An intelligent detection and diagnosis system for abnormal events in particleboard production

Abstract

The present invention relates to the field of particleboard production technology, and specifically to an intelligent detection and diagnosis system for abnormal events in particleboard production. The system includes a parameter time series acquisition module, an abnormal point identification module, a collaborative feature determination module, a synchronous trend analysis module, and an abnormal attribution and positioning module. The present invention establishes a dynamic linkage system of batch information and parameter fluctuations by performing continuous trend analysis and node sequence integration on various process parameters in the production process, realizes the synchronous identification of abnormal parameters and nodes, and uses the collaborative mutation of parameters and trend synchronization offset between process links as the core basis for abnormal diagnosis, improves the comprehensive accuracy of batch attribution, abnormal type and node positioning, enhances the ability to track abnormal events under multi-node and multi-parameter conditions, optimizes the integration of parameter data in the entire production process and abnormality traceability, reduces the quality risks caused by parameter fluctuations and process abnormalities, and optimizes the risk management and control capabilities in the entire production process.

Description

Technical Field

[0001] The present invention relates to the technical field of particleboard production, and in particular to an intelligent detection and diagnosis system for abnormal events in particleboard production. Background Art

[0002] The field of particleboard production mainly involves a series of process links such as wood raw material pretreatment, particle preparation, particle drying, gluing, laying and forming, hot pressing and curing, cooling and humidity control, surface treatment and finished product grading. This field focuses on core issues such as raw material selection, process parameter control, product quality inspection and production process management, covering all stages of the production process, and puts forward high requirements for the automation and intelligence of the process. Among them, the traditional particleboard production abnormal event intelligent detection and diagnosis system refers to the equipment operation abnormality, process parameter fluctuations, raw material ratio deviation, board defect problems in the production process. It usually adopts on-site manual inspections, manual recording and analysis of production data, fixed threshold alarms, static monitoring of process parameters, regular sample inspections, etc. to identify and judge, and preliminarily determine the type of abnormality and its cause through manual experience and statistical analysis.

[0003] Existing technologies rely on manual data collection, static parameter monitoring, and empirical judgment in the particleboard production process. It is difficult to map process parameter fluctuations and batch changes in real time, multi-node collaborative anomalies cannot be automatically perceived, and the on-site alarm method is limited to a single threshold setting. Production batch traceability is not comprehensive, and parameter or equipment anomalies in some process links are difficult to respond to in a timely manner, resulting in missed detection and misjudgment of abnormal events, unclear batch positioning, and affecting production stability and subsequent quality assurance. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an intelligent detection and diagnosis system for abnormal events in particleboard production.

[0005] In order to achieve the above-mentioned object, the present invention adopts the following technical solution: an intelligent detection and diagnosis system for abnormal events in particleboard production, the system comprising:

[0006] The parameter time series acquisition module is based on the wood shavings moisture content sampling channel, analyzes its change trend, compares the relationship between the sizing amount and the batch number, calculates the correspondence between the hot pressing temperature and the molding speed, determines the matching status of the process nodes, and integrates the time sequence of each parameter acquisition to obtain a parameter time series set;

[0007] The abnormal point identification module analyzes the fluctuation of the moisture content of the wood chips based on the parameter time series set, determines the corresponding relationship between the paving density and the batch number, compares the variation range of the hot pressing temperature and the forming speed, identifies the abnormal fluctuation of the batch parameters, and obtains the distribution characteristics of the abnormal parameters;

[0008] The collaborative feature determination module compares the change direction of the paving density and the spindle speed based on the abnormal parameter distribution characteristics, analyzes the inlet parameter combination of the wood shavings paving equipment and the hot pressing system, determines the collaborative state of the process parameters, and obtains the process collaborative mutation index;

[0009] The synchronization trend analysis module analyzes the changing trend of the glue application amount based on the process coordination mutation index, calculates the trend difference between the paving pressure and the spindle speed, compares the synchronization of process parameter fluctuations, determines the consistency of node trends, and obtains the trend offset characteristic value.

[0010] The improvements of the present invention are that the parameter time series set includes parameter type identification, time node mark, and sequence integrity information; the abnormal parameter distribution characteristics include abnormal type identification, abnormal occurrence node, and abnormal duration interval; the process collaborative mutation index includes collaborative abnormality identification, parameter linkage mode, and associated node characteristics; the trend offset feature includes trend consistency index, offset degree classification, and trend change identification.

[0011] The present invention is improved in that the parameter timing acquisition module includes:

[0012] The moisture content analysis submodule analyzes the moisture content changes of wood chips within a continuous time period based on the wood chip moisture content sampling channel. By classifying the direction changes of the collected signals at adjacent time nodes and comparing the growth or decrease in different time periods, the continuity and consistency of the moisture content changes are determined, and the fluctuation performance is screened to obtain the moisture content trend characteristics.

[0013] The glue attribution identification submodule compares the time distribution of the glue usage data and the corresponding batch number based on the moisture content trend characteristics, determines whether there is overlap or offset between the attribution of each set of glue usage information and the batch number, analyzes the matching of batch and glue data, identifies abnormal attribution relationships, and obtains glue attribution characteristics;

[0014] The parameter linkage integration submodule compares the changes of hot pressing temperature and molding speed at the same time node according to the glue attribution characteristics, determines the consistency of the change direction of each parameter, screens the linkage performance, and integrates the parameters and time nodes to obtain a parameter time series set.

[0015] The present invention is improved in that the outlier identification module includes:

[0016] The moisture content fluctuation identification submodule analyzes the variation trend of the wood shavings moisture content data in the target time period based on the parameter time series set, determines the variation direction and amplitude between continuous time series data, identifies segments where the amplitude of variation continuously rises or falls, and locates the time segments where concentrated variation occurs in the data sequence to obtain the moisture content mutation interval segment;

[0017] The parameter attribution relationship judgment submodule compares the corresponding relationship between the synchronized data of the glue application amount and the paving density and the batch number based on the moisture content mutation interval, determines whether the paving density sampling sequence is consistent with the batch of the mutation section, and obtains the parameter batch matching status distribution;

[0018] The batch fluctuation difference identification submodule compares the hot pressing temperature and forming speed of the corresponding batches according to the batch matching status distribution of the parameters, calculates the changes in the paving density and the spindle speed, calculates the fluctuation level of the parameters of each batch, and judges the fluctuation performance between batches to obtain the abnormal parameter distribution characteristics.

