An artificial intelligence-based automobile production line equipment energy efficiency monitoring and optimization method
By constructing a set of task time period structure items and equipment power consumption curves, identifying power mutation segments and generating risk records, the problem of tracing the source of energy consumption anomalies was solved, and the refined analysis of task time periods and the synchronous scheduling of conflicting tasks were realized, thereby improving the energy efficiency of the automobile production line.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- PRIME MINISTER YUN (SHANGHAI) INFORMATION TECHNOLOGY CO LTD
- Filing Date
- 2026-01-28
- Publication Date
- 2026-05-26
AI Technical Summary
Existing methods struggle to trace abnormal energy consumption in automotive production line equipment back to specific task stages, resulting in a lack of targeted task scheduling and load allocation, as well as a lack of automated collaborative scheduling for conflicting tasks, thus reducing energy efficiency.
By constructing a set of task time period structure items and device power consumption curves, power mutation segments are identified, risk records are generated, and scheduling feature groups are constructed based on the task time period conflict graph. These are then input into the synchronous scheduling model to generate device scheduling instructions.
It enables refined analysis of power changes during task time periods in the production process, improves the ability to capture abnormal power fluctuations and synchronize load coordination of conflicting tasks, and optimizes the overall energy efficiency of the equipment.
Smart Images
Figure CN122085907A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of energy efficiency control technology, specifically to an artificial intelligence-based method for monitoring and optimizing the energy efficiency of automotive production line equipment. Background Technology
[0002] In automobile production lines, energy costs account for a large proportion of overall production and operating expenses. Existing methods generally involve deploying power monitoring modules on various equipment to collect power data and combining this data with task status information generated by the production execution system to calculate the equipment's energy efficiency level, thereby achieving a macro-level assessment of the equipment's long-term energy consumption. However, these methods typically only analyze the equipment as a whole, making it difficult to link power data to the specific task operation process in a fine-grained manner. This results in the inability to accurately identify abnormal power fluctuations within a task phase, thus lacking targeted handling of energy consumption anomalies in task scheduling and load allocation. To address the issue of anomalies in energy consumption not being traced back to specific task phases, existing methods have begun to attempt to identify fluctuations in the power behavior of devices during task switching phases and to correlate power fluctuations with device status records using timestamps, thereby improving the detection capability of abnormal power consumption. However, due to the lack of joint analysis of the intensity and duration of power fluctuations during task switching phases, these methods cannot effectively distinguish between reasonable fluctuations caused by sudden load increases and redundant fluctuations caused by resource waste, thus making it difficult to provide accurate risk basis for subsequent energy efficiency optimization. Furthermore, existing methods typically rely on manual experience to formulate task scheduling strategies after detecting energy consumption anomalies. They lack the overall modeling capability for the conflict relationships between tasks across multiple devices over time periods, making it impossible to achieve automated collaborative scheduling of conflicting tasks. This results in the task scheduling process struggling to balance energy consumption and load efficiency, thereby reducing the overall energy efficiency of the production line. Therefore, there is an urgent need to propose a method for monitoring and optimizing the energy efficiency of automotive production line equipment that can achieve detailed analysis of energy consumption behavior during task time periods, risk marking of abnormal fluctuations, and synchronized scheduling of conflicting tasks. Summary of the Invention
[0003] To address the aforementioned technical problems, this invention provides an artificial intelligence-based method for monitoring and optimizing the energy efficiency of automotive production line equipment, the method comprising: S11: Obtain power time series data and equipment task status data of various types of automobile production line equipment in their respective master control cycles. Based on the power time series data and equipment task status data, construct a set of task time period structure items and equipment power consumption curves. S12, Based on the power consumption curve of the device, identify the power mutation segment during the task switching period, calculate and generate the score result corresponding to each segment, and filter the power redundancy mutation segment according to the score result to generate a risk record containing the task number, device number and mutation duration. S13, Generate a task time period conflict diagram that marks task time periods by using the task time period structure item set and the risk record; S14. Based on the task time period conflict diagram and the task time period structure item set, construct a scheduling feature group, and input the scheduling feature group into a predetermined synchronous scheduling model to generate equipment scheduling instructions.
