An ai-based industrial intelligent diagnosis management and control method and platform

Abstract

This invention relates to the field of industrial control technology and provides an AI-based industrial intelligent diagnostic and control method and platform, comprising the following steps: collecting physical sensor data and business event data to determine the life elements in four dimensions: load stress, maintenance and repair, repair and remodeling, and environmental erosion, and constructing a cross-dimensional causal life narrative chain; based on the life narrative chain, using a preset state card library and a multi-dimensional erosion rate formula, calculating the lifespan consumption rate, and combining counterfactual reasoning and lifespan case anchor point calibration to generate a lifespan image containing the remaining lifespan range; calculating the remaining lifespan credit limit based on the lifespan image, and classifying the equipment into a safe autonomous zone, a concern negotiation zone, or a mandatory protection request zone according to the lifespan consumption rate and the degree of interaction deterioration; matching the task damage vector with the equipment's endurance envelope, and remodeling the task parameters and conducting multi-machine lifespan mutual assistance negotiation. This invention achieves autonomous closed-loop control oriented towards the balancing of fleet lifespan consumption.

Description

Technical Field

[0001] This invention relates to the field of industrial control technology, specifically to an AI-based industrial intelligent diagnostic control method and platform. Background Technology

[0002] In the long-term operation of industrial equipment, its actual service life is determined not only by the original design and manufacturing quality, but also, to a greater extent, by acquired factors such as operating load, maintenance timeliness, repair quality, and environmental corrosion. Existing technologies mostly rely on trend predictions of single physical states such as vibration and temperature, rarely systematically integrating multi-source information such as maintenance records, repair interventions, and load history. At the management level, conventional practices simply involve setting alarm thresholds and shutting down the machine, lacking the ability to dynamically adjust task intensity based on actual service life credit and flexibly balance service life consumption among machine groups, leading to production conflicts such as premature repairs or operation with defects. Therefore, there is a need to provide an AI-based industrial intelligent diagnostic and management method and platform to address these issues. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of this invention is to provide an AI-based industrial intelligent diagnosis and control method and platform to solve the problems existing in the above-mentioned background technology.

[0004] This invention is implemented as follows: an AI-based industrial intelligent diagnostic and control method, the method comprising the following steps: Collect equipment physical sensor data and business event data to determine the four dimensions of life elements: load stress, maintenance and repair, repair and reconstruction, and environmental erosion, and construct a cross-dimensional causal life narrative chain. Based on the life narrative chain, the life consumption rate is calculated using a preset state card library and a multidimensional erosion rate formula. Combined with counterfactual reasoning and life case anchor point calibration, a life image containing the remaining life interval is generated. The remaining lifetime credit is calculated based on the lifetime image, and the equipment is classified into a safe autonomous zone, a concern negotiation zone, or a mandatory protection request zone according to the lifetime consumption rate and the degree of interaction deterioration. By matching the mission damage vector with the equipment endurance envelope, and through mission parameter reshaping and multi-aircraft lifespan mutual assistance negotiation, autonomous closed-loop management for balancing fleet lifespan consumption is achieved. The mission damage vector describes the combination of consumption intensity caused by the mission to various dimensions of the equipment, and the equipment endurance envelope is the upper limit of mission intensity determined by the lifespan credit limit and the decay mode in the lifespan image.

[0005] As a further aspect of this invention, the steps of identifying four dimensions of life elements—load stress, maintenance and repair, remodeling, and environmental erosion—and constructing a cross-dimensional causal life narrative chain specifically include: The load stress index L is determined based on the real-time current ratio, impact frequency, and duration. The maintenance perfection index M is determined based on the timeliness of maintenance, the completeness of the project, and the compliance of the process. The repair and remodeling quality index R is determined based on the repair type, component quality grade, and installation accuracy. The environmental erosion index E is determined based on the workshop temperature, humidity, dust, and concentration of corrosive gases. By taking each change of each basic life element as a node, and adding wear-accelerating edges, early-stage repair vulnerability edges, and erosion-lubrication failure edges, a life narrative chain with temporal and causal relationships is formed.

[0006] As a further aspect of the present invention, the step of calculating the lifespan consumption rate based on the aforementioned life narrative chain, using a preset state card library and a multidimensional erosion rate formula, specifically includes: Retrieve cards from the preset status card library that match the combination of L, M, R, and E, and read the baseline consumption constant ω0 recorded by the card. Substituting ω0, L, M, R, and E into the multidimensional erosion rate formula: , where Ψ represents the lifespan consumption rate, a, b and c are power exponents, ε is the amplification factor of environmental damage to the load, and κ is the multidimensional interaction deterioration factor. The lifespan consumption rate is integrated and accumulated over time to obtain the cumulative lifespan consumption. The actual remaining lifespan is obtained by subtracting the cumulative lifespan consumption from the preset initial total lifespan of the equipment.