[0019] The present invention is improved in that the collaborative feature determination module includes:

[0020] The direction matching submodule compares the change direction of the paving density with the change direction of the spindle speed feedback signal based on the distribution characteristics of the abnormal parameters, determines whether the trends of the two are consistent within the same batch, optimizes the judgment criteria for trend offset, and obtains the trend direction matching coefficient;

[0021] The combination relationship submodule calls the trend matching coefficient, compares the synchronous change characteristics of paving speed, roller pressing gap, hot pressing initial temperature and plate feeding pressure, identifies the key parameter combination of synchronous change characteristics, optimizes the interaction relationship between parameters, and obtains the parameter combination change range;

[0022] The collaborative status submodule calculates the collaborative change level of the paving density change and the spindle speed offset based on the variation range of the parameter combination, compares the trend fluctuation and linkage frequency between nodes, and obtains the process collaborative mutation index.

[0023] The present invention is improved in that the synchronization trend analysis module includes:

[0024] The trend calculation submodule analyzes the time series data of the glue application amount based on the process synergy mutation index, determines the change direction of the glue application amount under each batch number, compares the change rate in the time period before and after the mutation node, and calculates the trend slope interval of the glue application amount of each batch over time;

[0025] The pressure-speed comparison submodule calculates the variation range of the paving pressure and the spindle speed in the same time period according to the trend slope interval, compares the synchronicity of the pressure and speed trends, and obtains the trend difference range;

[0026] The trend consistency determination submodule calls the trend difference amplitude, analyzes the number of intersections and synchronization intervals of each parameter trend on the time axis according to the distribution direction difference under the different process nodes, identifies the consistent segments of trend fluctuations between parameters, and obtains the trend offset feature.

[0027] The present invention is improved in that the system further comprises:

[0028] The abnormal attribution and positioning module determines the abnormal attribution of the sudden change node of wood shavings moisture content, paving density and spindle speed change based on the trend offset feature, identifies the batch and event information of the abnormal node, integrates the abnormal correlation information of the process parameters, and obtains the batch abnormality feature positioning information;

[0029] The batch abnormality feature location information includes a batch location identifier, an abnormal event type, and a node affiliation identifier.

[0030] The present invention is improved in that the abnormal attribution positioning module includes:

[0031] The abnormal fluctuation attribution submodule analyzes the fluctuation relationship between the sudden change node of the particle moisture content and the change of the paving density based on the trend deviation characteristic, compares the linkage between the spindle speed and other process parameters, identifies and classifies the nodes of abnormal fluctuation, and obtains the abnormal fluctuation index;

[0032] The batch positioning and integration submodule analyzes and integrates the event information and process parameters of each batch based on the abnormal fluctuation index, determines the batch association when the abnormal fluctuation occurs, locates the abnormal batch and process node, and obtains the batch abnormality feature positioning data.

[0033] Compared with the prior art, the advantages and positive effects of the present invention are:

[0034] In the present invention, by conducting continuous trend analysis and node sequence integration on various process parameters in the production process, a dynamic linkage system of batch information and parameter fluctuations is established, and the synchronous identification of abnormal parameters and nodes is realized. The coordinated mutation of parameters and trend synchronization offset between process links are used as the core basis for abnormal diagnosis, and the comprehensive accuracy of batch attribution, abnormal type and node positioning is improved. The ability to track abnormal events under multi-node and multi-parameter conditions is enhanced, and the integration of parameter data in the entire production process and abnormality traceability are optimized, thereby reducing the quality risks caused by parameter fluctuations and process abnormalities and optimizing the risk management and control capabilities in the entire production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a system flow chart of the present invention;

[0036] Figure 2 This is a flow chart of the parameter timing acquisition module in the present invention;

[0037] Figure 3 This is a flow chart of the outlier identification module in the present invention;

[0038] Figure 4 This is a flow chart of the collaborative feature determination module in the present invention;

[0039] Figure 5This is a flow chart of the synchronization trend analysis module in the present invention;

[0040] Figure 6 This is a flow chart of the abnormal attribution positioning module in the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0042] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the devices or elements referred to must have a specific direction, be constructed and operate in a specific direction, and therefore should not be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined. Example

[0043] See also Figure 1 The present invention provides a technical solution: an intelligent detection and diagnosis system for abnormal events in particleboard production, comprising:

[0044] The parameter time series acquisition module is based on the wood particle moisture content sampling channel. It analyzes the changing trend of the wood particle moisture content acquisition signal, compares the attribution between the sizing amount input channel data and the batch number, calculates the corresponding relationship between the hot pressing temperature change and the molding speed feedback signal, determines the matching status of the process node information, and integrates the time sequence of each parameter acquisition to obtain the parameter time series set.

[0045] The outlier identification module analyzes the fluctuation of wood particle moisture content within the target time period based on the parameter time series set, determines the correspondence between the paving density sampling sequence and the batch number, compares the variation range of hot pressing temperature and forming speed, identifies abnormal fluctuations between batch parameters, and obtains the distribution characteristics of abnormal parameters;

[0046] The collaborative feature determination module compares the change direction of the paving density data and the spindle speed feedback based on the distribution characteristics of abnormal parameters, analyzes the combined characteristics of the wood shavings paving equipment parameters and the parameters of the hot pressing system inlet area, determines the collaborative state between each set of process parameters, optimizes the correlation of node parameter combinations, and obtains the process collaborative mutation index;

[0047] The synchronization trend analysis module analyzes the changing trend of glue application based on the process coordination mutation index, calculates the trend difference between the paving pressure detection channel and the spindle speed control unit, compares the synchronization status of each process parameter fluctuation, determines the consistency between node trends, and obtains the trend offset characteristic value;

[0048] The abnormal attribution and positioning module analyzes the distribution information of batch abnormal data based on the trend offset feature, determines the abnormal attribution of the sudden change node of wood particle moisture content, the direction change of paving density and the combination of strong and weak spindle speed, identifies the batch and event information of the abnormal node, integrates the abnormal correlation information of process parameters, and obtains the batch abnormality feature positioning information.

