A self-diagnosis method for abnormal working conditions of a feed processing process

Abstract

The application relates to the technical field of industrial production process control, in particular to a self-diagnosis method for abnormal working conditions in a feed processing process. The method comprises the following steps: acquiring an acoustic wave signal sequence of a crushing device of a feed processing line, determining a high-frequency energy value and a low-frequency energy value by using the acoustic wave signal sequence; determining a conversion ratio sequence by using the ratio of the high-frequency energy value and the low-frequency energy value; acquiring an ultrasonic amplitude value sequence of the crushing device, determining a change rate sequence of the attenuation degree value of the ultrasonic amplitude value sequence; determining an evolution slope of the conversion ratio sequence and an attenuation numerical value increment of the change rate sequence, determining a continuous time consumption feature that the evolution slope and the attenuation numerical value increment keep corresponding trends in the same time window; and determining an abnormal self-diagnosis result of the crushing device by using the combination of the evolution slope, the attenuation numerical value increment and the continuous time consumption feature. Through the technical scheme, the abnormality of the feed crushing device can be found in time, and the safety of feed production is ensured.

Description

Technical Field

[0001] This application relates to the field of industrial production process control technology, and in particular to a self-diagnosis method for abnormal operating conditions in a feed processing process. Background Technology

[0002] Bio-feed, relying on the advantage of improving nutrient utilization through microbial fermentation and modification, has become a mainstream product type in the field of large-scale livestock and poultry farming. In the continuous processing production line of bio-feed, the crushing equipment is the key pre-processing unit at the front end, which undertakes the task of crushing and screening bulk raw materials such as corn, soybean meal and bran. The uniformity of the particle size of the crushed raw materials directly determines the reaction stability of subsequent processes such as batching and mixing, closed fermentation and conditioning and granulation. It is the core link in controlling the live bacteria content, digestibility and absorption rate and the qualification rate of the finished bio-feed product.

[0003] Currently, most crushing equipment in the industry is of high-speed hammer or toothed disc structure. It is subject to the long-term impact of material impact, hard object collision, feed flow fluctuation, liner wear and hammer damage, etc. The operating conditions are complex and variable, and the routine operation and maintenance management is difficult. At present, the traditional mode of monitoring the operating status of feed crushing equipment is generally adopted, which is to use a single vibration sensor for fixed-point collection, current threshold over-limit alarm or manual periodic disassembly and inspection.

[0004] The processing of biological feed involves complex changes in form, such as the crushing of wet materials and the conveying of highly viscous dust. Biological feed itself is rich in free water and fermentation metabolites. The equipment is exposed to a harsh environment with high concentrations of adhesive dust and strong background noise for a long time. Using a single-dimensional mechanical vibration monitoring method has a low accuracy rate in identifying hidden mechanical faults such as early minor wear and screen breakage. It is difficult to detect abnormalities in the feed crushing equipment in a timely manner, making it difficult to ensure the safety and stability of the feed production process. Summary of the Invention

[0005] To promptly detect abnormalities in feed grinding equipment and ensure the safety and stability of the feed production process, this application provides a self-diagnostic method for abnormal operating conditions in the feed processing process. The method includes: acquiring the acoustic signal sequence of the grinding equipment in the feed processing line; converting the acoustic signal sequence to the frequency domain to generate an acoustic spectrum; determining the high-frequency energy value corresponding to mechanical collision and the low-frequency energy value corresponding to material friction in the acoustic spectrum; using the ratio of the high-frequency energy value to the low-frequency energy value as the energy conversion ratio; and splicing the energy conversion ratios according to the sequence acquisition order to obtain a conversion ratio sequence; acquiring the ultrasonic amplitude sequence of the grinding equipment; and analyzing the ultrasonic amplitude... The amplitude value sequence is denoised to determine the attenuation value, and the attenuation change rate of the attenuation value is determined. The attenuation change rate is spliced ​​according to the sequence acquisition order to obtain the change rate sequence. The conversion ratio sequence and the change rate sequence are aligned in the timestamp dimension to determine the evolution slope of the aligned conversion ratio sequence and the attenuation value increment of the aligned change rate sequence. The duration characteristic of the evolution slope and the attenuation value increment maintaining the corresponding trend within the same time window is determined. The evolution slope, attenuation value increment and duration characteristic are combined to construct a multidimensional spatiotemporal state quantity. The multidimensional spatiotemporal state quantity is used to determine the abnormal self-diagnosis result of the crushing equipment.