[0004] Furthermore, the steps for constructing the task time period structure item set and the device power consumption curve include: S111, by using the start and end timestamps corresponding to different task numbers in the device task status data, perform task segment extraction operation to generate a set of task time periods; S112, Perform segmented statistical processing based on the task time period set, obtain the triplet of power mean, power gradient change value and duration corresponding to each task time period, and construct a triplet set based on multiple triplets; S113, based on the set of triples, performs threshold interval merging processing and outputs a set of task time period calibration intervals; S114, perform segment mapping operation based on the task time period calibration interval set and power time series data to generate the task time period structure item set and device power consumption curve.
[0005] Furthermore, the steps for generating the task time period set are as follows: a1. Based on the start and end timestamps corresponding to each task number in the device task status data, construct a mapping record from task number to time index. a2, extract the start and end time ranges sequentially through the mapping record from task number to time index, and mark the power sampling segment corresponding to each task time period; a3 combines all task numbers with their corresponding power sampling segments to construct a task time period set.
[0006] Furthermore, the logic for generating the task time period structure item set and the device power consumption curve is as follows: b1, based on each item in the task time period calibration interval set, extract the task number, start timestamp and end timestamp contained therein, and construct the task time window set; b2, Match and compare the sampling timestamps in the power time series data with the task time window set, classify each power sampling point into the task time window of its corresponding task time period, and construct a set of field combination items; b3, Based on the triplet corresponding to each task time period in the task time period calibration interval set, and combined with its task time period number and task number, construct a set of task time period structure items; b4 generates a device power consumption curve based on the task time window set and power time series data.
[0007] Furthermore, the logic for filtering power consumption redundancy mutation segments based on the scoring results is as follows: S121, Based on the continuous power sampling points during the task switching time period in the device power consumption curve, construct a set of candidate power change segments; S122, For each candidate segment in the set of candidate power mutation segments, calculate the power gradient change value and duration; S123, Based on the power gradient change value and duration, a scoring result is generated; S124 outputs the power mutation candidate segments whose scoring results exceed the predetermined first scoring threshold as power redundancy mutation segments.
[0008] Furthermore, the step of generating the scoring result is as follows: c1 extracts the power gradient change value and duration value of each candidate segment in the power mutation candidate segment set; c2, input the power gradient change value and duration value into the predetermined scoring function to generate the score corresponding to each candidate segment; c3 binds the score to the task number, device number, and time index of the corresponding candidate segment, generates a score record, and outputs the score record as the score result.
[0009] Furthermore, the steps for generating a risk record containing the task number, device number, and duration of the mutation are as follows: S125, based on the task number, device number and time index in the power redundancy mutation segment, extract the mutation duration; S126 integrates the task number, device number, and mutation duration, and outputs a risk record.
[0010] Furthermore, the steps for generating a task time period conflict map with marked task time periods are as follows: S131, Based on the task number, device number, start timestamp, and end timestamp in the task time period structure item set, construct a task time period mapping index table; S132, based on the device number and task number in the power redundancy mutation section, and by referring to the task time period reference index table, filter out task time period pairs with overlapping time intervals. The logic for filtering task time period pairs with overlapping time intervals is as follows: d1: Extract the start and end timestamps of each task time period from the task time period structure item set, compare the start and end timestamps of any two task time periods in turn, and perform time overlap judgment processing. d2, record task time interval pairs with a time interval overlap length greater than zero, and output the set of all task time interval pairs that meet the time overlap condition, the set of task time interval pairs includes multiple task time interval pairs; S133: Based on task time period pairs with overlapping time intervals, extract the task number, device number, and task time period number corresponding to each pair of task time periods, and generate task time period connection edge fields; use all task time period numbers as a graph node set, use the task time period connection edge fields as a graph edge set, and output the task time period conflict graph.
[0011] Furthermore, the step of constructing the scheduling feature group is as follows: S141, extract the task number, task time period number, and power gradient change value for each task time period in the task time period structure item set; S142, perform association matching between the task time period number in the task time period conflict diagram and the task time period structure item set to confirm the task time period that participated in the construction of the scheduling feature group; S143, combine the task number, task time period number and power gradient change value into a scheduling feature group structure item; S144: Input all scheduling feature group structure items into the preset synchronous scheduling model to generate device scheduling instructions.
