CN-121808708-B - Industrial equipment energy efficiency root cause analysis and optimization method and system based on knowledge graph and stream computing

CN121808708BCN 121808708 BCN121808708 BCN 121808708BCN-121808708-B

Abstract

The invention provides an industrial equipment energy efficiency root cause analysis and optimization method and system based on knowledge graph and stream calculation, wherein the method comprises the steps of extracting nodes with structure, operation or mechanism relation from the knowledge graph, and calculating energy efficiency causal constraint weight; the method comprises the steps of obtaining running stream data, calculating trigger judgment quantity by combining weight and jitter suppression items, generating an energy efficiency alignable event, calculating equivalent overlapping length through a translation time interval within an allowable delay range, obtaining event alignment strength, obtaining stable alignment strength through exponential sliding average smoothing, generating candidate energy efficiency root cause alignment results, calculating relative optimization priority based on the stable alignment strength, and matching and issuing an optimization action instruction. The invention utilizes the knowledge graph to define the causal boundary, solves the problems of inaccurate positioning and repeated fluctuation of conclusion of energy efficiency root cause in industrial scene through anti-jitter triggering, uncertain delay alignment and stability smoothing mechanism in stream calculation, and realizes closed-loop control of energy efficiency optimization.

Inventors

Cai Hujia
PANG HUAXIN
ZHENG TIANFENG
HUANG ZHIWEN

Assignees

广州云天数据技术有限公司

Dates

Publication Date: 20260508
Application Date: 20260309

Claims (10)

