore, the target subsequence pair is subsequence 1, and the search density is 3 / 4.

[0082] S240. Pollution source tracing indicators are determined based on exploration parameters, exploration length, and exploration density. Among them, pollution source tracing indicators are positively correlated with exploration parameters, negatively correlated with exploration length, and positively correlated with exploration density.

[0083] For example, pollution source tracing index = exploration parameter * (1 / exploration length) * exploration density.

[0084] It is important to note that the pollution source tracing indicators in this application are not directly used to guide the scope of pollution source tracing investigations in the physical world, but rather to assess the cognitive state and operational efficiency of target users (operation and maintenance personnel) when operating the software system front end. Their core purpose is to determine whether users are conducting effective and orderly source tracing analysis.

[0085] Therefore, within the context of this system, an excessively long and disordered sequence of site clicks (i.e., the longer the search length), the more likely it is that the user is trying aimlessly, without having formed a clear source tracing hypothesis, resulting in low operational efficiency. Therefore, setting the search length to a negative correlation with pollution source tracing indicators aligns with human-computer interaction cognitive logic and helps identify this inefficient exploratory behavior.

[0086] When the system identifies that the target user is in an efficient and directional tracing state through pollution source tracing indicators (at which point the indicator value is high, the tracing length is relatively short but orderly), it will proactively provide intelligent assistance such as multi-parameter comparison of upstream stations to help the target user efficiently expand the tracking range.

[0087] After a user identifies an anomaly, the cause is typically analyzed, which involves verifying the causal relationship between parameters. For example, does an increase in ammonia nitrogen lead to a decrease in dissolved oxygen? Strong chemical and physical correlations exist between water quality parameters (e.g., COD and BOD5, ammonia nitrogen and dissolved oxygen), and analyzing these correlations is crucial for determining the type and extent of pollution. Experts will repeatedly and alternately examine the curves of two or more parameters, observing whether their trends match (both rising or falling simultaneously, or one increasing while the other decreases), to verify their hypotheses. In other words, if parameters are repeatedly compared, it indicates that the user is in the second stage (causal analysis stage). Therefore, it's essential to first determine if the user has entered the correlation comparison mode, and then determine the co-occurrence degree of correlation between parameters as the parameter correlation index.

[0088] Specifically, the process of determining the parameter correlation index includes S310~S360: S310, extract the operation objects corresponding to the water quality influencing factors of the session object to obtain the sequence of influencing objects.

[0089] S320, Determine whether there are multiple elements alternating in the sequence of affected objects.

[0090] S330, if there are multiple elements that appear alternately, then the multiple elements that appear alternately are identified as related elements.

[0091] S340, determine the first occurrence number of each related element in the sequence of affected objects.

[0092] S350, determine the second number of times that multiple related elements alternate in the sequence of affected objects.

[0093] S360 sums up multiple first numbers to obtain a sum value, and the ratio between the second number and the sum value formed by multiple first numbers is determined as the parameter correlation index.

[0094] Throughout the entire operation sequence, count the number of times operation object A is followed by operation object B (Count(A->B)) and the number of times operation object B is followed by operation object A (Count(B->A)). This yields the second number of alternating occurrences. The number of occurrences of multiple operation objects A forms the first number of occurrences of operation object A (Count(A)) and the number of occurrences of multiple operation objects B forms the first number of occurrences of operation object B (Count(B)). The parameter correlation index is [Count(A->B)+ Count(B->A)] / [Count(A)+ Count(B)].

[0095] Anomalies often stem from complex causes, sometimes involving a combination of factors (e.g., endogenous pollution + exogenous input + changes in hydrological conditions). In such cases, analysis can easily stall. When conventional approaches are blocked, experts may exhibit inefficient behaviors such as constantly changing search keywords, aimlessly clicking on different parameters and sites, and repeatedly viewing the same data, leading to erroneous exploration. Therefore, identifying these analytical stall characteristics can help determine if a user is in a state of analytical stagnation.

[0096] Specifically, the process for determining the stagnation index includes steps S410 to S460: S410, extract the operation objects of type search from the session object to obtain multiple search objects.

[0097] S420, determine the search entropy of the session object based on the information entropy of multiple search objects.