[0049] The parameter time series set includes parameter type identification, time node mark, and sequence integrity information. The abnormal parameter distribution characteristics include abnormal type identification, abnormal occurrence node, and abnormal duration interval. The process collaborative mutation index includes collaborative abnormality identification, parameter linkage mode, and associated node characteristics. The trend offset feature includes trend consistency index, offset degree classification, and trend change identification. The batch abnormality feature location information includes batch location identification, abnormal event type, and node attribution identification.

[0050] In Module 1, the particle moisture content sampling channel refers to a dedicated collection line or equipment installed on the particleboard production line for real-time detection and recording of the moisture content of particleboard raw materials; the glue dosage input channel refers to a sensor, data interface or system path used to monitor and record the amount of adhesive added, which can collect the glue dosage used for each batch of materials; the batch number is a unique code assigned to each batch of raw materials or finished products entering the production process, which is used for production process tracking, parameter binding and traceability; the ownership situation refers to the state in which parameter data such as glue dosage and moisture content are mapped and archived one by one with the corresponding batch number in the data management system to ensure that each data point clearly belongs to a batch; the corresponding relationship is used to describe the mapping or association between two parameters, such as the mutual influence and correlation trend between hot pressing temperature changes and forming speed feedback over time; process node information refers to the relevant parameters, status, and identification information of key processes or process stages (such as gluing, paving, hot pressing, etc.) on the particleboard production line, which is used to clarify the location and link of data occurrence.

[0051] In Module 2, fluctuation performance refers to the continuous change of particle moisture content during the production process within a specific time period, including its dynamic trends such as growth, decrease or mutation; paving density refers to the mass density per unit volume or per unit area of ​​the particle raw materials laid on the slab during production, which is an important process parameter affecting the performance of the board; the corresponding situation refers to the attribution and distribution status of the paving density acquisition sequence under the batch number mapping, ensuring that the density data and production batch data can be matched and tracked with each other; the amplitude of change is used to describe the amount of change in the parameter within a specific time period or batch, such as the fluctuation range of the hot pressing temperature or forming speed; abnormal fluctuation refers to the abnormal and drastic changes in certain parameters (such as moisture content, hot pressing temperature) in a short period of time, which exceeds the normal production fluctuation range.

[0052] In Module 3, the paving density data is the time series data actually collected, reflecting the density changes in each batch of paving process; the spindle speed refers to the actual speed of the rotating shaft related to the main power output of the equipment, and is an important parameter for measuring the operating status of the molding equipment; the combined feature is to group the changing trends of multiple parameters such as paving density, spindle speed, hot pressing inlet temperature, etc. at the same node or in the same time segment, for comprehensive analysis of the process status; the collaborative status indicates whether multiple parameters (such as paving density and spindle speed) show similar trends, synchronous changes, mutual response or abnormal linkage when producing the same batch.

[0053] In module 4, the change trend refers to the change direction of a certain process parameter (such as the amount of glue applied) over the production time, which can be rising, falling, stable or sudden, etc.; the paving pressure detection channel refers specifically to a special sensor or signal acquisition circuit used to monitor the pressure changes in the paving process of wood shavings raw materials in real time; the spindle speed control unit is a device that adjusts the spindle speed and provides real-time feedback to ensure that the equipment speed is stable within the process requirements; the consistency between node trends is to compare whether the change trends of various parameters at different process nodes (such as paving, hot pressing) of the same batch are synchronized, convergent or different.

[0054] In Module 5, anomaly attribution is to locate abnormal fluctuations in parameters during the production process and determine which process node and batch the anomaly belongs to. An abnormal node refers to a production process link or process point that is determined to have undergone abnormal changes during parameter trend analysis or abnormal event tracing.

[0055] See also Figure 2 , the parameter timing acquisition module includes:

[0056] The moisture content analysis submodule analyzes the moisture content changes of wood chips within a continuous time period based on the wood chip moisture content sampling channel. By classifying the direction changes of the collected signals at adjacent time nodes and comparing the growth or decrease in different time periods, the continuity and consistency of the moisture content changes are determined, and the fluctuation performance is screened to obtain the moisture content trend characteristics.

[0057] Real-time data extraction is performed on the online moisture detection equipment installed on the particleboard conveyor line. During the execution process, the continuously collected data are first numbered point by point in chronological order and divided into several time periods. Each time period covers a fixed time length (such as 60 seconds). The direction of the numerical change trend between adjacent collection points in each section is classified. When the moisture content value at a certain time point is less than that of the next point, it is marked as an upward direction; if it is greater, it is marked as a downward direction; if the absolute value of the difference between the two is less than or equal to 0.2%, it is marked as stable. Then, the direction change sequence in multiple time periods is identified in turn, and the fluctuation segment is combined. The start and end time, change direction and change amplitude of each fluctuation segment are marked to determine whether the maximum change in the fluctuation segment exceeds 1.5%. This threshold is based on the The moisture content of the wood shavings raw material is set within a normal floating range of 10% to 13%. If it exceeds the normal floating range, it is classified as an abnormal segment, and the time interval between the fluctuation segments is calculated. If the interval between any two segments does not exceed 3 minutes, it is marked as continuous fluctuation. The directions of the continuous fluctuation segments are further compared to see if they are consistent. If the directions are all rising or all falling, they are considered to be consistent. The starting value and the ending value in the fluctuation segment are then extracted to calculate the rate of change and obtain the trend slope. When the slope is higher than 0.05% per minute, it is judged to have an obvious trend. The above process is used to screen out the significant, continuous and consistent direction fluctuation segments in the batch, and the segments are organized into moisture content trend characteristic information, recording their start and end time, trend direction, change amplitude and batch code.

[0058] The glue attribution identification submodule compares the time distribution of glue usage data and the corresponding batch number based on the moisture content trend characteristics, determines whether there is overlap or offset between the attribution of each set of glue usage information and the batch number, analyzes the matching of batch and glue data, identifies abnormal attribution relationships, and obtains glue attribution characteristics;

[0059] Extract the glue usage data that overlaps with its time range, compare the timestamp of each glue application record with the moisture content trend time period, calculate the time difference between the glue application data record time point and the midpoint of the trend segment, if the absolute value of the time difference is less than or equal to 2 minutes, it is considered to belong to the batch corresponding to the trend segment, and the record with a time difference of more than 2 minutes is marked as an offset record. The 2-minute threshold is set according to the actual response lag time of the glue application equipment. Through historical data statistics, it can be obtained that more than 90% of normal data are concentrated within 1 minute. Then, the glue application records in each batch are averaged. The mean is calculated and the difference is compared with the previous batch. When the change in the average glue dosage between batches exceeds 3%, it is determined that there is a mutation in the glue dosage of this batch. The 3% judgment standard is set according to the production ratio regulations. For example, the glue dosage in a batch record is 68 kg, and the previous batch is 70.5 kg, then the difference is 2.5 kg, corresponding to a ratio of about 3.5%, which exceeds the threshold and is marked as an attribution anomaly. Combined with the two conditions of time attribution and mutation of dosage between batches, all records that meet the attribution offset or mutation amplitude are marked as glue attribution abnormal data.