[0012] Furthermore, the generation steps of the synchronization scheduling model include: Acquire historical task time period scheduling data, and divide the historical task time period scheduling data into training set and test set. The historical task time period scheduling data includes a set of task time period structure items, a task time period conflict diagram and corresponding device scheduling instructions. A synchronous scheduling classification network is constructed. The task time period structure item set and task time period conflict graph in the training set are used as input to the synchronous scheduling classification network, and the corresponding device scheduling instructions are used as output to train the synchronous scheduling classification network to obtain the initial device scheduling instruction classification network. The initial equipment scheduling instruction classification network is validated using a test set. The initial equipment scheduling instruction classification network with a classification accuracy greater than or equal to the preset scheduling accuracy threshold is used as the synchronous scheduling model.
[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention constructs a set of task time period structure items and equipment power consumption curves based on power time series data and equipment task status data, so as to achieve refined analysis of power change behavior in each task time period during the production process, thereby improving the traceability of power changes during task switching phases and enhancing the ability to capture abnormal power fluctuations. In addition, the present invention also constructs a risk record by identifying candidate power mutation segments based on the device power consumption curve and combining them with the score results generated by the calculation, so as to enable targeted marking of task time periods with power consumption redundancy fluctuations during task switching, thereby improving the accuracy of task time period conflict determination and enhancing the ability to analyze energy consumption interference between task time periods. Furthermore, this invention constructs a scheduling feature group based on the task time period conflict diagram and the task time period structure item set, and inputs it into the synchronous scheduling model to generate equipment scheduling instructions, so as to realize the synchronous load coordination of task time periods with conflicting relationships, thereby improving the energy efficiency balance of task allocation, and thus optimizing the energy efficiency utilization level of the overall equipment of the automobile production line. In summary, this invention constructs an energy efficiency analysis system that integrates power time series data and task status data, enabling accurate identification of energy consumption anomalies and dynamic scheduling of conflicting tasks during the operation of automotive production line equipment. This ensures energy efficiency balance while improving overall scheduling coordination and resource utilization. Attached Figure Description
[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.
[0015] Figure 1 The flowchart illustrates an artificial intelligence-based method for monitoring and optimizing the energy efficiency of automotive production line equipment, as provided in Embodiment 1 of the present invention. Detailed Implementation
[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0017] Please see Figure 1 As shown in the figure, this embodiment discloses a method for monitoring and optimizing the energy efficiency of automobile production line equipment based on artificial intelligence. The method includes: S11: Obtain power time series data and equipment task status data of various types of automobile production line equipment in their respective master control cycles. Based on the power time series data and equipment task status data, construct a set of task time period structure items and equipment power consumption curves. It should be noted that the power time series data is based on the sampling period. The power sampling points are recorded sequentially, with each power sampling point corresponding to a power value. The power is collected in real time by intelligent power monitoring modules installed on various equipment on the automobile production line. Each device is equipped with an independent sampling loop. The sampled power value is the actual power consumption of the device per unit time. The sampled power values are arranged in chronological order to form a continuous power time series data, which is then uploaded to the production energy consumption scheduling platform for recording and analysis. The equipment task status data is generated through the task instruction issuance and feedback mechanism in the Production Execution System (MES). The recorded content includes task number, task issuance time, task execution start and end time, equipment number to which the task belongs, and task status field. This data can be associated and matched with power data by timestamp and equipment number.
[0018] Specifically, the steps for constructing the task time period structure item set and device power consumption curve include: S111, by using the start and end timestamps corresponding to different task numbers in the device task status data, perform task segment extraction operation to generate a set of task time periods; Specifically, the steps for generating the task time period set are as follows: a1. Based on the start and end timestamps corresponding to each task number in the device task status data, construct a mapping record from task number to time index. Equipment task status data is automatically generated by the MES system. Whenever the system dispatches a task to a production device, it generates a unique task number, along with the task dispatch time and estimated end time. After the task is completed, the equipment controller reports the actual completion time to the MES system, forming a complete task execution record; The task number is paired with its corresponding actual start and end timestamps to form a mapping record from task number to time index.
[0019] a2, extract the start and end time ranges sequentially through the mapping record from task number to time index, and mark the power sampling segment corresponding to each task time period; Specifically, the start and end times of each task number are analyzed sequentially, and these are used as an index window to extract all power values within the corresponding time interval from the power time series data. It should be noted that this extraction operation uses a time-based matching window. All power values falling within this interval will be assigned to the power sampling segment of this task and bound to the original task number.