[0098] In other words, search entropy is the semantic relevance mentioned above.

[0099] S430, identify the atomic objects of the same operation type and consecutive behavior in time in the session object as consecutive operation objects.

[0100] S440, determine the standard deviation of the time interval between multiple consecutive operation objects based on the timestamp of the behavior atomic object.

[0101] S450, operation quality is determined based on standard deviation, and operation quality is negatively correlated with standard deviation.

[0102] Operation quality = 1 / [1 + (operation rhythm)], where operation rhythm is the standard deviation of the time interval between all consecutive operations within a session.

[0103] S461, extract all manipulated water quality parameters from the session object to obtain a parameter set.

[0104] S462, the information entropy of the parameter set is determined as the parameter entropy.

[0105] S463, based on the session object, the sum of the number of operations of the top N parameters with the highest operation frequency in the parameter set is used as the ratio to the total number of operations of the parameters in the parameter set to obtain the focus rate.

[0106] S464, the hypothetical convergence is determined based on the parameter entropy and the focusing rate, wherein the hypothetical convergence is negatively correlated with the parameter entropy and positively correlated with the focusing rate.

[0107] S470. The analysis stagnation index is determined based on search entropy, operation quality, parameter correlation index and hypothesis convergence. The analysis stagnation index is positively correlated with search entropy, negatively correlated with operation quality, negatively correlated with parameter correlation index and negatively correlated with hypothesis convergence.

[0108] The thinking is divergent (high search entropy), the behavior is chaotic (low operation quality), and all attempts at correlation analysis have failed (multiple parameter correlation indicators are low). Therefore, we can set the analysis stagnation index as (search entropy) * (1 - operation quality) * (1 - MAX(parameter correlation index)).

[0109] Furthermore, to distinguish between purposeful multivariate hypothesis testing and random, unfocused activity by target users, hypothesis convergence is introduced. In other words, analytical stagnation refers to a state where, during the analysis process, target users exhibit high-frequency exploratory actions (such as searching, switching parameters / sites) but are accompanied by extremely inefficient verification actions (such as comparison, correlation analysis), and these exploratory actions fail to form stable parameter correlations, causing the analysis process to oscillate between multiple hypotheses and stagnate. The essence of analytical stagnation is an imbalance between exploration and verification, rather than simply divergent exploratory behavior.

[0110] Purposeful multivariate hypothesis testing is accompanied by testing behavior, while unordered blind spots lack effective testing mitigation. Therefore, a hypothesis convergence metric can be added to measure whether the target user's exploratory behavior converges to a few core hypotheses, thereby distinguishing between purposeful multivariate hypothesis testing and unordered blind spots.

[0111] Specifically, all manipulated water quality parameters in the session object are extracted to form a parameter set. The information entropy of this parameter set is calculated and denoted as parameter entropy (E_params). A high parameter entropy value indicates that the parameters that the user focuses on are very scattered. The sum of the number of operations on the top N parameters with the highest operation frequency in the session (such as the top 3) is counted as the ratio of the total number of operations on the parameters and denoted as focus rate (R_focus). It is assumed that the convergence degree (C_hypothesis) = focus rate (R_focus) / (1 + parameter entropy (E_params)).

[0112] A higher convergence rate indicates that while users may be searching broadly, their focus is clearly concentrated on a few core parameters, which aligns with the characteristics of purposeful multivariate hypothesis testing. Conversely, a low convergence rate suggests that users' attention is scattered, more akin to random, unfocused clicking.

[0113] At this point, the analysis stagnation index = (search entropy) * (1 - operation quality) * (1 - MAX(parameter correlation index)) * (1 - hypothesis convergence). In this formula, (1 - MAX(parameter correlation index)) and (1 - hypothesis convergence) are two key effectiveness decay factors. Even if the search entropy is high and the operation quality is slightly poor, as long as the user eventually finds a strong parameter correlation (high MAX(parameter correlation index) value) or focuses their behavior on a few hypotheses (high hypothesis convergence value), the overall analysis stagnation index value will be lowered, and the system will not easily determine stagnation.