[0007] As a further aspect of the present invention, the step of generating a lifetime image containing the remaining lifetime range by combining counterfactual reasoning and lifetime case anchor calibration specifically includes: Copy the current life narrative chain, delete the specified maintenance or repair event, generate an uninterrupted branch, deduce along the branch according to the degradation rules and compare its remaining lifespan difference with the real branch to obtain the intervention gain or scar label; Retrieve anchor point cases that have completed their entire lifecycle, use event sequence similarity matching to find the most similar anchor point, project the total lifecycle onto the current equipment according to the narrative stage ratio, and determine the remaining lifecycle range; The remaining lifetime range, gain label, and scratch label are integrated into the lifetime image data package.

[0008] As a further aspect of the present invention, the step of classifying equipment according to its lifespan consumption rate and interaction degradation degree specifically includes: Extracting the interaction deterioration term from the multidimensional erosion rate formula The value of is used as a measure of the degree of interaction degradation; Preset warning lines, critical thresholds, and first and second thresholds for interaction degradation items; When the remaining lifespan credit limit is higher than the warning line, the interaction deterioration item is less than the first threshold, and there are no damage tags, the equipment enters the safe autonomous zone and can directly undertake missions. When the remaining life credit is below the warning line, or the interaction deterioration item is between the first and second thresholds, or there is a scar tag, the equipment enters the attention negotiation zone; When the remaining lifetime credit limit is lower than the critical threshold, or the interaction deterioration item is greater than the second threshold, the equipment enters the mandatory protection request zone, locks the equipment, and refuses processing tasks.

[0009] As a further aspect of the present invention, the task damage vector is matched with the equipment withstand capability envelope, and the steps of task parameter reshaping and multi-aircraft lifespan mutual assistance negotiation are specifically included: Establish a task damage vector library consisting of load stress index, environmental erosion index, and task duration; Based on the equipment's current remaining lifespan credit, decay mode, and lifespan consumption rate, generate a tolerance envelope that limits the upper limit of each dimension of the acceptable damage vector. The task damage vector is matched with the withstand envelope of candidate equipment to screen compliant equipment. If the specified equipment envelope is insufficient, the process knowledge base is invoked to adjust the task parameters and generate the reshaped damage vector. If critical processes must be performed by equipment in the concern negotiation zone that may still be overdrawn after remodeling, the equipment agent broadcasts the process damage vector to be shared and the mutual assistance points bounty to the fleet to achieve a balance of fleet life consumption.

[0010] Another objective of this invention is to provide an AI-based industrial intelligent diagnostic and control platform, the platform comprising: The narrative chain construction module is used to collect equipment physical sensor data and business event data, determine the life elements in four dimensions: load stress, maintenance and repair, repair and reconstruction, and environmental erosion, and construct a life narrative chain with cross-dimensional causal relationships. The lifetime image module is used to calculate the lifetime consumption rate based on the life narrative chain, using a preset state card library and a multidimensional erosion rate formula, and generate a lifetime image containing the remaining lifetime range by combining counterfactual reasoning and lifetime case anchor point calibration. The equipment classification module is used to calculate the remaining lifetime credit limit based on the lifetime image, and classify the equipment into a safe autonomous zone, a concern negotiation zone, or a mandatory protection request zone according to the lifetime consumption rate and the degree of interaction deterioration. The task management module is used to match the task damage vector with the equipment endurance envelope. Through task parameter reshaping and multi-machine lifespan mutual assistance negotiation, it realizes autonomous closed-loop management for the balance of fleet lifespan consumption. The task damage vector describes the combination of consumption intensity caused by the task to various dimensions of the equipment. The equipment endurance envelope is the upper limit of task intensity determined by the lifespan credit limit and the decay mode in the lifespan image.

[0011] Compared with the prior art, the beneficial effects of the present invention are: By extracting the four dimensions of life elements—load stress, maintenance and repair, remodeling, and environmental erosion—and constructing a life narrative chain with causal relationships, the complete service history of equipment is transformed into a reasonable atlas structure. Utilizing a pre-set state card library and a multi-dimensional erosion rate formula, the independent influences and synergistic deterioration effects among load, maintenance, repair, and environment are uniformly quantified into lifespan consumption rates. Combined with counterfactual reasoning and lifespan anchor point calibration, a lifespan image containing uncertain intervals is generated, completely avoiding the unknown risks of black-box model predictions.

[0012] By quantifying remaining lifetime as a consumable lifetime credit limit and dynamically classifying equipment into three control zones—safe and autonomous, attentive negotiation, or mandatory protection—based on real-time consumption rates and the degree of interaction deterioration, the risk status becomes clear and actionable. Furthermore, by defining task damage vectors and equipment tolerance envelopes, intelligent task-equipment matching is achieved. The introduction of task parameter reshaping and distributed multi-machine lifetime mutual assistance negotiation enables autonomous closed-loop control that balances fleet lifetime consumption while ensuring production. Attached Figure Description

[0013] Figure 1 This is a flowchart of an AI-based industrial intelligent diagnostic and control method.