[0060] The parameter linkage integration submodule compares the changes of hot pressing temperature and molding speed at the same time node according to the glue attribution characteristics, determines the consistency of the change direction of each parameter, screens the linkage performance, and integrates the parameters and time nodes to obtain the parameter time series set;

[0061] The corresponding hot pressing temperature and molding speed data are extracted in sequence, and the two types of parameters are aligned on the time axis to construct parameter pairs at the same time point. For each set of data, the temperature change direction and speed change direction of the two points before and after are first determined to be consistent. If the directions are the same, they are classified as unidirectional linkage. The proportion of records with consistent linkage in the entire set of data is then counted. When this ratio is not less than 0.75, the batch is marked as a high-consistency batch. This judgment is based on the empirical rule that temperature and speed usually change positively correlated under normal production conditions. Afterwards, all time nodes with high consistency characteristics are organized into event records, and the corresponding parameter names, values, and timestamps are centralized and integrated into sequence entries. The parameters are merged into a unified parameter time series set in chronological order. If there is a record at a certain time point where the temperature and speed change simultaneously and in the same direction, it is included in the linkage event sequence, forming a trend data set of hot pressing temperature and molding speed at the same time point for subsequent multi-parameter collaborative judgment.

[0062] See also Figure 3 , the outlier identification module includes:

[0063] The moisture content fluctuation identification submodule analyzes the variation trend of wood shavings moisture content data in the target time period based on the parameter time series set, determines the direction and magnitude of the variation between consecutive time series data, identifies segments where the amplitude of the variation continuously increases or decreases, and locates the time segments where concentrated changes occur in the data sequence to obtain the moisture content mutation interval;

[0064] The time series moisture content data generated during the production process of each batch of wood shavings are processed sequentially. The moisture content value at each time node in the sequence is read in turn. The difference operation is performed on the values ​​between two adjacent time nodes, and the positive and negative signs of the difference are recorded. If the current node value minus the previous node value is greater than 0, it is marked as an upward direction. If it is less than 0, it is a downward direction. If the absolute value is less than 0.2%, it is considered a stable segment. Then, the point segments with continuous and consistent direction marks are combined into a trend segment. Each trend segment records its start time, end time, direction type and maximum change amplitude. In this process, the change amplitude of each segment is accumulated and summed. If the amplitude exceeds 1.5% continuously and lasts for more than 3 minutes, it is marked as a significant trend segment. The amplitude threshold of 1.5% is based on the normal moisture content of wood shavings. The fluctuation range is 10% to 13%, and the average offset value is taken after statistical calculation plus a safety margin. Further, continuously rising or falling segments are searched in all trend segments. If the maximum difference between them exceeds 2%, it is recorded as a strong trend segment. For example, the original moisture content data of a batch is 12.4%, 12.9%, 13.5%, 13.8%, 13.9%, 14.3%, and 14.5%. In the sequence, it rises continuously from the first value to the last value with a difference of 2.1% and a duration of 5 minutes, which meets the mutation judgment condition. Then, the segments with the above conditions in all sequences are scanned, and their start and end time nodes in the parameter time series set are recorded. Combined with their batch number and time label, the time segment with mutation in the continuous trend segment of moisture content in the batch is located, which constitutes the moisture content mutation interval segment.

[0065] The parameter attribution relationship judgment submodule compares the corresponding relationship between the synchronized data of the glue application amount and the paving density and the batch number based on the moisture content mutation interval, determines whether the paving density sampling sequence is consistent with the batch in the mutation section, and obtains the parameter batch matching status distribution;

[0066] Extract the glue application amount collection data and paving density collection data within the time period, record their timestamps, data values ​​and belonging batch numbers respectively, and for each glue application record and paving density record, find out whether its timestamp is within the mutation interval time range. If it exists, extract its corresponding batch number and compare it with the batch number of the mutation interval. If the batch number of the glue application or density data is consistent with the mutation interval, it is considered that the parameter is consistent with the mutation section. If the batch numbers are inconsistent but the time coverage overlaps, it is marked as a conflict of ownership. Then, the ownership of the glue application data and paving density data in each mutation section is counted, and the number of consistency and conflict are recorded respectively. The proportion of consistent ownership is calculated and output by classification. The sections with a ratio less than 0.8 are considered to have an unstable matching status. The attribution ratio is determined based on the boundary value set by adjusting the attribution matching rate of more than 90% of the historical data of the production system. Specifically, if the mutation section time is from 10:00 to 10:10, the corresponding batch number is B103, and the gluing data time in this time period is 10:01 and 10:04, respectively, with the corresponding batch numbers B103 and B102, then the first gluing record has the same attribution, and the second record has an inconsistent batch number and is marked as an attribution conflict. The batch attribution comparison results of all gluing data and paving density data in each mutation section are integrated to generate the batch matching status distribution information of each parameter.

[0067] The batch fluctuation difference identification submodule compares the hot pressing temperature and forming speed of the corresponding batches according to the parameter batch matching status distribution, calculates the changes in paving density and spindle speed, and uses the formula:

[0068] ;

[0069] Calculate the fluctuation level of parameters for each batch , and judge the fluctuation performance between batches to obtain the abnormal parameter distribution characteristics, among which, Indicates the The hot pressing temperature data at all times reflects the temperature status of the hot pressing link during the production process. Indicates the The molding speed data at each moment reflects the equipment operation speed in the molding process. Indicates the The paving density data at each moment reflects the density of the raw materials in the paving process of the production line. Indicates the The spindle speed data at each moment reflects the real-time operating status of the main power output components of the equipment. Indicates the The change range of moisture content within a period of time reflects the change of moisture content of raw materials during the production process. Indicates the total number of time periods involved in the calculation, that is, the length of the data time series. Indicates a minimum stabilization factor set to prevent the denominator from reaching zero.