[0114] The analysis of stagnation indicators will only increase significantly when the exploration behavior simultaneously satisfies divergence (high search entropy), low quality (low operation quality), ineffectiveness (low parameter correlation index), and lack of focus (low hypothesis convergence), so as to accurately capture the true stagnation state.

[0115] Preferably, the process of determining the convergence index includes steps S510 to S560: S510, Extract the operation type from the session object.

[0116] S520, Determine whether a preset end type exists in the operation type.

[0117] The default end type can be "Save", "Generate Report", or other similar operations.

[0118] S530, if there is a preset end type in the operation type, the session object is divided into an early session object and a late session object according to the timestamp.

[0119] S540, classify the operation objects according to the operation type.

[0120] S550, based on the classified operation objects, determine the early information entropy corresponding to each type of operation in the early session object and the late information entropy corresponding to the later session object.

[0121] S560 defines the ratio between the sum of multiple previous information entropies and the sum of multiple later information entropies as the convergence criterion.

[0122] The ratio between the sum of multiple earlier information entropies and the sum of multiple later information entropies; a high value (close to 1) indicates that the user is ending the analysis.

[0123] The analysis is truly complete only when both critical termination operations (the occurrence of a pre-defined termination type) and convergence of the focus (a reduction in information entropy) are satisfied.

[0124] The process of determining the abnormal cognitive stage of the target user and providing intelligent assistance to the target user through front-end window push based on the abnormal cognitive stage includes S610: S610, determine if the pollution source tracing index is greater than the preset pollution source tracing threshold; if the pollution source tracing index is greater than the preset pollution source tracing threshold, then execute S611~S612; if the pollution source tracing index is less than or equal to the preset pollution source tracing threshold, then execute S620.

[0125] S611 identifies the stations corresponding to the water quality anomaly warnings as abnormal stations.

[0126] S612 pushes a jump link in the front-end window that compares multiple parameters of multiple sites between the abnormal site and the previous preset sites.

[0127] The user is efficiently performing source tracing analysis and strongly suspects an upstream issue. The system should proactively push a multi-parameter comparison view of the upstream sites and prompt, "Source tracing analysis in progress. Do you want to calculate the theoretical arrival time of the contaminant plume at each site?"

[0128] S620, determine whether the parameter correlation index is greater than the preset parameter correlation threshold; if the parameter correlation index is greater than the preset parameter correlation threshold, then execute S621~S622; if the parameter correlation index is less than or equal to the preset parameter correlation threshold, then execute S630.

[0129] S621, Generate an overlay comparison chart of multiple related elements and determine the correlation coefficients among the multiple related elements.

[0130] S622 pushes a jump link with overlaid comparison charts and correlation coefficients in the front-end window.

[0131] The user is seriously verifying the causal relationship between two parameters. The system should automatically generate a superimposed curve and scatter plot of these two parameters, calculate their correlation coefficient, and directly push the results to the user.

[0132] S630, determine whether the analysis stagnation index is greater than the preset analysis stagnation threshold; if the analysis stagnation index is greater than the preset analysis stagnation threshold, then execute S631~S632; if the analysis stagnation index is less than or equal to the preset analysis stagnation threshold, then execute S640.

[0133] S631, acquire external events and abnormal parameters when a water quality anomaly warning occurs.

[0134] S632 pushes jump links for external events and exception parameters to the foreground window.

[0135] Users are struggling to find effective clues and are in a state of confusion. The system should activate the "thinking exploration mode" and proactively push information such as: external events, for example, "When the anomaly occurred, there were no work order records within 3 kilometers, but there was 2mm of rainfall."; unconventional parameters, for example, "We suggest you pay attention to the historical changes in oxidation-reduction potential (ORP)."; and historical cases, for example, "We found 3 historical cases with similar data patterns, and the ultimate cause was 'sediment disturbance' in all of them."

[0136] S640, determine whether the convergence index is greater than the preset convergence threshold; if the convergence index is greater than the preset convergence threshold, then execute S641~S642.

[0137] S641 generates a report draft based on multiple session objects.

[0138] S642 pushes a jump link to the report draft in the front-end window.

[0139] The draft report archives user actions for future reference.

[0140] If the convergence index is less than or equal to the preset convergence threshold, remain silent to avoid interference.