[0014] Figure 2 This is a flowchart for constructing a life narrative chain in an AI-based industrial intelligent diagnostic and control method.

[0015] Figure 3 This is a flowchart illustrating the calculation of lifespan consumption rate in an AI-based industrial intelligent diagnostic and control method.

[0016] Figure 4 This is a flowchart for determining a lifetime image in an AI-based industrial intelligent diagnostic and control method.

[0017] Figure 5 This is a flowchart illustrating the classification of equipment in an AI-based industrial intelligent diagnostic and control method.

[0018] Figure 6 This is a flowchart illustrating mutual consultation in an AI-based industrial intelligent diagnostic and control method.

[0019] Figure 7 This is a structural block diagram of an AI-based industrial intelligent diagnostic and control platform. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0021] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0022] like Figure 1 As shown in the figure, this embodiment of the invention provides an AI-based industrial intelligent diagnosis and control method, the method comprising the following steps: S100 collects physical sensor data and business event data from equipment to determine the four dimensions of life elements: load stress, maintenance and repair, repair and remodeling, and environmental erosion, and constructs a cross-dimensional causal life narrative chain.

[0023] In this embodiment of the invention, physical sensor data (load status, environmental information) and business event data (maintenance information, repair information) are first collected. Then, based on the collected data, four quantifiable indices, namely load stress, maintenance and repair, repair and remodeling, and environmental erosion, are defined as life elements. Each change of each index and its mutual influence relationship are woven into a life narrative chain with temporal and causal logic.

[0024] S200: Based on the life narrative chain, using a preset state card library and a multidimensional erosion rate formula, calculate the life consumption rate, and combine counterfactual reasoning and life case anchor point calibration to generate a life image containing the remaining life interval.

[0025] In this embodiment of the invention, based on the life narrative chain, a baseline consumption constant is obtained by querying a pre-established state card library, and then the current lifetime consumption rate is dynamically calculated by substituting it into the multidimensional erosion rate formula. Finally, a lifetime image containing the remaining lifetime range is obtained by using counterfactual deduction and anchor point calibration of historical cases.

[0026] S300 calculates the remaining lifetime credit limit based on the lifetime image, and classifies the equipment into a safe autonomous zone, a concern negotiation zone, or a mandatory protection request zone based on the lifetime consumption rate and the degree of interaction deterioration.

[0027] In this embodiment of the invention, the remaining lifespan in the lifespan image is converted into a plannable lifespan credit limit. This can be achieved by extracting the lower limit of the remaining lifespan range in the lifespan image and performing a conservative calculation accordingly. For example, 1 hour of remaining lifespan is converted into 100 credits. Based on the real-time consumption rate and the degree of interaction degradation in the formula, equipment is automatically classified into a safe autonomous zone, a concern negotiation zone, or a mandatory protection request zone, thereby forming a clear basis for hierarchical control.

[0028] The S400 matches the mission damage vector with the equipment's endurance envelope, and through mission parameter reshaping and multi-aircraft lifespan mutual assistance negotiation, achieves autonomous closed-loop management for balancing the lifespan consumption of the fleet.

[0029] In this embodiment of the invention, a damage vector of the proposed production task is uploaded. The task damage vector describes the combination of consumption intensity caused by the task to various dimensions of the equipment. An equipment withstand capability envelope is determined, which is the upper limit of task intensity jointly determined by the lifetime credit limit and the decay pattern in the lifetime mapping. The task damage vector is matched with the current equipment withstand capability envelope to screen feasible equipment. When the match is insufficient, process parameters can be actively adjusted to reshape the task damage. If necessary, high-damage processes can be shared within the fleet through negotiation, achieving a balance and autonomous closed loop in the overall lifetime consumption of the fleet.

[0030] like Figure 2 As shown, in a preferred embodiment of the present invention, the steps of identifying four dimensions of life elements—load stress, maintenance and repair, remodeling, and environmental erosion—and constructing a cross-dimensional causal life narrative chain specifically include: S101, determine the load stress index L based on the real-time current ratio, impact frequency and duration; S102, the maintenance perfection index M is determined based on the timeliness of maintenance, the completeness of the project, and the compliance of the process. S103, the repair and remodeling quality index R is determined based on the repair type, component quality grade and installation accuracy; S104, the environmental erosion index E is determined based on the workshop temperature, humidity, dust and corrosive gas concentrations; S105 takes each change of each life element as a node, adds wear acceleration edge, early maintenance vulnerability edge and erosion-lubrication failure edge, forming a life narrative chain with temporal and causal relationships.