[0070] The fluctuation level of each batch parameter refers to the range of changes or degree of fluctuation between various process parameters (such as hot pressing temperature, forming speed, paving density, spindle speed, etc.) collected from different batches during the production process. The fluctuation reflects the instability or abnormal fluctuations in the production process and helps detect potential abnormalities.

[0071] Map the assigned batch number information to each process parameter one by one. Select the data sequence length of batch B01 production time period as 5 sample points. Obtain the actual measurement value of each parameter at each time point in turn. Set the hot pressing temperature sequence within this batch as: ; Molding speed is: ; Pavement density is: ;

[0072] The spindle speed is: ;

[0073] The moisture content changes as follows: ;

[0074] Normalization is performed on each parameter to unify the scale for calculation. The Min-Max normalization method is used and the upper and lower limits of each parameter are set. The normalization range of hot pressing temperature is , the forming speed is , the pavement density is , the spindle speed is , the moisture content changes to , the normalized processing results are as follows:

[0075] ;

[0076] ;

[0077] ;

[0078] ;

[0079] ;

[0080] Substitute the above normalized results into the calculation formula, where, , stability factor value , calculate the molecular part:

[0081] Item 1: ;

[0082] Item 2: ;

[0083] Item 3: ;

[0084] Item 4: ;

[0085] Item 5: ;

[0086] The sum of the molecules is:

[0087] ;

[0088] The denominator is:

[0089] ;

[0090] The calculation results are:

[0091] ;

[0092] The results show that the multi-parameter collaborative fluctuation value of batch B01 under normalized dimension is , which is significantly lower than the benchmark fluctuation determination threshold , which means that this batch does not show a mutation trend between key process parameters, the changes between hot pressing temperature and forming speed are basically stable, the difference fluctuation between paving density and spindle speed is also very small, and the range of change of raw material moisture is within the normal deviation range. This numerical result can be directly used as the input feature quantity of batch stability in subsequent process collaborative state analysis for further processing.

[0093] See also Figure 4 , the collaborative feature determination module includes:

[0094] The direction matching submodule compares the changing direction of the paving density with the changing direction of the spindle speed feedback signal based on the distribution characteristics of abnormal parameters, determines whether the trends of the two are consistent within the same batch, optimizes the judgment criteria for trend offset, and obtains the trend direction matching coefficient;

[0095] Extract the time period and batch number of each batch that is marked as abnormal parameters, and obtain the time series data of paving density and spindle speed in the corresponding batch from the parameter time series set. During the execution process, traverse the batch numbers in turn, and judge the trend direction of the two types of parameters in the same time period according to the time alignment method. Calculate the numerical difference of paving density for every two adjacent time points. If the latter point minus the previous point is greater than zero, the segment is marked as an upward trend. If it is less than zero, it is marked as a downward trend. If the absolute difference does not exceed 0.1 kilograms per cubic meter, it is judged to be a stable trend. The trend judgment of the spindle speed data also adopts the same standard. Then, pair the two groups of trend marks in each time period. If the trend direction is the same, it is recorded as a consistent trend. If it is opposite, it is marked as inconsistent. The state is eliminated as a neutral segment that is not included in the judgment, and then the ratio of the number of time periods with consistent trends in the entire batch to the total number of judgment time periods is counted as the trend direction consistency ratio. This ratio value is the trend direction matching coefficient, and the coefficient range is set between 0 and 1. When the value is greater than or equal to 0.75, it is judged that the trend match is good. The threshold is a standard value set based on the mean value of the direction consistency ratio in the historical stable batch statistics. For example, a batch has a total of 20 judgment time periods, of which 15 are judged to have consistent trends. The matching coefficient is 0.75, and the batch is marked as having a good trend direction match. If the matching coefficient is less than 0.5, it is considered that there is a significant deviation from the trend, and the batch will be recorded as a matching anomaly. The matching coefficients and matching judgment results of each batch constitute the trend direction matching coefficient data set.

[0096] The combination relationship submodule calls the trend matching coefficient to compare the synchronous change characteristics of paving speed, roller gap, hot pressing initial temperature and plate feeding pressure, identify the key parameter combinations of synchronous change characteristics, optimize the interaction between parameters, and obtain the parameter combination change range;

[0097] Parameter call processing is performed on the batch numbers with good matching one by one, and the time series data related to the four parameters of paving speed, roller gap, initial hot pressing temperature and plate feeding pressure in the corresponding batch are extracted. These four groups of parameters are paired one by one in chronological order, and the numerical change trend of each parameter is analyzed at the same time node. The positive and negative signs of the difference between each two adjacent time nodes are judged, and the local trend direction of the parameter is identified accordingly. In each time period, if the trend direction marks of the four parameters are exactly the same, it is marked as a trend synchronization segment. If three parameters are consistent and the remaining one is stable or in a different direction, it is recorded as a partial synchronization segment. If the directions of the four parameters diverge, it is recorded as a trend separation segment. Each type of segment is counted and its time range is recorded. Then, the average change amplitude of the four types of parameters in the trend synchronization segment is calculated separately. The amplitude is the average of the absolute differences between adjacent time points. For example, in a certain batch, the initial hot pressing temperature rises from 170°C to 176℃, the change is 6℃. If this change spans 3 time periods, the average change range is 2℃. Continue to summarize the parameter change ranges of all time periods, and calculate the difference between each parameter and the average change value of its three parameters in the synchronization segment. When the difference value is less than the set deviation threshold, the parameter is determined to be a key synchronization parameter. The deviation threshold is set to twice the standard deviation of the historical fluctuation of each parameter. In particleboard production, the standard deviation of the paving speed is usually 0.15 meters per minute, so the synchronization identification deviation threshold is 0.3 meters per minute. When the average change of the paving speed of a batch in the synchronization segment is 0.8 meters per minute, and it is consistent with the changes of the other three parameters, with a difference of 0.1 meters per minute, the paving speed is marked as a key synchronization parameter. Identify the combination of key synchronization change parameters in batches with good trend matching, and output the overall change range of each type of combination in the synchronization segment as the parameter combination change range result.