1. The industrial equipment energy efficiency root cause analysis and optimization method based on knowledge graph and stream calculation is characterized by comprising the following steps: extracting candidate equipment nodes and energy efficiency related nodes from an industrial equipment knowledge graph, wherein the candidate equipment nodes and target equipment nodes have structural connection, operation dependence or mechanism action relation; counting the number of the effective relations and carrying out normalization calculation to obtain the energy efficiency causal constraint weight; acquiring operation flow data of industrial equipment, and calculating operation variation amplitude and persistence factors in a sliding time window; calculating original event triggering intensity, wherein the original event triggering intensity is the product of the energy efficiency causal constraint weight, the operation change amplitude and the persistence factor; Calculating a jitter suppression term, wherein the jitter suppression term is the product of the absolute value of the triggering intensity difference value of the original event at adjacent time and a preset jitter suppression coefficient; Generating a trigger judgment amount which is the difference value of the original event trigger intensity minus the jitter suppression term; If the triggering judgment amount reaches a triggering threshold value, generating an industrial equipment energy efficiency alignable event carrying a time interval; calculating the equivalent overlapping interval length through a translation time interval within an allowable delay range, wherein the equivalent overlapping interval length is the maximum overlapping length of an operation change event and an energy efficiency performance change event; generating event alignment intensity, wherein the event alignment intensity is the product of a first ratio and the energy efficiency causal constraint weight, and the first ratio is the ratio of the equivalent overlapping interval length to the union interval length; Performing accumulated smoothing treatment on the event alignment intensity by using an exponential sliding average recursion formula to obtain stable alignment intensity; If the stable alignment strength reaches a candidate judgment threshold, generating a candidate energy efficiency root cause alignment result; Calculating normalized relative optimization priorities based on the stable alignment intensities; And matching a target optimization action from an optimization action library according to the relative optimization priority, and issuing an execution instruction.
2. The industrial equipment energy efficiency root cause analysis and optimization method based on knowledge graph and stream computation according to claim 1, wherein the normalized computation obtains an energy efficiency causal constraint weight, comprising: taking the number of effective relations between the candidate equipment nodes and the energy efficiency related nodes as molecules; taking the sum of the number of the effective relations of the candidate equipment nodes pointing to all energy efficiency related nodes as a denominator; and calculating the quotient of the numerator and the denominator to obtain dimensionless relative constraint weight as the energy efficiency causal constraint weight.
3. The industrial equipment energy efficiency root cause analysis and optimization method based on knowledge graph and stream calculation according to claim 1, wherein the calculation of the operation variation amplitude comprises: If the point location data type is continuous numerical value type, calculating a difference value of adjacent sampling values, and carrying out normalization processing on the difference value by utilizing a preset allowable fluctuation range or historical stable segment statistic to obtain an operation variation amplitude; if the point data type is discrete state type, counting the state switching times in the sliding time window, and normalizing the state switching times by utilizing a preset allowable switching frequency upper limit to obtain the operation change amplitude.
4. The industrial equipment energy efficiency root cause analysis and optimization method based on knowledge graph and stream computation according to claim 1, wherein the computing the equivalent overlap interval length comprises: judging whether a direct overlapping part exists between a time interval of the operation change event and a time interval of the energy efficiency expression change event; if a direct overlapping part exists, determining the length of the direct overlapping part as the equivalent overlapping interval length; if the direct overlapping part does not exist but the interval falls into the allowable delay range, shifting one time interval in the allowable delay range until the maximum overlapping part is formed, and determining the length of the maximum overlapping part as the equivalent overlapping interval length.
5. The method for analyzing and optimizing root cause of industrial equipment energy efficiency based on knowledge graph and flow calculation according to claim 1, wherein the obtaining stable alignment strength comprises: Acquiring a preset smoothing coefficient; calculating a first product, wherein the first product is the product of the preset smoothing coefficient and the stable alignment strength at the last moment; calculating a second numerical value, wherein the second numerical value is a difference value between 1 and the preset smoothing coefficient; calculating a second product, wherein the second product is the product of the second numerical value and the event alignment intensity at the current moment; And taking the sum of the first product and the second product as the stable alignment strength at the current moment.
6. The method for analyzing and optimizing root cause of industrial equipment energy efficiency based on knowledge-graph and flow calculation according to claim 1, wherein the calculating relative optimization priority comprises: selecting stable alignment strength of a specific candidate root cause relationship as a molecule; calculating the sum of stable alignment intensities of all candidate root cause relations under the same energy efficiency node as a denominator; dividing the numerator by the denominator to obtain a value in the form of a duty cycle as a relative optimization priority.
7. The industrial equipment energy efficiency root cause analysis and optimization method based on knowledge graph and stream computation according to claim 1, wherein the optimization action library comprises a parameter adjustment action, an operation mode switching action, a load distribution adjustment action and an operation and maintenance treatment action; The parameter adjustment action corresponds to performing a write point operation to a control system; The operation mode switching action corresponds to issuing a mode switching instruction to the control system; the load distribution adjustment action corresponds to issuing a load distribution instruction to the scheduling system; the operation and maintenance treatment action corresponds to generating an operation and maintenance work order or pushing alarm information.
8. The knowledge-graph and flow-computing-based industrial equipment energy efficiency root cause analysis and optimization method according to claim 1, wherein the method further comprises: And generating an operation adjustment result record, wherein the operation adjustment result record at least comprises a device node identifier, an associated energy efficiency node identifier, an action type identifier, an action issuing time stamp and an action parameter identifier which correspond to the executed action.
9. The method for analyzing and optimizing the root cause of energy efficiency of industrial equipment based on knowledge graph and stream calculation according to claim 1, wherein the length of the union interval is the length of the union of the time interval of running change event and the time interval of energy efficiency expression change event.
10. Industrial equipment energy efficiency root cause analysis and optimization system based on knowledge graph and stream calculation, which is characterized by comprising: The constraint information generation module is configured to extract candidate equipment nodes and energy efficiency related nodes from the industrial equipment knowledge graph, wherein the candidate equipment nodes and the target equipment nodes have structural connection, operation dependence or mechanism action relation; the system comprises an event generation module, an event suppression item, a trigger judgment amount and an industrial equipment energy efficiency alignable event, wherein the event generation module is configured to acquire industrial equipment operation flow data, calculate operation change amplitude and persistence factors in a sliding time window, calculate original event trigger intensity, calculate jitter suppression items, generate trigger judgment amounts, and generate industrial equipment energy efficiency alignable events in a carrying time interval if the trigger judgment amounts reach a trigger threshold, wherein the original event trigger intensity is the product of the energy efficiency causality constraint weight, the operation change amplitude and the persistence factors; The root cause alignment analysis module is configured to calculate an equivalent overlapping interval length through a translation time interval within an allowable delay range, wherein the equivalent overlapping interval length is the maximum overlapping length of an operation change event and an energy efficiency performance change event, generate event alignment strength, wherein the event alignment strength is the product of a first ratio and the energy efficiency causal constraint weight, the first ratio is the ratio of the equivalent overlapping interval length to a union interval length, and perform accumulated smoothing on the event alignment strength by utilizing an exponential sliding average recursion formula to obtain stable alignment strength; and the optimization execution module is configured to calculate normalized relative optimization priority based on the stable alignment intensity, match target optimization actions from an optimization action library according to the relative optimization priority and issue execution instructions.