[0031] In this embodiment of the invention, the load stress index L is obtained according to a preset rule based on the ratio of the actual current to the rated current of the equipment, the frequency of impacts during operation, and the duration of a single impact. A higher value indicates a heavier load and a more severe impact. The maintenance perfection index M is determined based on the timeliness of maintenance execution, the completeness of the prescribed maintenance items, and the compliance of the operating procedures. A higher M value corresponds to complete timeliness, no omissions, and excellent processes. The maintenance remodeling quality index R is comprehensively evaluated based on the type of maintenance event (replacement, repair, welding, etc.), the quality grade of the replaced parts (original, aftermarket, repaired), and the precision indicators during installation (such as alignment adjustment and dynamic balance level). A higher R value indicates better maintenance results. The environmental erosion index E is determined using data from temperature, humidity, dust concentration, and corrosive gas sensors deployed in the workshop. A lower E value indicates a better environment. Finally, each change in L, M, R, and E is treated as a node and arranged chronologically. Simultaneously, based on domain knowledge, three types of associated edges are automatically added between nodes: when a low M period is immediately followed by a high L event, an "accelerated wear edge" is added; when a high L or high E event occurs within a vulnerable window period after a low R maintenance, an "early maintenance vulnerability edge" is added; and when E is high and the specified oil has not been changed during maintenance, an "erosion-lubrication failure edge" is added. All these nodes and edge types together constitute a graph structure, namely the equipment's life narrative chain.

[0032] like Figure 3 As shown, in a preferred embodiment of the present invention, the step of calculating the lifespan consumption rate based on the life narrative chain, using a preset state card library and a multidimensional erosion rate formula, specifically includes: S201, retrieve a card from the preset status card library that matches the combination of L, M, R and E, and read the baseline consumption constant ω0 recorded by the card; S202, Substitute ω0, L, M, R, and E into the multidimensional erosion rate formula: , where Ψ represents the lifespan consumption rate, a, b and c are power exponents, ε is the amplification factor of environmental damage to the load, and κ is the multidimensional interaction deterioration factor. S203, the lifespan consumption rate is integrated and accumulated over time to obtain the cumulative lifespan consumption, and the actual remaining lifespan is obtained by subtracting the cumulative lifespan consumption from the preset initial total lifespan of the equipment.

[0033] In this embodiment of the invention, the preset state card library is a multidimensional lookup table. Each card corresponds to a set of discrete combinations of L, M, R, and E, and also corresponds to a baseline consumption constant and a dominant decay mode. Based on the actual values ​​of the four primitives, the most matching card is retrieved, and a pre-calibrated baseline consumption constant ω0 is directly read from the card. ω0 represents the baseline lifespan points consumed per hour by the equipment under standard operating conditions in this typical state. Then, the multidimensional erosion rate formula is substituted into the formula to calculate Ψ, where the power exponents a, b, and c respectively characterize the independent nonlinear enhancement effects of load, lack of maintenance (2-M), and incomplete repair (2-R). This exponential term is a multidimensional interactive deterioration term, specifically capturing the additional consumption multiple caused by the synergistic surge of several factors when maintenance is inadequate and repairs are incomplete, while simultaneously subjected to high loads (L) and high corrosion (E). Through this formula, the coupling effect between maintenance, repairs, load, and environment is calculated accurately in one go. Then, the above calculation is repeatedly performed according to a set time cycle to obtain Ψ within each cycle, and Ψ is multiplied by the time cycle length and accumulated to the cumulative lifespan consumption value. When the equipment is put into production, an initial total lifespan point is preset; subtracting the continuously accumulating lifespan consumption from this total point yields the actual remaining lifespan at the current moment.

[0034] like Figure 4 As shown, in a preferred embodiment of the present invention, the step of generating a lifetime image containing the remaining lifetime range by combining counterfactual reasoning and lifetime case anchor calibration specifically includes: S204, copy the current life narrative chain, delete the specified maintenance or repair event, generate an uninterrupted branch, deduce along the branch according to the degradation rule and compare its remaining lifespan difference with the real branch to obtain the intervention gain or scar label; S205: Retrieve anchor point cases that have completed their entire lifecycle, use event sequence similarity matching to find the most similar anchor point, project the total lifecycle onto the current equipment according to the narrative stage ratio, and determine the remaining lifecycle range. S206 integrates the remaining lifetime range, gain tag, and scratch tag into the lifetime image data package.