[0006] In this way, by separately collecting acoustic signal sequences and ultrasonic amplitude sequences that characterize mechanical actions and constructing multidimensional spatiotemporal state quantities, it is helpful to improve the accuracy of self-diagnosis of hidden faults.

[0007] Optionally, acquiring the acoustic signal sequence of the crushing equipment in the feed processing line, and acquiring the ultrasonic amplitude sequence of the crushing equipment, includes: continuously acquiring the vibration waves excited by the mechanical collision of the inner cavity through an acoustic receiver attached to the surface of the crushing equipment shell to obtain the acoustic signal sequence; continuously emitting ultrasonic waves toward the powder falling channel through ultrasonic probes arranged on both sides of the discharge chute at the bottom of the crushing equipment, and receiving the residual ultrasonic waves after they penetrate laterally into the powder falling channel; and performing energy attenuation conversion processing on the residual ultrasonic waves to obtain the ultrasonic amplitude sequence.

[0008] Optionally, the high-frequency energy value corresponding to the mechanical collision and the low-frequency energy value corresponding to the material friction are determined in the acoustic spectrum diagram, including: determining the hard impact segment as the high-frequency interval in the acoustic spectrum diagram, and performing amplitude integration calculation on the energy distribution within the high-frequency interval to obtain the high-frequency energy value; determining the flexible friction segment as the low-frequency interval in the acoustic spectrum diagram, and performing amplitude integration calculation on the energy distribution within the low-frequency interval to obtain the low-frequency energy value.

[0009] In this way, by distinguishing between the hard impact section and the flexible friction section, and performing amplitude integration calculations separately to obtain different energy characterization values, we can avoid background noise interference caused by friction of biological materials and reduce errors in the signal analysis process.

[0010] Optionally, the ultrasonic amplitude sequence is denoised to determine the attenuation value, including: segmenting the ultrasonic amplitude sequence into multiple time-domain data segments according to a predetermined step size; performing an arithmetic mean calculation operation on the ultrasonic amplitude points within each time-domain data segment to obtain the average attenuation amplitude corresponding to the time-domain data segment, and using the average attenuation amplitude as the attenuation value of the corresponding time node.

[0011] Optionally, determining the duration of the corresponding trend of the evolution slope and the decay increment within the same time window includes: obtaining the absolute timestamp shared by the aligned conversion ratio sequence and the rate of change sequence; using the absolute timestamp as a matrix mapping index; performing planar pairing and assembly of the evolution slope and decay increment corresponding to the same absolute timestamp to obtain a two-dimensional coupling matrix; using the two-dimensional coupling matrix to track the change trajectory of the evolution slope and the decay increment on the time axis; and determining the duration of the corresponding trend of the evolution slope and the decay increment within the same time window based on the change trajectory.

[0012] In this way, by establishing a two-dimensional coupling matrix to pair and assemble the evolution slopes and attenuation increments of two independent dimensions in a plane, the temporal correlation between acoustic features and biological material falling features can be intuitively mapped, which helps to improve the coherence of anomaly feature tracking.

[0013] Optionally, the evolution slope, attenuation increment, and duration feature are combined to construct a multidimensional spatiotemporal state variable, including: when the evolution slope is within a preset gentle slope range and the attenuation increment is within a preset slow increment range, the duration feature is marked as a continuous duration value; the absolute value of the conversion ratio of the crushing equipment at the current monitoring node is determined; and the continuous duration value, evolution slope, attenuation increment, and absolute value of the conversion ratio are parameterized and concatenated to construct a first state variable as a multidimensional spatiotemporal state variable.