[0141] This application embodiment responds to water quality anomaly warnings by monitoring the target user's actions in the front-end window to obtain a session object. Based on the session object, it determines pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and judgment convergence indicators. These indicators are then compared sequentially with their corresponding thresholds. The indicator that first exceeds the threshold is identified as the target indicator. This follows the classic workflow of water experts, from anomaly discovery to pollution source tracing, correlation verification between pollution parameters, and attribution decision-making. By quantifying the target user's behavioral characteristics at each stage, the application accurately determines the target user's cognitive stage and determines the front-end window push accordingly, providing the target user with appropriate intelligent assistance and improving the efficiency of analyzing and handling water quality anomalies in water conservancy project management.

[0142] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0143] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for managing water conservancy projects, characterized in that, include: In response to water quality anomaly warnings, monitor the target user's actions in the front-end window and obtain the session object; Based on the session object, determine the pollution source tracing indicators, parameter correlation indicators, analysis stagnation indicators, and analysis convergence indicators; The pollution source tracing index, the parameter correlation index, the analysis stagnation index, and the judgment convergence index are compared with the threshold corresponding to the index in turn; The indicator that first exceeds the threshold of the indicator is identified as the target indicator; The push of the front-end window is determined based on the target indicators; The step of responding to a water quality anomaly warning by monitoring the target user's actions in the front-end window and obtaining a session object includes: In response to a water quality anomaly warning, listen for front-end events and obtain the original event object; The operation object and operation type are parsed from the original event object to obtain multiple behavior atomic objects; Determine the timestamp of each behavior atomic object; Adjacent behavior atomic objects whose timestamp difference is greater than a preset threshold are identified as divisible behavior atomic objects. The divisible behavior atomic object is segmented to obtain a session object; in, The process of determining the pollution source tracing indicators includes: The exploration parameters are determined based on the session object. These exploration parameters are used to measure the consistency between the target user's exploration direction in the spatial path and the expected direction of the upstream flow. The exploration length is determined based on the session object, and the exploration length is used to characterize the span of the exploration path of the target user; The exploration density is determined based on the session object, and the exploration density is used to measure the orderliness of the target user's exploration in the spatial path; The pollution source tracing index is determined based on the exploration parameters, the exploration length, and the exploration density, wherein the pollution source tracing index is positively correlated with the exploration parameters, negatively correlated with the exploration length, and positively correlated with the exploration density. The step of determining the exploration parameters based on the session object includes: Extract multiple sites from the session object to obtain a site sequence formed in chronological order; A movement vector is formed based on the movement directions corresponding to two adjacent stations in the station sequence; Determine the average vector among multiple movement vectors; The starting and ending points of the water flow are determined from the station sequence based on the actual water flow direction. The vector formed by pointing from the water flow termination point to the water flow starting point is determined as the reverse water flow vector; The cosine similarity between the average vector and the countercurrent vector is determined as the search parameter; The step of determining the probe density based on the session object includes: The multiple actual stations through which the water flows from the water flow termination point to the water flow start point are determined as a reverse water flow sequence, in the opposite direction of the actual water flow. Extract multiple character-type subsequences between the upstream flow sequence and the station sequence, wherein the character-type subsequence is a sequence formed by multiple elements that exist simultaneously in the upstream flow sequence and the station sequence and have the same sequential order; Mutually exclusive character subsequences among multiple character subsequences are defined as subsequence pairs, wherein the number of elements in each of the two character subsequences of the subsequence pair is greater than 1; The subsequence pair with the largest total number of elements among multiple subsequence pairs is identified as the target subsequence pair; The search density is determined by the ratio between the sum of the number of elements in the target subsequence pairs and the number of elements in the station sequence. The step of determining the probe length based on the session object includes: The search length is determined by the sum of the magnitudes of the multiple movement vectors of the station sequence. in, The process of determining the parameter correlation index includes: Extract the operation objects corresponding to the water quality influencing factors of the session object to obtain the sequence of influencing objects; Determine whether there are multiple elements alternating in the sequence of affected objects; If multiple elements appear alternately, then the multiple alternating elements are identified as related elements; Determine the first occurrence number of each relevant element in the sequence of affected objects; Determine the second number of times that multiple related elements alternate in the sequence of affected objects; The sum of multiple first numbers is obtained, and the ratio between the second number and the sum of multiple first numbers is determined as the parameter correlation index. in, The process of determining the analytical stagnation index includes: Extract the operation objects of type search from the session object to obtain multiple search objects; The search entropy of the session object is determined based on the information entropy of the multiple search objects; The atomic objects of the same operation type and consecutive in time in the session object are identified as consecutive operation objects; Determine the standard deviation of the time interval between multiple consecutive operation objects based on the timestamp of the behavior atomic object; Operational quality is determined based on the standard deviation, and the operational quality is negatively correlated with the standard deviation; Extract all manipulated water quality parameters from the session object to obtain a parameter set; The information entropy of the parameter set is determined as the parameter entropy; The focus rate is obtained by taking the sum of the number of operations of the top N most frequently operated parameters in the parameter set based on the session object, and the ratio of this sum to the total number of operations of all parameters in the parameter set. The hypothetical convergence is determined based on the parameter entropy and the focusing rate, wherein the hypothetical convergence is negatively correlated with the parameter entropy and positively correlated with the focusing rate; The analysis stagnation index is determined based on the search entropy, the operation quality, the parameter correlation index, and the hypothesis convergence. The analysis stagnation index is positively correlated with the search entropy, negatively correlated with the operation quality, negatively correlated with the parameter correlation index, and negatively correlated with the hypothesis convergence. in, The process of determining the convergence index includes: Extract the operation type from the session object; Determine whether a preset end type exists among the operation types; If a preset end type exists among the operation types, the session object is divided into an early session object and a late session object according to the timestamp; The operation objects are classified according to the operation type; Based on the classified operation objects, determine the corresponding early information entropy in the early session object and the corresponding late information entropy in the late session object for each type of operation. The ratio between the sum of multiple previous information entropies and the sum of multiple later information entropies is determined as the convergence criterion.