[0035] In this embodiment of the invention, the current real life narrative chain is replicated, and a specified maintenance or repair event is deleted from the replicated chain (i.e., it is assumed that the intervention never occurred), generating a non-intervention branch. Then, according to preset degradation rules, the changes in M ​​and R are deduced on this branch, and the remaining lifespan under the non-intervention scenario is simulated using the same multidimensional erosion rate formula. The real remaining lifespan is compared with the simulated non-intervention remaining lifespan: if the real lifespan is significantly longer, the intervention brought a positive gain; if the real lifespan does not extend or even shows abnormal decline, it indicates that the repair may have introduced hidden damage. The gain or damage is quantified as a specific label and attached to the lifespan image. The degradation rules are as follows: 1. Natural degradation deduction of M (maintenance completeness): In the real branch, each time maintenance is performed, the M value is re-evaluated to a higher value (e.g., 0.9) based on the quality of maintenance completion. In the non-intervention branch, because it is assumed that maintenance was not performed, M is not re-evaluated, but rather decays continuously over time from the previous value. For example, if maintenance is not performed for more than twice the scheduled maintenance period, the M value is reduced by 0.2 from the current value; if environmental corrosion (E) is greater than 0.7, the degradation rate doubles. Thus, the computer gradually reduces the M value in the uninterrupted branch according to the time cycle. 2. Deduction of missing and residual damage in R (Repair and Remodeling Quality): In the uninterrupted branch, assuming that maintenance was not performed, the R value will remain at the low value before maintenance, or continue to decrease along the preset curve due to the development of the fault. More importantly, the damage before maintenance (such as "early pitting of bearings") will continue to exist. The degradation rule will instruct the computer to automatically upgrade these damages, for example, changing "early pitting" to "expanding spalling," and accordingly lowering the R value or health status.

[0036] In this embodiment of the invention, an anchor point case library is established in advance, storing the complete life narrative chains and final total lifespan of several pieces of equipment of the same type that have been operated to the point of scrapping. Through similarity matching of event sequences (comparing the types, order, and proportions of events they experienced), several anchor points most similar to the current equipment's narrative are identified. Referring to the proportion of the anchor point's total lifespan in the similar narrative phase, the anchor point's total lifespan is projected onto the current equipment, resulting in multiple possible remaining lifespan prediction values. These values ​​constitute a range interval, thus yielding the remaining lifespan interval. For example, if the proportion of anchor point A's total lifespan in the similar narrative phase is 25%, and the equipment has accumulated 1500 hours of operation when the corresponding event occurs, then the equipment's estimated total lifespan = 1500 / 25% = 6000 hours. The remaining lifespan predicted by anchor point A is 6000 - 1500 = 4500 hours. Finally, the remaining lifespan interval, gain tags, and damage tags are packaged together to generate a structured lifespan mapping data package.

[0037] like Figure 5As shown, in a preferred embodiment of the present invention, the step of classifying equipment according to its lifespan consumption rate and interaction degradation degree specifically includes: S301, Extracting the interactive degradation term from the multidimensional erosion rate formula. The value of is used as a measure of the degree of interaction degradation; S302, preset warning line, critical threshold, and first and second thresholds for interaction deterioration items; S303: When the remaining lifespan credit limit is higher than the warning line, the interaction deterioration item is less than the first threshold, and there are no damage tags, the equipment enters the safe autonomous zone and can directly undertake missions. S304. When the remaining life credit is below the warning line, or the interaction deterioration item is between the first and second thresholds, or there is a scar tag, the equipment enters the attention negotiation zone. S305: When the remaining life credit limit is lower than the critical threshold, or the interaction deterioration item is greater than the second threshold, the equipment enters the mandatory protection request zone, locks the equipment, and refuses processing tasks.

[0038] In this embodiment of the invention, the multidimensional erosion rate formula is specifically used to... The value is extracted and used as a direct quantitative indicator of the current degree of interaction deterioration. A value of 1 indicates no interaction deterioration, while a value greater than 1 indicates a synergistic acceleration effect; the larger the value, the more severe the deterioration. Two lifetime credit thresholds (warning line and critical threshold) and two thresholds for interaction deterioration items (first threshold and second threshold) are pre-set. Then, the equipment can be classified. When the equipment is classified into the safe autonomous zone, it indicates that the equipment is in good health and can undertake various production tasks without restriction. When the equipment enters the mandatory protection request zone, the equipment's processing capacity is immediately locked, all production tasks are rejected, and an emergency maintenance request is sent to the maintenance system with detailed lifetime image information until manual intervention or repair is completed.

[0039] like Figure 6 As shown, in a preferred embodiment of the present invention, the task damage vector is matched with the equipment withstand capability envelope, and the steps of task parameter reshaping and multi-aircraft lifespan mutual assistance negotiation are specifically included: S401, Establish a task damage vector library consisting of load stress index, environmental erosion index and task duration; S402 generates a tolerance envelope that limits the upper limit of each dimension of the acceptable damage vector, based on the equipment's current remaining life credit, decay mode, and life consumption rate. S403 matches the mission damage vector with the withstand envelope of candidate equipment to screen compliant equipment. S404 If the specified equipment envelope is insufficient, the process knowledge base is called to adjust the task parameters and generate the reshaped damage vector. S405 If a critical process must be performed by equipment in the concern negotiation zone that may still be overdrawn after remodeling, the equipment agent broadcasts the process damage vector to be shared and the mutual assistance points bounty to the fleet to achieve a balance of fleet life consumption.