[0014] Optionally, the abnormal self-diagnosis result of the crushing equipment can be determined by using multi-dimensional spatiotemporal state variables, including: determining the continuous duration value and the absolute value of the conversion ratio recorded in the first state variable; when it is confirmed that the continuous duration value is greater than the cycle duration in the preset degradation rule conditions and the absolute value of the conversion ratio is greater than the preset efficiency critical point, the matching and hitting degradation rule conditions are determined, and the wear exceeding the limit result is output as the abnormal self-diagnosis result of the crushing equipment.

[0015] In this way, by combining the first state quantity with specific long-term continuous duration values ​​and the absolute value of the conversion ratio for analysis, false alarms caused by relying on a single instantaneous data are avoided, and the accuracy of identifying progressive degradation wear is improved when the equipment processes corrosive fermented biomass.

[0016] Optionally, the evolution slope, decay increment, and duration feature are combined to construct a multidimensional spatiotemporal state variable, including: when the evolution slope exceeds a preset abrupt change slope threshold and the decay increment simultaneously exceeds a preset step increment threshold within a preset evaluation time window, determining the over-limit start point and over-limit end point of the over-limit jump; calculating the jump duration value by subtracting the over-limit start point from the over-limit end point, and marking the duration feature as the jump duration value; and concatenating the jump duration value, the evolution slope exceeding the preset abrupt change slope threshold, and the decay increment exceeding the preset step increment threshold to construct a second state variable as a multidimensional spatiotemporal state variable.

[0017] Optionally, the abnormal self-diagnosis result of the crushing equipment is determined by using multidimensional spatiotemporal state variables, including: determining the jump time value recorded inside the second state variable; if the jump time value is less than the mutation time in the preset degradation rule conditions, determining the matching mutation rule conditions, and outputting the screen breakage result as the abnormal self-diagnosis result of the crushing equipment.

[0018] Optionally, the method further includes: applying an amplitude demodulation algorithm to the acoustic signal sequence to determine the low-frequency envelope excited by the internal drive rotor of the crushing equipment in a single rotation cycle, and statistically analyzing the envelope repetition frequency of the low-frequency envelope within a preset evaluation time; comparing the envelope repetition frequency with the current motor speed frequency of the crushing equipment; when it is confirmed that the envelope repetition frequency matches the motor speed frequency, determining the peak-valley amplitude difference of the low-frequency envelope; and when it is confirmed that the peak-valley amplitude difference shows an expanding trend within a consecutive preset number of cycles, outputting the rotor imbalance result as the abnormal self-diagnosis result of the crushing equipment.

[0019] Optionally, the method further includes: acquiring a thermal imaging image of the exterior of the top feed channel of the crushing equipment, determining the spatial geometric contour of the continuous high-temperature zone induced by the friction between the falling material and the inner wall of the channel from the thermal imaging image; when it is confirmed that the spatial geometric contour shrinks in the preset lower half of the feed channel and generates high-heat agglomerates within a specific height of the feed channel, the agglomeration blockage abnormality is output as the abnormal self-diagnosis result of the crushing equipment.

[0020] The technical solutions provided by the embodiments of this application may include the following beneficial effects: by acquiring the acoustic signal sequence and ultrasonic amplitude sequence of the crushing equipment in the feed processing line, the conversion ratio sequence and the rate of change sequence can be determined respectively. Based on the two different characteristics of acoustic wave characteristics and amplitude characteristics, the continuous time consumption characteristics of maintaining the corresponding direction within the same time window can more accurately realize the self-diagnosis of abnormal operating conditions of the crushing equipment, and ensure the safety and stability of the feed production process.

[0021] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating a self-diagnosis method for abnormal operating conditions in a feed processing process according to an exemplary embodiment; Figure 2 This is a waveform diagram showing the acoustic signal sequence and low-frequency envelope characteristics of the crushing equipment. Detailed Implementation

[0023] To promptly detect abnormalities in feed grinding equipment and ensure the safety and stability of the feed production process, this application provides a self-diagnosis method for abnormal operating conditions in the feed processing process. Figure 1 This is a flowchart illustrating a self-diagnosis method for abnormal operating conditions in a feed processing process, according to an exemplary embodiment. Figure 1 As shown, the method includes the following steps.