2. The method as described in claim 1, characterized in that, The step of determining the push of the front-end window based on the target metric includes: If the pollution source tracing index is greater than the preset pollution source tracing threshold, then the station corresponding to the water quality anomaly warning will be identified as an abnormal station. The front-end window pushes a jump link for comparing multiple parameters of the abnormal site and a preset number of sites before the abnormal site.

3. The method as described in claim 2, characterized in that, Before determining the subsequence pair with the largest total number of elements among multiple subsequence pairs as the target subsequence pair, the method further includes: If the character subsequence does not form a subsequence pair, then an empty sequence is configured to form a subsequence pair with the character subsequence.

4. The method as described in claim 3, characterized in that, The step of determining the push of the front-end window based on the target metric further includes: If the pollution source tracing index is less than or equal to the preset pollution source tracing threshold, then determine whether the parameter association index is greater than the preset parameter association threshold. If the parameter correlation index is greater than the preset parameter correlation threshold, then an overlay comparison chart between the multiple related elements is generated, and the correlation coefficient between the multiple related elements is determined. The overlay comparison chart and the correlation coefficient are pushed to the front-end window via a jump link.

5. The method as described in claim 4, characterized in that, The step of determining the push of the front-end window based on the target metric further includes: If the parameter correlation index is less than or equal to the preset parameter correlation threshold, then determine whether the analysis stagnation index is greater than the preset analysis stagnation threshold. If the analysis stagnation index is greater than the preset analysis stagnation threshold, then the external events and abnormal parameters at the time the water quality anomaly warning occurs are obtained; The external event and exception parameters are pushed to the front-end window as jump links.

6. The method as described in claim 5, characterized in that, The step of determining the push of the front-end window based on the target metric further includes: If the analysis stagnation index is less than or equal to the preset analysis stagnation threshold, then determine whether the analysis convergence index is greater than the preset analysis convergence threshold. If the assessment convergence index is greater than the preset assessment convergence threshold, a report draft is generated based on multiple session objects. A jump link to the report draft is pushed in the front-end window.

7. A water resources and water conservancy project management system, characterized in that, The system is used to perform the method described in any one of claims 1 to 6.

Images