[0098] The collaborative status submodule calculates the collaborative change level of the paving density change and the spindle speed offset based on the parameter combination change amplitude, compares the trend fluctuation and linkage frequency between nodes, and uses the formula:

[0099] ;

[0100] Get process synergy mutation index ,in, Represents the number of linkage nodes within the range of the combination change, Representative The change in pavement density in each node, Representative The spindle speed offset in each node, Representative The linkage structure parameters corresponding to the variation range of the parameter combination within each node, Representative The spindle control feedback delay within each node, Representative The difference between the paving feedback signal and the spindle change within each node.

[0101] The process synergy mutation index is a key indicator used to evaluate the coordinated changes between multiple process parameters in the particleboard production process. By comprehensively analyzing the changes in multiple parameters in the production process (such as paving density, spindle speed, hot pressing temperature, etc.), it is judged whether there are synchronous fluctuations, consistent trends or abnormal linkages between the parameters. It measures whether the changes in parameters are coordinated between different process nodes (such as paving, hot pressing, etc.) and whether there are abnormal mutations or abnormal fluctuations.

[0102] First extract the The change of pavement density in the linkage node , during the production process of batch number 20240518, 、 、 Pavement density values ​​were collected separately 、 、 , the density change in the two time periods is 、 In order to eliminate the dimensionality effect of different parameter units, the data is normalized using the maximum and minimum normalization method, taking the historical density range 、 , calculated as:

[0103] ;

[0104] ;

[0105] Then extract the spindle speed offset , at the same time point, the spindle speed is 、 、 , and get the change 、 , the speed reference range is 、 , normalized to:

[0106] ;

[0107] ;

[0108] Filter to meet and The node is used as a collaborative analysis node. The conditions are met, so , and then extract the linkage structure parameters corresponding to the parameter combination variation range , the standard deviation of density and speed after normalization is used to estimate the standard deviation of pavement density. , then after normalization it is The standard deviation of the speed is , normalized to ,get:

[0109] ;

[0110] Spindle control feedback delay Indicates the delay from signal transmission to actual adjustment. The node delay is , the maximum response time is , normalized to:

[0111] ;

[0112] The difference between the paving feedback signal and the spindle change It is calculated by the Euclidean distance between two normalized curves. Let the two vectors be 、 ,calculate:

[0113] ;

[0114] right , substitute: 、 、 、 、 ,have to:

[0115] ;

[0116] ;

[0117] ;

[0118] The results show that the On each node, the collaborative mutation index , showing that there is a sudden inconsistent synergistic relationship between the paving density and the spindle speed at this time point, with typical parameter linkage imbalance characteristics, which can be used for subsequent trend offset feature extraction and abnormal node positioning. The formula is based on the product term Reflects the linkage strength of parameters, square root term Measures the strength and delay of the disturbance structure, the denominator Ensure that the differences between fluctuation trends have a suppressive effect on the overall calculation results, thereby achieving stable and balanced collaborative mutation identification.

[0119] See also Figure 5 , the synchronization trend analysis module includes:

[0120] The trend calculation submodule analyzes the time series data of the glue application amount based on the process synergy mutation index, determines the change direction of the glue application amount under each batch number, compares the change rate in the time period before and after the mutation node, and calculates the trend slope interval of the glue application amount of each batch over time;

[0121] Extract the time series data of glue usage in the batch according to the batch number, read and sort the timestamp and glue usage value of each data record, and construct the change trajectory of glue usage on the time axis. Then take the glue usage difference between adjacent time points as the unit time change. If the value of the latter data point is greater than the previous data point, the segment is marked as an upward trend, otherwise it is a downward trend. If the glue usage difference between the two points is within the range of ±1kg and the glue usage of the batch fluctuates around the total amount of 70kg, it is considered to be a stable trend. Then group the data segments with continuous and consistent change directions in each batch, extract the start and end time and glue usage in each group, and calculate the change rate of each group. The rate is the total glue usage difference divided by the duration length. For example, the glue usage of a batch increased from 66kg to 74kg between 8:00 and 8:10, and the rate is 0.8k g per minute. This rate is used to compare the trend difference before and after the mutation node of the batch. Then, with the mutation node as the center, two time periods are taken forward and backward respectively, and the calculation of the glue change rate is repeated. If the rate difference before and after is greater than 1.5kg per minute, it is recorded as a trend mutation segment. The threshold value comes from the maximum load change boundary determined by the stability of the glue application control rate within the range of ±1kg per minute of the glue application equipment. Finally, the glue usage data of each batch is calculated to obtain the trend rate of each segment, and all rate values ​​are sorted within the batch. The difference between the highest rate and the lowest rate is the glue trend slope interval of the batch. For example, if the maximum rate is 1.8kg per minute and the minimum rate is 0.2kg per minute, the trend slope interval is 1.6kg per minute. All batch numbers and their corresponding slope intervals are sorted to construct a glue trend slope interval data table.

[0122] The pressure-speed comparison submodule calculates the variation of paving pressure and spindle speed in the same time period based on the trend slope interval, and compares the synchronization of pressure and speed trends using the formula:

[0123] ;

[0124] Get the trend difference amplitude ,in, Representative Pavement pressure detection data at each time node, Representative The spindle speed feedback data at each time node, is the total number of time nodes;

[0125] The trend difference amplitude refers to the difference measure between the change trends of different process parameters (such as paving pressure and spindle speed), which reflects the fluctuation synchronization of these two parameters in the production process.

[0126] The paving pressure and spindle speed are monitored in real time by sensors and the production control system respectively. If the paving pressure changes within four time nodes during the gluing process, and the spindle speed also changes at the corresponding time nodes, the change range of the paving pressure and spindle speed needs to be calculated. The paving pressure and spindle speed data from time node 1 to time node 4 are as follows:

[0127] Time node 1: paving pressure = 1000N, spindle speed = 1500RPM;

[0128] Time node 2: paving pressure = 1100N, spindle speed = 1600RPM;

[0129] Time node 3: paving pressure = 1200N, spindle speed = 1700RPM;

[0130] Time node 4: paving pressure = 1300N, spindle speed = 1800RPM;

[0131] Normalize the paving pressure and spindle speed data:

[0132] Normalization of paving pressure:

[0133] Maximum value: 1300N, minimum value: 1000N;

[0134] Normalize the pressure value at each time node:

[0135] Time node 1: ;

[0136] Time node 2: ;

[0137] Time node 3: ;

[0138] Time Node 4: ;

[0139] Normalization of spindle speed:

[0140] Maximum: 1800RPM, Minimum: 1500RPM;

[0141] Normalize the speed value at each time node:

[0142] Time node 1: ;

[0143] Time node 2: ;

[0144] Time node 3: ;

[0145] Time Node 4: ;

[0146] The normalized data is as follows:

[0147] Pavement pressure: [0, 0.3333, 0.6667, 1.0];

[0148] Spindle speed: [0, 0.3333, 0.6667, 1.0];

[0149] Substitute into the formula to calculate:

[0150] ;

[0151] ;

[0152] This result shows that the trend difference , indicating that the changing trends of paving pressure and spindle speed within the selected time node are completely consistent, which means that there is no deviation between paving pressure and spindle speed within this time period, indicating that the changes of these two process parameters within this time period are completely synchronized.