[0024] In step S101, the high-frequency energy value and low-frequency energy value of the crushing equipment are determined.

[0025] Acquire the acoustic signal sequence of the crushing equipment in the feed processing line, convert the acoustic signal sequence to the frequency domain to generate an acoustic spectrum, and determine the high-frequency energy value of the corresponding mechanical collision and the low-frequency energy value of the corresponding material friction in the acoustic spectrum.

[0026] The acoustic signal sequence is used to reflect the vibration characteristics of the internal mechanical components of the crushing equipment. As the core cutting device of the biological feed processing line, the crushing equipment mainly relies on the high-speed rotating metal hammers inside to continuously impact the agricultural raw materials and fermentation substrates entering the cavity.

[0027] Probiotic communities and fermented nutrient solutions may be added to biological feed. Compared with traditional dry feed, biological feed has a higher moisture content and more complex rheological properties. Inside the crushing equipment under rated load, frequent and dense solid collisions will occur, and the high-speed rotating metal components will continuously impact the raw grain particles head-on.

[0028] Moist biological materials form a soft, cushioning layer with strong sound absorption in the crushing chamber. When the original vibration waves generated by mechanical collisions penetrate this cushioning layer, acoustic attenuation and phase shift may occur.

[0029] The acoustic signal sequence of the crushing equipment in the feed processing line can record the changes in the pressure wave amplitude of the mechanical action inside the crushing equipment. The acoustic signal sequence records the original waveform characteristics that reflect the smoothness of rotor operation and the stress crushing state of biological materials.

[0030] The acquisition process of acoustic signal sequences can be recorded by high-frequency discretization following the Nyquist sampling law, so that the continuous mechanical oscillation physical process can be transformed into a digital measurement sequence that can be recognized and analyzed by the processor array computing unit.

[0031] Converting a sound wave signal sequence to the frequency domain to generate an acoustic spectrum map can reconstruct and map the original one-dimensional time-scale vibration waveform to a range containing multiple different oscillation frequency distributions. The superposition and mixing of various frequency components in the time-domain waveform makes it impossible to directly observe microscopic changes. By applying algorithms such as the Fast Fourier Transform, the time-domain data points can be deconstructed into an acoustic spectrum map with a frequency domain distribution composed of different sinusoidal fundamental waves.

[0032] Acoustic spectrograms can intuitively show the intensity of energy accumulation and distribution caused by different actions in different Hertz frequency bands, making it easy to separate and analyze complex sound sources that are mixed and intertwined.

[0033] By determining the high-frequency energy value corresponding to mechanical collision and the low-frequency energy value corresponding to material friction in the acoustic spectrum, the high-frequency energy value is associated with the short, harsh sound wave frequency band excited by the high-speed impact of a high-hardness metal hammer on the screen structure or hard grain impurities; the low-frequency energy value is associated with the dull, low sound wave frequency band generated by the tangential sliding friction of soft and somewhat viscous fermented powder biological materials on the screen surface and by being pushed, rolled and squeezed by strong wind pressure.

[0034] The viscous resistance of biological feed is usually greater than that of conventional dried raw materials, and the envelope of low-frequency friction sound patterns usually exhibits a broader base distribution. The sound frequency bands corresponding to mechanical collisions and material friction can be determined by testing the audio near the crushing equipment of the feed processing line in advance. For example, audio acquisition and statistical analysis of audio frequency bands can be performed for two states: no-load operation and loading with biological feed, to determine the different sound frequency bands that can distinguish between mechanical collisions and material friction. The embodiments of this application will not be described in detail here.

[0035] In one embodiment, acquiring the acoustic signal sequence of the crushing equipment in a feed processing line and acquiring the ultrasonic amplitude sequence of the crushing equipment includes: continuously acquiring the vibration waves excited by mechanical collisions in the inner cavity through an acoustic receiver attached to the surface of the crushing equipment shell to obtain the acoustic signal sequence; continuously emitting ultrasonic waves toward the falling channel of the powder through ultrasonic probes arranged on both sides of the discharge chute at the bottom of the crushing equipment, and receiving the residual ultrasonic waves after they penetrate laterally through the falling channel of the powder; and performing energy attenuation conversion processing on the residual ultrasonic waves to obtain the ultrasonic amplitude sequence.