[0020] a3, combine all task numbers with their corresponding power sampling segments to construct a task time period set; In a specific embodiment, the structure of the task formed by combining each task number extracted in step a2 with the structure of its corresponding power sampling segment is as follows: Task Number:
[0021] Start timestamp:
[0022] End timestamp:
[0023] Power sampling segment:
[0024] Combine the structures of all tasks in list form to construct a set of task time periods.
[0025] S112, Perform segmented statistical processing based on the task time period set, obtain the triplet of power mean, power gradient change value and duration corresponding to each task time period, and construct a triplet set based on multiple triplets; In one specific embodiment, the power sampling segment of each record in the task time period set is processed as follows: The average power value of all power values in the power sampling segment is calculated to generate the average power value for the task time period. Calculate the difference between any two adjacent power values, and divide by . The power change rate is generated by averaging all power change rates within the task time period to generate the average change rate, which is then used as the power gradient change value. The duration of the task time period is generated by subtracting the start timestamp from the end timestamp of the task time period. The power mean, power gradient change, and duration are combined into a triplet structure, represented as: ; Combine the triple structures of all task time periods to generate a set of triples; S113, based on the set of triples, performs threshold interval merging processing and outputs a set of task time period calibration intervals; In a specific embodiment, the merging determination parameters are based on preset merging determination parameters, which include a mean tolerance threshold, a gradient change threshold, and a duration difference threshold. Iterate through the sets of triplets for consecutive task time periods in the power triplet set. If the difference between each component in the triplet set of adjacent task time periods is less than the corresponding threshold, then classify them into the same task time period calibration interval. This can be understood as: performing interpolation on two adjacent task time periods to generate two task time periods. The difference between the values of the power gradient change and the duration; The difference between the power gradient change value and the duration is compared with the corresponding mean tolerance threshold, gradient change threshold and duration difference threshold. If they are less than the mean tolerance threshold, gradient change threshold and duration difference threshold, they are classified into the same task time period calibration interval. If multiple task time periods exhibit stable power, slow fluctuations, and similar durations, they are considered to be of the same type; and all time period divisions are summarized into a task time period calibration interval set. It should be noted that the merging decision parameters were obtained based on historical experimental data analysis. S114, Perform segment mapping operation based on the task time period calibration interval set and power time series data to generate the task time period structure item set and device power consumption curve; Specifically, the logic for generating the task time period structure item set and the device power consumption curve is as follows: b1, based on each item in the task time period calibration interval set, extract the task number, start timestamp and end timestamp contained therein, and construct the task time window set; b2, Match and compare the sampling timestamps in the power time series data with the task time window set, classify each power sampling point into the task time window of its corresponding task time period, and construct a set of field combination items; Specifically, for each sampling point in the power time series data If its sampling timestamp The task time window that falls within a certain task time period [ If the sampling point is matched with the corresponding task number, then the sampling point will be assigned to the task number. and task time period number Bind the fields to form a composite item: Represented as: It should be noted that: For the first The start timestamp of each task's time period. For the first The end timestamp of each task's time period; Generate a set of field combination items based on multiple sets of field combination items; b3, Based on the triplet corresponding to each task time period in the task time period calibration interval set, and combined with its task time period number and task number, construct a set of task time period structure items; Each task time period number is associated with its corresponding average power, power gradient change value, and duration, and a task number is added to form a set of task time period structure items, represented as follows:
[0026] In the task time period structure item set, For the first The average power over a given task time period For the first The power gradient change value over a task time period. For the first The duration of each task's time period; b4 generates a device power consumption curve based on the task time window set and power time series data; Specifically, based on the start and end timestamps of each item in the task time window set as the judgment interval, all power sampling points whose sampling timestamps fall within the judgment interval are selected from the power time series data; For each power sampling point that falls within the judgment interval, obtain the corresponding task number and task time period number to mark the task time period; The labeled power sampling points are integrated in ascending order of sampling timestamps. A continuous power output trend is constructed using a time-power mapping structure to generate a power response trajectory. All power sampling points in the power response trajectory are segmented according to their corresponding task number and task time period number, and the task time period is used as the segment boundary to generate a power response trajectory with a clear stage affiliation, which serves as the device power consumption curve for the device.