Description

Industrial equipment energy efficiency root cause analysis and optimization method and system based on knowledge graph and stream computing Technical Field The invention belongs to the field of industrial Internet of things, and particularly relates to an industrial equipment energy efficiency root cause analysis and optimization method and system based on knowledge graph and stream computing. Background In the industrial field, the energy efficiency level of equipment directly relates to energy utilization efficiency, running cost and system safety, and especially in the scenes of electric power, energy, rail transit, large industrial parks and the like, the equipment quantity is large, the system coupling is strong, the running working condition continuously changes, so that the energy efficiency abnormality often has the characteristics of cross-equipment propagation and cross-time appearance. The existing energy efficiency analysis technology mainly relies on an energy consumption monitoring system, historical data statistical analysis or rule threshold judgment to identify and evaluate the abnormality of a single device or a local system, and partial systems are introduced into flow calculation to realize online acquisition and real-time analysis, and meanwhile attempt to describe the structural relationship, operation dependence and process logic of the device by utilizing a knowledge graph for auxiliary positioning. However, a common bottleneck still exists in engineering application, namely on one hand, on the one hand, the streaming analysis is generally based on the association of a fixed time window or a neighboring moment, the time approximate synchronization or delay of the running change of default equipment and the energy efficiency change can be covered by the fixed window, but the influence of actions such as starting and stopping, adjustment, working condition switching, load distribution and the like on the energy efficiency in actual running often presents uncertain delay and gradual manifestation, the delay amplitude changes along with the changes of working conditions and equipment states, the association under the fixed window is inferred to be easy to generate mismatch, and thus the situation is manifested as drifting and conclusion repetition, on the other hand, the engineering of the knowledge graph stays on the static modeling or relation query level, even if structural connection and dependent links exist in the graph, the relation existence is difficult to be converted into the relation establishment on the streaming time sequence, and when the sampling jitter, the control swing and the short-term disturbance exist in the field data, the analysis result is more easily influenced by transient fluctuation and lacks continuous consistency, so that the operation and maintenance staff are difficult to execute effective running adjustment or energy saving control according to the situation. Therefore, the key difficulty of the energy efficiency root cause analysis of the industrial equipment is not the lack of data or relationship description, but the fact that how to combine the corresponding relationship between the engineering causal boundary provided by the knowledge graph and the event on the time axis in the streaming computing environment, to process the dislocation problem caused by uncertain delay and transient disturbance, and to stably convert the analysis conclusion into executable optimization action, which is the core problem that is not effectively solved in the prior art. Disclosure of Invention The invention aims to design an industrial equipment energy efficiency root cause analysis and optimization method and system based on knowledge graph and stream calculation, which can realize fusion of causal boundary of the knowledge graph and stream time sequence alignment, so that a root cause analysis conclusion has continuous consistency and engineering operability. In order to achieve the above object, in a first aspect of the present invention, there is provided an industrial equipment energy efficiency root cause analysis and optimization method based on knowledge graph and stream computation, the method comprising: extracting candidate equipment nodes and energy efficiency related nodes from an industrial equipment knowledge graph, wherein the candidate equipment nodes and target equipment nodes have structural connection, operation dependence or mechanism action relation; counting the number of the effective relations and carrying out normalization calculation to obtain the energy efficiency causal constraint weight; acquiring operation flow data of industrial equipment, and calculating operation variation amplitude and persistence factors in a sliding time window; calculating original event triggering intensity, wherein the original event triggering intensity is the product of the energy efficiency causal constraint weight, the operation change amplitude and the per