[0036] In the biological feed processing workshop, due to the spread of fermentation gases and the pervasive high-humidity dust, the absorption effect of the external air medium on the high-frequency sound wave components can be continuously collected by an acoustic receiver attached to the surface of the crushing equipment shell. This creates a solid-state acoustic coupling channel between the sensing element and the metal shell, reducing or weakening the absorption of high-frequency vibration components by the external turbid air medium.

[0037] The acoustic receiver is fixed to the surface of the outer shell and can periodically acquire the transient impact of the inner cavity and the wave signal transmitted to the outer shell of the equipment, and present it as a sequence of sound wave signals that can be recognized by the digital processor. When the bio-fermentation powder passes through the discharge chute under the action of gravity, it is prone to slight agglomeration due to the moisture bridging effect. By using ultrasonic probes arranged on both sides of the discharge chute at the bottom of the crushing equipment, high-frequency directional ultrasonic beams are continuously emitted towards the powder falling channel, so that the ultrasonic beams can completely cover the cross-sectional area of ​​the wet powder falling freely in the horizontal space axis.

[0038] When ultrasound encounters suspended biological fermentation particles, it may undergo scattering and viscous absorption effects. Due to the high concentration of suspended wet dust in the falling channel, the ultrasound beam has an absorption and scattering effect. The residual ultrasound waves received after the powder falls laterally through the falling channel may carry spatial attenuation information about the material concentration and density.

[0039] To reduce the attenuation effect on ultrasound, the residual ultrasound can be processed by energy attenuation conversion. The initial driving power of the ultrasound transmitter is compared and mapped with the actual residual power captured by the receiver to obtain an ultrasound amplitude sequence that reflects the fluctuation of the degree of powder obstruction over time.

[0040] The obtained ultrasonic amplitude sequence, using the mechanical wave medium attenuation principle, enables continuous detection of the actual amount of biological material passing through the sealed discharge trough, which facilitates a more accurate diagnosis of the working conditions of the feed processing process.

[0041] In one embodiment, determining the high-frequency energy value of the corresponding mechanical collision and the low-frequency energy value of the corresponding material friction in the acoustic spectrum includes: determining the hard impact segment as the high-frequency range in the acoustic spectrum, and performing amplitude integration calculation on the energy distribution within the high-frequency range to obtain the high-frequency energy value; determining the flexible friction segment as the low-frequency range in the acoustic spectrum, and performing amplitude integration calculation on the energy distribution within the low-frequency range to obtain the low-frequency energy value.

[0042] The hard impact range can be pre-defined in the frequency band between 12,000 Hz and 18,000 Hz. The hard impact range is determined as the high-frequency range in the acoustic spectrum diagram. The high-frequency range accurately covers the crisp impact characteristic frequency band generated when the hammer breaks or the metal structure encounters a hard foreign object.

[0043] Biological materials typically contain some incompletely softened hard fibrous raw materials. The high-frequency sound waves emitted by the splitting of these hard fibers by the hammer blades are also covered within this calibrated high-frequency range.

[0044] The energy distribution within the high-frequency range is continuously integrated along the frequency coordinate axis. The high-frequency integral formula for calculating the high-frequency energy value is expressed as follows: , Indicates high-frequency energy value, This indicates the starting frequency value of the high-frequency range. This indicates the numerical value of the ending frequency in the high-frequency range. This represents the power spectral density function at the corresponding frequency in the acoustic spectrum.

[0045] By fully integrating and accumulating the power spectral density function in the high-frequency range, the high-frequency energy value can represent the total mechanical energy released by all abnormal rigid collisions that occur in the crushing equipment within a specific time window.