[0027] S12, Based on the power consumption curve of the device, identify the power mutation segment during the task switching period, calculate and generate the score result corresponding to each segment, and filter the power redundancy mutation segment according to the score result to generate a risk record containing the task number, device number and mutation duration. Specifically, the logic for filtering power consumption redundancy mutation segments based on the scoring results is as follows: S121, Based on the continuous power sampling points during the task switching time period in the device power consumption curve, construct a set of candidate power change segments; It should be noted that: the task switching period is determined by the start and end time index of the adjacent task numbers in the device task status data; each power change candidate segment contains a set of continuous power sampling points and their timestamps, which are used to describe the power change behavior during task switching; S122, For each candidate segment in the set of candidate power mutation segments, calculate the power gradient change value and duration; For each power change candidate segment in the power change candidate segment set, calculate the power value difference between any two adjacent power sampling points in sequence, and divide it by the sampling period to obtain the power change rate. Represented as: In the formula, For each candidate segment, the power gradient change value and duration are given. The sampling period is For the first Power values at each power sampling point For the first Power values at each power sampling point The total number of candidate segments; It should be noted that the power gradient change value in this step is the same as the power gradient calculation value mentioned above. This is repeated here only to avoid confusion. The duration is calculated by subtracting the start timestamp of the segment from the end timestamp of the candidate segment. Represented as: In the formula, For duration, This is the end timestamp of the candidate segment. This is the start timestamp of the candidate segment; S123, Based on the power gradient change value and duration, a scoring result is generated; Specifically, the steps for generating the scoring results are as follows: c1 extracts the power gradient change value and duration value of each candidate segment in the power mutation candidate segment set; c2, input the power gradient change value and duration value into the predetermined scoring function to generate the score corresponding to each candidate segment; c3 binds the score with the task number, device number and time index of the corresponding candidate segment, generates a score record, and outputs the score record as the score result; Represented as: In the formula, For the first The scoring record of each power sampling point For the first The duration of each power sampling point For duration reference, for The power gradient change value at each power sampling point As a reference quantity for power gradient, ; It should be noted that: power gradient reference quantity and duration reference The optimal settings are determined by the operator based on analysis of historical parameters; S124 outputs the power mutation candidate segments whose scoring results exceed the predetermined first scoring threshold as power redundancy mutation segments.
[0028] It should be noted that the first scoring threshold is determined by the production energy consumption scheduling platform based on historical samples and validation set performance, and can be set according to the comprehensive objectives of coverage and false alarm rate. Specifically, the steps for generating a risk record that includes the task number, device number, and duration of the mutation are as follows: S125, based on the task number, device number and time index in the power redundancy mutation segment, extract the mutation duration; S126 integrates the task number, device number, and mutation duration, and outputs a risk record; In a specific embodiment, the task number, device number and duration of each power redundancy mutation segment are integrated into a risk record and output as a risk record set. This step can be understood as follows: This step realizes the structured implementation from "curve-level anomaly segments" to "task and device dimension risk items", forming data units that can be directly consumed by the scheduling and alarm system.
[0029] S13, Generate a task time period conflict diagram that marks task time periods by using the task time period structure item set and the risk record; Specifically, the steps to generate a task time period conflict graph with marked task time periods are as follows: S131, Based on the task number, device number, start timestamp, and end timestamp in the task time period structure item set, construct a task time period mapping index table; It should be noted that the task time period reference index table provides an index range for subsequent time interval overlap judgment, which is used to quickly locate the task time periods corresponding to adjacent task numbers under the same device number; S132, based on the device number and task number in the power redundancy mutation section, and by referring to the task time period reference index table, filter out task time period pairs with overlapping time intervals. The logic for filtering task time period pairs with overlapping time intervals is as follows: d1: Extract the start and end timestamps of each task time period from the task time period structure item set, compare the start and end timestamps of any two task time periods in turn, and perform time overlap judgment processing. In a specific embodiment, if the start timestamp of the first task time period is less than the end timestamp of the second task time period, and the end timestamp of the first task time period is greater than the start timestamp of the second task time period, then it is determined that the two time intervals overlap. Represented as: In the formula, The length of the overlap between time intervals. and , These are the start and end timestamps for the two task time periods, respectively. If the overlap length of the time intervals is greater than 0, then it is determined that there is an overlap of time intervals; d2, record task time interval pairs with a time interval overlap length greater than zero, and output the set of all task time interval pairs that meet the time overlap condition, the set of task time interval pairs includes multiple task time interval pairs; S133: Based on task time period pairs with overlapping time intervals, extract the task number, device number, and task time period number corresponding to each pair of task time periods, and generate task time period connection edge fields; use all task time period numbers as a graph node set, use the task time period connection edge fields as a graph edge set, and output the task time period conflict graph.