[0153] The trend consistency determination submodule calls the trend difference amplitude and analyzes the number of intersections and synchronization intervals of each parameter trend on the time axis based on the distribution direction differences under different process nodes. It identifies the consistent segments of trend fluctuations between parameters and obtains the trend offset feature value.

[0154] Extract the parameter trend marks recorded in each key process node (such as gluing, paving, and hot pressing) from the parameter time series, arrange the trend direction of each type of parameter in chronological order, and record whether it is rising, falling, or stable. Compare the trend direction sequences between each two different parameters one by one on the time axis, and judge whether their trend directions are the same at each time point. When the trend directions are the same, they are marked as consistent points. If the directions are opposite, they are marked as intersection points. If any of them is in a stable state, the point is skipped and ignored. Then count the number of trend intersections on the entire time axis. This number is used to evaluate the deviation frequency of the change pace between parameters. For example, if the glue dosage and the initial temperature of hot pressing have a trend direction intersection 12 times within an hour, then a direction change hedge occurs every 5 minutes on average, which is much higher than that of stable parameters. The average crossover benchmark value three times per hour indicates that the trend is not synchronized. Then find the time period with completely consistent trends, that is, the trend direction of all participating parameters in this time period is consistent and the duration exceeds 3 minutes. This is recorded as the trend synchronization interval. The start and end time, participating parameters and consistent trend direction of the trend synchronization interval are recorded. If the total duration of the synchronization segment accounts for more than 60% of the total duration of the entire time axis, it is determined that the parameter group has a consistent trend feature. If it is less than 30%, it is determined that the trend fluctuation is inconsistent. The interval ratio judgment threshold is set based on historical data comparison statistics, among which the proportion of 90% of high-consistency batch synchronization segments is concentrated above 60%. Finally, the number of trend crossings, the start and end time of the trend synchronization interval, the identification of the participating parameters and the proportion are extracted to constitute the trend offset feature.

[0155] See also Figure 6 ,The abnormal attribution positioning module includes:

[0156] The abnormal fluctuation attribution submodule analyzes the fluctuation relationship between the sudden change nodes of particle moisture content and the change of paving density based on the trend offset characteristic, compares the linkage between the spindle speed and other process parameters, identifies and classifies the nodes of abnormal fluctuation, and obtains the abnormal fluctuation index;

[0157] All identified trend deviation batch data are batch numbered and grouped, and the marked particle moisture content mutation node time points and corresponding moisture content change amplitudes in the batch are extracted, and the nodes are arranged in chronological order to form a mutation sequence. Then, the sampling data sequence of the paving density acquisition channel is obtained within the same time range, and the density difference between two adjacent moments is judged. If the change amplitude exceeds 0.4 kilograms per cubic meter and the direction is consistent with the moisture content change, the segment is marked as a fluctuation coupling segment. The difference judgment threshold of 0.4 is set according to the average density change range in the normal production stable batch. Then, the fluctuation coupling segments corresponding to each mutation node are summarized, and the fluctuation start time, end time and fluctuation direction are recorded, and the paving density fluctuation is recorded. The sequence is linked with the spindle speed sampling sequence for linkage judgment, and the trend of the change direction of the spindle speed in the same time segment is extracted. If the speed direction is opposite to the density direction, it is marked as anti-linkage; if the directions are consistent, it is marked as positive linkage. If the fluctuation direction is not obvious or the absolute value of the change is less than 1.5 revolutions per minute, it is recorded as a neutral segment. The threshold is set to ±2 revolutions per minute with reference to the daily fluctuation range of the spindle speed. Finally, the coupling situation of each fluctuation segment is paired with the trend mark of the spindle speed and classified into three categories: consistent coupling, anti-coupling or invalid coupling. The frequency of each type of coupling in each batch is then counted. If the frequency of the same type of coupling accounts for more than half of all mutation nodes, it is considered that there is stable abnormal linkage behavior in the batch, which constitutes an abnormal fluctuation indicator.

[0158] The batch location and integration submodule analyzes and integrates the event information and process parameters of each batch based on abnormal fluctuation indicators, determines the batch association when abnormal fluctuations occur, locates abnormal batches and process nodes, and obtains batch abnormality feature location data;

[0159] The data is then compared with the time series of the remaining parameters within the batch. All process parameters that change during the time period corresponding to the abnormal node are extracted, including glue dosage, plate feed pressure, hot pressing initial temperature, and paving speed. For each parameter, an identifier for the magnitude and direction of change within the time period is constructed and aligned along the time axis. Segments within the same time period where three or more parameters change synchronously are marked, and the corresponding batch number is extracted. If the segment has been classified as a coupling segment in the abnormal fluctuation index, it is further confirmed as a valid segment belonging to the abnormal node. The ratio of the number of valid abnormal segments to the total number of time periods in the batch is calculated. If the ratio exceeds 0.3, the batch is marked as an abnormal batch. This threshold is determined based on the historical distribution of normal fluctuation batches in production, where the number of batches with a value above 0.25 accounts for 80% of the total number of abnormal batches. The process node information of all abnormal segments within the abnormal batch is then extracted, including the corresponding equipment number, corresponding process stage (such as paving, hot pressing, glue application), and sampling parameter name. This constitutes the abnormal node identifier and forms the batch abnormality feature location data.