[0046] During the kneading and crushing process, fermented biological materials mainly exhibit viscous frictional work. The flexible friction section can be calibrated in the frequency band range of 2000 Hz to 5000 Hz. In the acoustic spectrum diagram, the flexible friction section is determined as the low frequency range. The low frequency range is the characteristic frequency band of a large amount of wet powdery material that meets the qualified particle size standard and rubs against the edge of the mesh under the suction of the fan.

[0047] The energy distribution within the low-frequency range is continuously integrated along the frequency coordinate axis. The low-frequency integral formula for calculating the low-frequency energy value is expressed as follows: , Indicates low-frequency energy value. This indicates the starting frequency value of the low-frequency range. This indicates the numerical value of the termination frequency in the low-frequency range.

[0048] The low-frequency energy value is obtained by solving the low-frequency integral calculation formula. The low-frequency energy value can characterize the normal operation energy consumed in the current crushing stage due to the effective flow friction generated by the biological powder and the overcoming of the internal viscosity of the material itself.

[0049] In step S102, the conversion ratio sequence is determined based on the high-frequency energy value and the low-frequency energy value.

[0050] The ratio of high-frequency energy value to low-frequency energy value is used as the energy conversion ratio. The energy conversion ratios are spliced ​​together according to the sequence acquisition order to obtain the conversion ratio sequence.

[0051] In its healthy operating state, the pulverizing equipment primarily converts the mechanical power input from the motor into the frictional and kneading energy required to effectively pulverize biological materials containing a certain moisture content. The ratio of high-frequency energy value to low-frequency energy value is used as the energy conversion ratio, which is defined as follows: , Indicates the energy conversion ratio. Indicates high-frequency energy value, Representing low-frequency energy value, the energy conversion ratio characterizes the actual energy distribution ratio between the equipment's input power in ineffective mechanical hard impact and effective biological material crushing.

[0052] When the low-frequency energy value is lower than the preset noise limit, it can be determined that the crushing equipment is in an abnormal working condition of no-load or stall, and the energy conversion ratio at the corresponding moment is marked as invalid data and excluded from the subsequent processing of the conversion ratio sequence. The preset noise limit can be determined based on the statistical average value of low-frequency energy of multiple consecutive production cycles when the crushing equipment is running under no-load conditions.

[0053] During the processing of bio-feed, free organic acids and moisture may cause slight and continuous chemical corrosion on the surface of the metal hammer blades, accelerating the passivation and wear rate of the hammer blade striking edges. When the hammer blades inside the crushing equipment wear excessively, resulting in a decrease in cutting ability, the bio-fermented grain particles cannot be quickly sheared and crushed, but instead continue to be subjected to ineffective and rigid pushing and impact in the inner cavity. At this time, the high-frequency energy value corresponding to useless power consumption will increase significantly, while the low-frequency energy value corresponding to effective discharge friction remains unchanged, or even the low-frequency energy value of discharge friction may decrease due to the slowdown of discharge, resulting in an increase in the energy conversion ratio.

[0054] The timestamp of each data packet can be extracted, and the energy conversion ratios calculated at each time point can be smoothly spliced ​​and connected according to the continuous sequence of acquisition triggered by the timestamps. This allows the scattered and isolated ratio values ​​representing transient acoustic characteristics at isolated time points to be integrated into a conversion ratio sequence that continuously reflects the evolution trend of the working efficiency of the crushing equipment.

[0055] The energy conversion ratio is defined as the ratio of high-frequency energy value to low-frequency energy value. Considering that the low-frequency energy value may approach zero when the crushing equipment is running under no-load or in extreme working conditions such as stall, the energy conversion ratio may have a meaningless maximum value, which in turn interferes with the calculation of the subsequent evolution slope. Therefore, a preset noise floor value can be set as a judgment threshold.

[0056] The preset noise floor value can be determined in the following way: control the crushing equipment to run continuously for multiple complete production cycles under no-load conditions, collect the low-frequency energy value in each production cycle, calculate the arithmetic mean of all collected values, and use the arithmetic mean as the preset noise floor value.