[0030] S14, construct a scheduling feature group based on the task time period conflict diagram and the task time period structure item set, and input the scheduling feature group into a predetermined synchronous scheduling model to generate equipment scheduling instructions; Specifically, the steps for constructing the scheduling feature group are as follows: S141, extract the task number, task time period number, and power gradient change value for each task time period in the task time period structure item set; S142, perform association matching between the task time period number in the task time period conflict diagram and the task time period structure item set to confirm the task time period that participated in the construction of the scheduling feature group; S143, combine the task number, task time period number and power gradient change value into a scheduling feature group structure item; S144, input all scheduling feature group structure items into the preset synchronous scheduling model to generate equipment scheduling instructions; Specifically, the steps for generating the synchronization scheduling model include: Acquire historical task time period scheduling data, and divide the historical task time period scheduling data into training set and test set. The historical task time period scheduling data includes a set of task time period structure items, a task time period conflict diagram and corresponding device scheduling instructions. It should be noted that: the equipment scheduling instruction consists of discrete label fields such as priority level, execution order number, and start time period number; A synchronous scheduling classification network is constructed. The task time period structure item set and task time period conflict graph in the training set are used as input to the synchronous scheduling classification network, and the corresponding device scheduling instructions are used as output to train the synchronous scheduling classification network to obtain the initial device scheduling instruction classification network. The initial equipment scheduling instruction classification network is validated using a test set. The initial equipment scheduling instruction classification network with a classification accuracy greater than or equal to the preset scheduling accuracy threshold is used as the synchronous scheduling model. It should be noted that synchronous scheduling classification networks include, but are not limited to, models such as Temporal Graph Network, GraphConvolutional Network, or Long Short-Term Memor as the basic structure, used to simultaneously handle task time constraints and load characteristics.
[0031] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired or wireless network. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.
[0032] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only one method, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0033] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0034] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0035] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.
[0036] The above embodiments are only used to illustrate the technical methods 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 methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. An artificial intelligence-based automobile production line equipment energy efficiency monitoring and optimization method, characterized in that, The method includes: S11: Obtain power time series data and equipment task status data of various types of automobile production line equipment in their respective master control cycles. Based on the power time series data and equipment task status data, construct a set of task time period structure items and equipment power consumption curves. S12, Based on the power consumption curve of the device, identify the power mutation segment during the task switching period, calculate and generate the score result corresponding to each segment, and filter the power redundancy mutation segment according to the score result to generate a risk record containing the task number, device number and mutation duration. S13, Generate a task time period conflict diagram that marks task time periods by using the task time period structure item set and the risk record; S14. Based on the task time period conflict diagram and the task time period structure item set, construct a scheduling feature group, and input the scheduling feature group into a predetermined synchronous scheduling model to generate equipment scheduling instructions.
2. The artificial intelligence-based automobile production line equipment energy efficiency monitoring and optimization method according to claim 1, characterized in that, The steps for constructing the task time period structure item set and device power consumption curve include: S111, by using the start and end timestamps corresponding to different task numbers in the device task status data, perform task segment extraction operation to generate a set of task time periods; S112, Perform segmented statistical processing based on the task time period set, obtain the triplet of power mean, power gradient change value and duration corresponding to each task time period, and construct a triplet set based on multiple triplets; S113, based on the set of triples, performs threshold interval merging processing and outputs a set of task time period calibration intervals; S114, perform segment mapping operation based on the task time period calibration interval set and power time series data to generate the task time period structure item set and device power consumption curve. 3.The method of claim 2, wherein, The steps for generating the task time period set are as follows: a1. Based on the start and end timestamps corresponding to each task number in the device task status data, construct a mapping record from task number to time index. a2, extract the start and end time ranges sequentially through the mapping record from task number to time index, and mark the power sampling segment corresponding to each task time period; a3 combines all task numbers with their corresponding power sampling segments to construct a task time period set.
4. The artificial intelligence-based automobile production line equipment energy efficiency monitoring and optimization method according to claim 3, characterized in that, The logic for generating the task time period structure item set and the device power consumption curve is as follows: b1, based on each item in the task time period calibration interval set, extract the task number, start timestamp and end timestamp contained therein, and construct the task time window set; b2, Match and compare the sampling timestamps in the power time series data with the task time window set, classify each power sampling point into the task time window of its corresponding task time period, and construct a set of field combination items; b3, Based on the triplet corresponding to each task time period in the task time period calibration interval set, and combined with its task time period number and task number, construct a set of task time period structure items; b4 generates a device power consumption curve based on the task time window set and power time series data.