[0160] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An intelligent detection and diagnosis system for abnormal events in particleboard production, characterized in that: The system comprises: The parameter time series acquisition module is based on the wood shavings moisture content sampling channel, analyzes its change trend, compares the relationship between the sizing amount and the batch number, calculates the correspondence between the hot pressing temperature and the molding speed, determines the matching status of the process nodes, and integrates the time sequence of each parameter acquisition to obtain a parameter time series set; The abnormal point identification module analyzes the fluctuation of the moisture content of the wood chips based on the parameter time series set, determines the corresponding relationship between the paving density and the batch number, compares the variation range of the hot pressing temperature and the forming speed, identifies the abnormal fluctuation of the batch parameters, and obtains the distribution characteristics of the abnormal parameters; The collaborative feature determination module compares the change direction of the paving density and the spindle speed based on the abnormal parameter distribution characteristics, analyzes the inlet parameter combination of the wood shavings paving equipment and the hot pressing system, determines the collaborative state of the process parameters, and obtains the process collaborative mutation index; The collaborative feature determination module includes: The direction matching submodule compares the change direction of the paving density with the change direction of the spindle speed feedback signal based on the distribution characteristics of the abnormal parameters, determines whether the trends of the two are consistent within the same batch, optimizes the judgment criteria for trend offset, and obtains the trend direction matching coefficient; The combination relationship submodule calls the trend matching coefficient, compares the synchronous change characteristics of paving speed, roller pressing gap, hot pressing initial temperature and plate feeding pressure, identifies the key parameter combination of synchronous change characteristics, optimizes the interaction relationship between parameters, and obtains the parameter combination change range; The collaborative status submodule calculates the collaborative change level of the paving density change and the spindle speed offset based on the variation range of the parameter combination, compares the trend fluctuation and linkage frequency between nodes, and obtains the process collaborative mutation index; The synchronization trend analysis module analyzes the change trend of the glue application amount based on the process synergy mutation index, calculates the trend difference between the paving pressure and the spindle speed, compares the synchronization of process parameter fluctuations, determines the consistency of node trends, and obtains the trend offset characteristic value; The synchronization trend analysis module includes: The trend calculation submodule analyzes the time series data of the glue application amount based on the process synergy mutation index, determines the change direction of the glue application amount under each batch number, compares the change rate in the time period before and after the mutation node, and calculates the trend slope interval of the glue application amount of each batch over time; The pressure-speed comparison submodule calculates the variation range of the paving pressure and the spindle speed in the same time period according to the trend slope interval, compares the synchronicity of the pressure and speed trends, and obtains the trend difference range; The trend consistency determination submodule calls the trend difference amplitude, analyzes the number of intersections and synchronization intervals of each parameter trend on the time axis according to the distribution direction difference under the different process nodes, identifies the consistent segments of trend fluctuations between parameters, and obtains the trend offset feature.

2. The intelligent detection and diagnosis system for abnormal events in particleboard production according to claim 1 is characterized in that: The parameter time series set includes parameter type identification, time node mark, and sequence integrity information; the abnormal parameter distribution characteristics include abnormal type identification, abnormal occurrence node, and abnormal duration interval; the process collaborative mutation index includes collaborative abnormality identification, parameter linkage mode, and associated node characteristics; the trend offset feature includes trend consistency index, offset degree classification, and trend change identification.

3. The intelligent detection and diagnosis system for abnormal events in particleboard production according to claim 1 is characterized in that: The parameter timing acquisition module includes: The moisture content analysis submodule analyzes the moisture content changes of wood chips within a continuous time period based on the wood chip moisture content sampling channel. By classifying the direction changes of the collected signals at adjacent time nodes and comparing the growth or decrease in different time periods, the continuity and consistency of the moisture content changes are determined, and the fluctuation performance is screened to obtain the moisture content trend characteristics. The glue attribution identification submodule compares the time distribution of the glue usage data and the corresponding batch number based on the moisture content trend characteristics, determines whether there is overlap or offset between the attribution of each set of glue usage information and the batch number, analyzes the matching of batch and glue data, identifies abnormal attribution relationships, and obtains glue attribution characteristics; The parameter linkage integration submodule compares the changes of hot pressing temperature and molding speed at the same time node according to the glue attribution characteristics, determines the consistency of the change direction of each parameter, screens the linkage performance, and integrates the parameters and time nodes to obtain a parameter time series set.

4. The intelligent detection and diagnosis system for abnormal events in particleboard production according to claim 1 is characterized in that: The outlier identification module includes: The moisture content fluctuation identification submodule analyzes the variation trend of the wood shavings moisture content data in the target time period based on the parameter time series set, determines the variation direction and amplitude between continuous time series data, identifies segments where the amplitude of variation continuously rises or falls, and locates the time segments where concentrated variation occurs in the data sequence to obtain the moisture content mutation interval segment; The parameter attribution relationship judgment submodule compares the corresponding relationship between the synchronized data of the glue application amount and the paving density and the batch number based on the moisture content mutation interval, determines whether the paving density sampling sequence is consistent with the batch of the mutation section, and obtains the parameter batch matching status distribution; The batch fluctuation difference identification submodule compares the hot pressing temperature and forming speed of the corresponding batches according to the batch matching status distribution of the parameters, calculates the changes in the paving density and the spindle speed, calculates the fluctuation level of the parameters of each batch, and judges the fluctuation performance between batches to obtain the abnormal parameter distribution characteristics.

5. The intelligent detection and diagnosis system for abnormal events in particleboard production according to claim 1 is characterized in that: The system further comprises: The abnormal attribution and positioning module determines the abnormal attribution of the sudden change node of wood shavings moisture content, paving density and spindle speed change based on the trend offset feature, identifies the batch and event information of the abnormal node, integrates the abnormal correlation information of the process parameters, and obtains the batch abnormality feature positioning information; The batch abnormality feature location information includes a batch location identifier, an abnormal event type, and a node affiliation identifier.

6. The intelligent detection and diagnosis system for abnormal events in particleboard production according to claim 5 is characterized in that: The abnormal attribution positioning module includes: The abnormal fluctuation attribution submodule analyzes the fluctuation relationship between the sudden change node of the particle moisture content and the change of the paving density based on the trend deviation characteristic, compares the linkage between the spindle speed and other process parameters, identifies and classifies the nodes of abnormal fluctuation, and obtains the abnormal fluctuation index; The batch positioning and integration submodule analyzes and integrates the event information and process parameters of each batch based on the abnormal fluctuation index, determines the batch association when the abnormal fluctuation occurs, locates the abnormal batch and process node, and obtains the batch abnormality feature positioning data.

Images