[0057] During real-time monitoring, when the low-frequency energy value at a certain moment is detected to be lower than or equal to the preset noise threshold, the calculated energy conversion ratio can be marked as invalid data and removed when constructing the conversion ratio sequence, thereby avoiding interference from extreme operating conditions on subsequent evolution trend analysis.

[0058] In step S103, the rate of change sequence of the ultrasonic amplitude sequence of the crushing equipment is determined.

[0059] The ultrasonic amplitude sequence of the crushing equipment is acquired, the ultrasonic amplitude sequence is denoised to determine the attenuation value, the attenuation change rate of the attenuation value is determined, and the attenuation change rate is spliced ​​according to the sequence acquisition order to obtain the change rate sequence.

[0060] In one embodiment, denoising the ultrasonic amplitude sequence to determine the attenuation value includes: segmenting the ultrasonic amplitude sequence into multiple time-domain data segments according to a predetermined step size; performing an arithmetic mean calculation operation on the ultrasonic amplitude points within each time-domain data segment to obtain the average attenuation amplitude corresponding to the time-domain data segment; and using the average attenuation amplitude as the attenuation value of the corresponding time node.

[0061] Ultrasonic amplitude sequences are used to characterize the attenuation characteristics of sound wave energy passing through the pulverized material flow, the process of biological feed falling and discharging, mixed with tiny water droplet agglomerates containing moisture and large particles of semi-finished pulverized material.

[0062] To eliminate short-term, severe spikes in the ultrasonic signal caused by transient turbulent vortices in the high-pressure airflow inside the discharge trough, or by a few high-viscosity large particles accidentally flying over and blocking the ultrasonic probe, the ultrasonic amplitude sequence can be windowed and truncated at a predetermined step size of, for example, 300 milliseconds, thereby cutting the original data sequence into multiple time-domain data segments that are correlated with each other in time.

[0063] The segmented processing mechanism enables efficient local batch processing operations within the boundaries of limited memory resources, avoiding the serious memory overflow risk caused by global sequence scanning.

[0064] For each set of time-domain data segments containing hundreds of densely sampled ultrasonic amplitude points, the arithmetic mean calculation operation is performed by summing and dividing by the total number of samples. This avoids short-term high-frequency random noise interference amplitude and obtains the average attenuation amplitude corresponding to the time-domain data segment that can stably represent the overall biological material concentration level within the time slice.

[0065] The average attenuation amplitude, which avoids random disturbances and burr interference, is used as the attenuation value at the corresponding time node, so that the attenuation value can more stably reflect the actual density evolution state of the biomaterial inside the powder falling channel.

[0066] Determine the rate of change of the attenuation value with respect to the time coordinate system. The formula for calculating the rate of change of attenuation is as follows: , Indicates the rate of decay. This represents the difference in attenuation values ​​between two adjacent time points. It represents the time interval between two adjacent time points.

[0067] For bio-powders with sticky wall-hanging properties, the uniformity of discharge speed is a direct indicator of whether material blockage occurs in the entire equipment. The attenuation change rate, by calculating the difference between adjacent amplitudes, can reflect the dynamic acceleration or lag deceleration trend of the material flow at the discharge port.

[0068] The extracted discrete slope parameters are smoothly spliced ​​along the time axis according to the sensor sampling time scale, i.e. the sequence acquisition order, to obtain a rate of change sequence representing the unobstructed outflow state of moist biological materials.

[0069] Energy attenuation conversion can be based on the attenuation characteristics of ultrasound in wet powder suspension media. During the process of ultrasound penetrating the powder falling channel, there is an attenuation relationship between the initial driving amplitude of the transmitting end and the actual remaining amplitude of the receiving end. The attenuation relationship is affected by the material concentration, ambient temperature and ambient humidity.

[0070] Specifically, the higher the material concentration, the more suspended particles the ultrasound encounters during penetration, resulting in stronger scattering and absorption effects and a lower residual amplitude at the receiving end; the higher the ambient temperature, the more intense the thermal motion of the medium molecules, causing the ultrasound energy to dissipate faster; and the higher the ambient humidity, the more significant the absorption effect of water molecules on ultrasound.