[0040] In this embodiment of the invention, a multidimensional damage vector is predefined for each type of production task. This vector consists of three components: the load stress index L (the average or peak load intensity of the task), the environmental erosion index E, and the task duration T. Each piece of equipment dynamically generates a tolerance envelope based on its current remaining lifetime credit, the dominant degradation mode indicated by the lifetime mapping (such as wear or fatigue), and the lifetime consumption rate. This envelope explicitly defines the upper limits of the L, E, and T components in the acceptable task damage vector. A rule mapping table can be pre-defined, containing the upper limits of L, E, and T corresponding to each combination of credit range, degradation mode, and consumption rate range. During production scheduling, the damage vector of the task to be assigned is compared item by item with the tolerance envelope of the candidate equipment. Equipment whose components do not exceed the upper limit of the envelope is considered compliant and can be directly assigned tasks. If a critical process must be performed by a specific piece of equipment, but the damage vector of the original task exceeds the tolerance envelope of that equipment, a task parameter reshaping process is triggered. The process knowledge base is invoked to attempt to adjust process parameters (e.g., changing one large depth of cut to two small depths of cut, or reducing the feed rate and extending the processing time) without altering the final machining quality, generating a new damage vector. When a piece of equipment is in the concern negotiation zone, and after task refactoring assessment, executing the entire critical process would still result in credit overdraft, mutual assistance negotiation is initiated. The equipment's agent broadcasts a sharing request to the cluster, clearly listing the damage vector of the process to be shared and the mutual assistance points reward offered for this sharing. Upon receiving the request, agents of other equipment in the cluster, based on their remaining credit and future scheduled tasks, autonomously decide whether to respond and how many points to request in return. This mechanism allows for cross-equipment balancing of the overall lifespan consumption of the cluster.

[0041] like Figure 7 As shown in the figure, this embodiment of the invention also provides an AI-based industrial intelligent diagnostic and control platform, the platform comprising: The narrative chain construction module 100 is used to collect equipment physical sensor data and business event data, determine the life elements in four dimensions: load stress, maintenance and repair, repair and remodeling, and environmental erosion, and construct a life narrative chain with cross-dimensional causal relationships. The lifetime image module 200 is used to calculate the lifetime consumption rate based on the life narrative chain, using a preset state card library and a multidimensional erosion rate formula, and generate a lifetime image containing the remaining lifetime range by combining counterfactual reasoning and lifetime case anchor point calibration. The equipment classification module 300 is used to calculate the remaining lifetime credit limit based on the lifetime image, and classify the equipment into a safe autonomous zone, a concern negotiation zone, or a mandatory protection request zone according to the lifetime consumption rate and the degree of interaction deterioration. The task management module 400 is used to match the task damage vector with the equipment endurance envelope. Through task parameter reshaping and multi-machine lifespan mutual assistance negotiation, it realizes autonomous closed-loop management for the balance of fleet lifespan consumption. The task damage vector describes the combination of consumption intensity caused by the task to various dimensions of the equipment. The equipment endurance envelope is the upper limit of task intensity determined by the lifespan credit limit and the decay mode in the lifespan image.

[0042] In a preferred embodiment of the present invention, the narrative chain construction module 100 includes: The load stress element is used to determine the load stress index L based on the real-time current ratio, impact frequency, and duration. The maintenance completeness unit is used to determine the maintenance completeness index M based on maintenance timeliness, project completeness, and process compliance. The repair and remodeling unit is used to determine the repair and remodeling quality index R based on the repair type, component quality grade, and installation accuracy. The environmental erosion unit is used to determine the environmental erosion index E based on workshop temperature, humidity, dust and corrosive gas concentrations; The life narrative chain unit is used to take each change of each life element as a node, add wear-accelerating edge, early maintenance vulnerability edge, and erosion-lubrication failure edge to form a life narrative chain with temporal and causal relationships.

[0043] In a preferred embodiment of the present invention, the lifetime imaging module 200 includes: The card matching unit is used to retrieve cards that match the combination of L, M, R and E from a preset state card library, and read the baseline consumption constant ω0 recorded for that card. The consumption rate unit is used to substitute ω0, L, M, R, and E into the multidimensional erosion rate formula: , where Ψ represents the lifespan consumption rate, a, b and c are power exponents, ε is the amplification factor of environmental damage to the load, and κ is the multidimensional interaction deterioration factor. The remaining lifespan unit is used to integrate and accumulate the lifespan consumption rate over time to obtain the cumulative lifespan consumption. The actual remaining lifespan is obtained by subtracting the cumulative lifespan consumption from the preset initial total lifespan of the equipment.