5. The artificial intelligence-based automobile production line equipment energy efficiency monitoring and optimization method according to claim 4, characterized in that, The logic for filtering power redundancy mutation segments based on the scoring results is as follows: S121, Based on the continuous power sampling points during the task switching time period in the device power consumption curve, construct a set of candidate power change segments; S122, For each candidate segment in the set of candidate power mutation segments, calculate the power gradient change value and duration; S123, Based on the power gradient change value and duration, a scoring result is generated; S124 outputs the power mutation candidate segments whose scoring results exceed the predetermined first scoring threshold as power redundancy mutation segments.
6. The method for energy efficiency monitoring and optimization of automobile production line equipment based on artificial intelligence according to claim 5, characterized in that, The steps for generating the scoring results are as follows: c1 extracts the power gradient change value and duration value of each candidate segment in the power mutation candidate segment set; c2, input the power gradient change value and duration value into the predetermined scoring function to generate the score corresponding to each candidate segment; c3 binds the score to the task number, device number, and time index of the corresponding candidate segment, generates a score record, and outputs the score record as the score result.
7. The method for energy efficiency monitoring and optimization of automobile production line equipment based on artificial intelligence according to claim 6, characterized in that, The steps to generate a risk record containing the task number, device number, and duration of the mutation are as follows: S125, based on the task number, device number and time index in the power redundancy mutation segment, extract the mutation duration; S126 integrates the task number, device number, and mutation duration, and outputs a risk record.
8. The method for energy efficiency monitoring and optimization of automobile production line equipment based on artificial intelligence according to claim 7, characterized in that, The steps to generate a task time period conflict map with marked task time periods are as follows: S131, Based on the task number, device number, start timestamp, and end timestamp in the task time period structure item set, construct a task time period mapping index table; S132, based on the device number and task number in the power redundancy mutation section, and by referring to the task time period reference index table, filter out task time period pairs with overlapping time intervals. The logic for filtering task time period pairs with overlapping time intervals is as follows: d1: Extract the start and end timestamps of each task time period from the task time period structure item set, compare the start and end timestamps of any two task time periods in turn, and perform time overlap judgment processing. d2, record task time interval pairs with a time interval overlap length greater than zero, and output the set of all task time interval pairs that meet the time overlap condition, the set of task time interval pairs includes multiple task time interval pairs; S133: Based on task time period pairs with overlapping time intervals, extract the task number, device number, and task time period number corresponding to each pair of task time periods, and generate task time period connection edge fields; use all task time period numbers as a graph node set, use the task time period connection edge fields as a graph edge set, and output the task time period conflict graph.
9. The method for energy efficiency monitoring and optimization of automobile production line equipment based on artificial intelligence according to claim 8, characterized in that, The steps for constructing the scheduling feature group are as follows: S141, extract the task number, task time period number, and power gradient change value for each task time period in the task time period structure item set; S142, perform association matching between the task time period number in the task time period conflict diagram and the task time period structure item set to confirm the task time period that participated in the construction of the scheduling feature group; S143, combine the task number, task time period number and power gradient change value into a scheduling feature group structure item; S144: Input all scheduling feature group structure items into the preset synchronous scheduling model to generate device scheduling instructions.
10. The method for energy efficiency monitoring and optimization of automobile production line equipment based on artificial intelligence according to claim 9, characterized in that, The steps for generating the synchronization scheduling model include: Acquire historical task time period scheduling data, and divide the historical task time period scheduling data into training set and test set. The historical task time period scheduling data includes a set of task time period structure items, a task time period conflict diagram and corresponding device scheduling instructions. A synchronous scheduling classification network is constructed. The task time period structure item set and task time period conflict graph in the training set are used as input to the synchronous scheduling classification network, and the corresponding device scheduling instructions are used as output to train the synchronous scheduling classification network to obtain the initial device scheduling instruction classification network. The initial equipment scheduling instruction classification network is validated using a test set. The initial equipment scheduling instruction classification network with a classification accuracy greater than or equal to the preset scheduling accuracy threshold is used as the synchronous scheduling model.