[0044] In a preferred embodiment of the present invention, the lifetime imaging module 200 further includes: The branch deduction unit is used to replicate the current life narrative chain, delete specified maintenance or repair events, generate uninterrupted branches, and deduce along the branches according to the degradation rules and compare their remaining lifespan differences with the real branches to obtain intervention gains or scar labels. The lifespan interval unit is used to retrieve anchor point cases that have completed their entire lifespan. It uses event sequence similarity matching to find the most similar anchor point, projects the total lifespan onto the current equipment according to the narrative stage ratio, and determines the remaining lifespan interval. The lifetime mapping unit is used to integrate the remaining lifetime range, gain label, and scratch label into the lifetime mapping data packet.

[0045] The above description only details the preferred embodiments of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0046] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0047] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the disclosure in the specification and embodiments. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

Claims

1. An AI-based industrial intelligent diagnostic and control method, characterized in that, The method includes the following steps: Collect equipment physical sensor data and business event data to determine the life elements in four dimensions: load stress, maintenance and repair, repair and reconstruction, and environmental erosion, and construct a life narrative chain with cross-dimensional causal relationships. Based on the life narrative chain, the life consumption rate is calculated using a preset state card library and a multidimensional erosion rate formula. Combined with counterfactual reasoning and life case anchor point calibration, a life image containing the remaining life interval is generated. The remaining lifetime credit is calculated based on the lifetime image, and the equipment is classified into a safe autonomous zone, a concern negotiation zone, or a mandatory protection request zone according to the lifetime consumption rate and the degree of interaction deterioration. By matching the mission damage vector with the equipment endurance envelope, and through mission parameter reshaping and multi-machine lifetime mutual assistance negotiation, the mission damage vector describes the combination of consumption intensity caused by the mission to various dimensions of the equipment. The equipment endurance envelope is the upper limit of mission intensity determined by the lifetime credit limit and the decay mode in the lifetime image.

2. The AI-based industrial intelligent diagnosis and control method according to claim 1, characterized in that, The steps involved in identifying the four dimensions of life elements—load stress, maintenance and repair, remodeling, and environmental erosion—and constructing a cross-dimensional causal life narrative chain include: The load stress index L is determined based on the real-time current ratio, impact frequency, and duration. The maintenance perfection index M is determined based on the timeliness of maintenance, the completeness of the project, and the compliance of the process. The repair and remodeling quality index R is determined based on the repair type, component quality grade, and installation accuracy. The environmental erosion index E is determined based on the workshop temperature, humidity, dust, and concentration of corrosive gases. By taking each change of each basic life element as a node, and adding wear-accelerating edges, early-stage repair vulnerability edges, and erosion-lubrication failure edges, a life narrative chain with temporal and causal relationships is formed.

3. The AI-based industrial intelligent diagnosis and control method according to claim 2, characterized in that, Based on the aforementioned life narrative chain, the steps for calculating the lifespan consumption rate using a preset state card library and a multidimensional erosion rate formula specifically include: Retrieve cards from the preset state card library that match the combination of L, M, R, and E, and read the baseline consumption constant ω0 recorded by the card. Substituting ω0, L, M, R, and E into the multidimensional erosion rate formula: , where Ψ represents the lifespan consumption rate, a, b and c are power exponents, ε is the amplification factor of environmental damage to the load, and κ is the multidimensional interaction deterioration factor. The lifespan consumption rate is integrated and accumulated over time to obtain the cumulative lifespan consumption. The actual remaining lifespan is obtained by subtracting the cumulative lifespan consumption from the preset initial total lifespan of the equipment.

4. The AI-based industrial intelligent diagnosis and control method according to claim 3, characterized in that, The steps for generating a lifetime map containing the remaining lifetime range, combining counterfactual reasoning and lifetime case anchor calibration, specifically include: Copy the current life narrative chain, delete the specified maintenance or repair event, generate an uninterrupted branch, deduce along the branch according to the degradation rules and compare its remaining lifespan difference with the real branch to obtain the intervention gain or scar label; Retrieve anchor point cases that have completed their entire lifecycle, use event sequence similarity matching to find the most similar anchor point, project the total lifecycle onto the current equipment according to the narrative stage ratio, and determine the remaining lifecycle range; The remaining lifetime range, gain label, and scratch label are integrated into the lifetime image data package.

5. The AI-based industrial intelligent diagnosis and control method according to claim 4, characterized in that, The steps for classifying equipment based on its lifespan consumption rate and the degree of interaction degradation specifically include: Extracting the interaction deterioration term from the multidimensional erosion rate formula The value of is used as a measure of the degree of interaction degradation; Preset warning lines, critical thresholds, and first and second thresholds for interaction degradation items; When the remaining lifespan credit limit is higher than the warning line, the interaction deterioration item is less than the first threshold, and there are no damage tags, the equipment enters the safe autonomous zone and can directly undertake missions. When the remaining life credit is below the warning line, or the interaction deterioration item is between the first and second thresholds, or there is a scar tag, the equipment enters the attention negotiation zone; When the remaining lifetime credit limit is lower than the critical threshold, or the interaction deterioration item is greater than the second threshold, the equipment enters the mandatory protection request zone, locks the equipment, and refuses processing tasks.

6. The AI-based industrial intelligent diagnosis and control method according to claim 2, characterized in that, Matching the mission damage vector with the equipment's endurance envelope, through steps such as mission parameter reshaping and multi-aircraft lifespan cooperative negotiation, specifically includes: Establish a task damage vector library consisting of load stress index, environmental erosion index, and task duration; Based on the equipment's current remaining lifespan credit, decay mode, and lifespan consumption rate, generate a tolerance envelope that limits the upper limit of each dimension of the acceptable damage vector. The task damage vector is matched with the withstand envelope of candidate equipment to screen compliant equipment. If the specified equipment envelope is insufficient, the process knowledge base is invoked to adjust the task parameters and generate the reshaped damage vector. If critical processes must be performed by equipment in the concern negotiation zone that may still be overdrawn after remodeling, the equipment agent broadcasts the process damage vector to be shared and the mutual assistance points bounty to the fleet to achieve a balance of fleet life consumption.

7. An AI-based industrial intelligent diagnostic and control platform, characterized in that, The platform includes: The narrative chain construction module is used to collect equipment physical sensor data and business event data, determine the life elements in four dimensions: load stress, maintenance and repair, repair and reconstruction, and environmental erosion, and construct a life narrative chain with cross-dimensional causal relationships. The lifetime image module is used to calculate the lifetime consumption rate based on the life narrative chain, using a preset state card library and a multidimensional erosion rate formula, and generate a lifetime image containing the remaining lifetime range by combining counterfactual reasoning and lifetime case anchor point calibration. The equipment classification module is used to calculate the remaining lifetime credit limit based on the lifetime image, and classify the equipment into a safe autonomous zone, a concern negotiation zone, or a mandatory protection request zone according to the lifetime consumption rate and the degree of interaction deterioration. The task management module is used to match the task damage vector with the equipment endurance envelope. Through task parameter reshaping and multi-machine lifespan mutual assistance negotiation, the task damage vector describes the combination of consumption intensity caused by the task to various dimensions of the equipment. The equipment endurance envelope is the upper limit of task intensity determined by the lifespan credit limit and the decay mode in the lifespan image.

8. The AI-based industrial intelligent diagnostic and control platform according to claim 7, characterized in that, The narrative chain construction module includes: The load stress element is used to determine the load stress index L based on the real-time current ratio, impact frequency, and duration. The maintenance completeness unit is used to determine the maintenance completeness index M based on maintenance timeliness, project completeness, and process compliance. The repair and remodeling unit is used to determine the repair and remodeling quality index R based on the repair type, component quality grade, and installation accuracy. The environmental erosion unit is used to determine the environmental erosion index E based on workshop temperature, humidity, dust and corrosive gas concentrations; The life narrative chain unit is used to take each change of each life element as a node, add wear-accelerating edge, early maintenance vulnerability edge, and erosion-lubrication failure edge to form a life narrative chain with temporal and causal relationships.

9. The AI-based industrial intelligent diagnostic and control platform according to claim 8, characterized in that, The lifetime imaging module includes: The card matching unit is used to retrieve cards that match the combination of L, M, R and E from a preset state card library, and read the baseline consumption constant ω0 recorded for that card. The consumption rate unit is used to substitute ω0, L, M, R, and E into the multidimensional erosion rate formula: , where Ψ represents the lifespan consumption rate, a, b and c are power exponents, ε is the amplification factor of environmental damage to the load, and κ is the multidimensional interaction deterioration factor. The remaining lifespan unit is used to integrate and accumulate the lifespan consumption rate over time to obtain the cumulative lifespan consumption. The actual remaining lifespan is obtained by subtracting the cumulative lifespan consumption from the preset initial total lifespan of the equipment.

10. The AI-based industrial intelligent diagnostic and control platform according to claim 9, characterized in that, The lifetime mapping module also includes: The branch deduction unit is used to replicate the current life narrative chain, delete specified maintenance or repair events, generate uninterrupted branches, and deduce along the branches according to the degradation rules and compare their remaining lifespan differences with the real branches to obtain intervention gains or scar labels. The lifespan interval unit is used to retrieve anchor point cases that have completed their entire lifespan. It uses event sequence similarity matching to find the most similar anchor point, projects the total lifespan onto the current equipment according to the narrative stage ratio, and determines the remaining lifespan interval. The lifetime mapping unit is used to integrate the remaining lifetime range, gain label, and scratch label into the lifetime mapping data packet.

Images