[0071] To establish a reliable attenuation conversion relationship, ultrasonic penetration experiments can be conducted in advance in the workshop environment where the crushing equipment is located, using standard biological feed samples with different concentration gradients. The emission amplitude and reception amplitude of each concentration sample under different temperature and humidity conditions can be recorded to establish a control data table with material concentration as the main variable and ambient temperature and humidity as compensation variables. During online monitoring, the corresponding compensation parameters are selected from the control data table based on the real-time measured ambient temperature and humidity, and the remaining amplitude at the receiving end is converted into an ultrasonic amplitude sequence reflecting the material concentration level.

[0072] In step S104, the duration characteristics of the conversion ratio sequence and the rate of change sequence are determined.

[0073] Align the conversion ratio sequence and the rate of change sequence along the timestamp dimension, determine the evolution slope of the aligned conversion ratio sequence and the decay increment of the aligned rate of change sequence, and determine the sustained time consumption characteristics of the evolution slope and decay increment maintaining the corresponding trend within the same time window.

[0074] The data throughput rate and buffer delay of the internal rigid collision ratio acoustic data collected by the acoustic receiver and the external material flow state space data collected by the ultrasonic probe may differ. Aligning the conversion ratio sequence and the rate of change sequence with the reference position in the timestamp dimension can avoid the timing phase disorder that may be caused by asynchronous sampling of multiple sensors.

[0075] The evolution slope of the aligned conversion ratio sequence is determined. The evolution slope is fitted by linear regression to avoid high-frequency noise. The obtained evolution slope can reflect the changes of the grinding equipment in the feed processing process within a macroscopic observation period of, for example, 10 minutes, so as to determine whether the grinding equipment has a relatively slow deterioration or an instantaneous deterioration due to component failure.

[0076] The evolution slope is obtained by performing linear regression fitting on the conversion ratio sequence within a preset evaluation time window. Specifically, the timestamps within the preset evaluation time window can be used as the horizontal axis, and the energy conversion ratio at the corresponding time can be used as the vertical axis to plot multiple discrete data points in a two-dimensional coordinate system. Straight lines are fitted to these discrete data points to minimize the sum of the squares of the vertical distances from the fitted straight line to each data point. The slope of the fitted straight line is defined as the evolution slope.

[0077] The duration of the preset evaluation time window is pre-calibrated based on the rotor speed and material throughput characteristics of the crushing equipment. For example, it can be set to cover the time span of at least 300 complete rotor rotation cycles of the crushing equipment, or set to the order of tens of seconds to several minutes, such as 10 minutes, to ensure that the fitting results can reflect the macroscopic operating trend of the crushing equipment rather than instantaneous noise fluctuations.

[0078] The decay increment of the aligned rate of change sequence is determined. The decay increment is calculated by integral accumulation. The decay increment reflects the actual cumulative change in the density of high humidity dust inside the discharge channel within the observation time segment.

[0079] The attenuation increment is obtained by accumulating the rate of change sequence segment by segment within a preset evaluation time window. Specifically, the average rate of change between two adjacent time nodes within the preset evaluation time window can be taken, and then the average value can be multiplied by the corresponding time interval to obtain a single-segment accumulated value.

[0080] The same operation is repeated for all adjacent time nodes within the preset evaluation time window in chronological order to obtain multiple single-segment cumulative values. The sum of all single-segment cumulative values ​​is used to obtain the attenuation value increment. The segment-by-segment accumulation adopts the method of averaging adjacent change rates to smooth the high-frequency noise fluctuations in the change rate sequence, so that the attenuation value increment can stably reflect the overall cumulative change in the concentration of material inside the discharge channel within the preset evaluation time window.

[0081] The continuous time consumption characteristics of the evolution slope and the attenuation value increment being determined to maintain the same direction of increase or the opposite direction of decrease within the same time window are identified. The continuous time consumption characteristics can be determined by setting a time threshold to eliminate the interference of short-term operating condition fluctuations caused by the instantaneous switching of different biomaterial formulations, and to identify the long-term degradation process of actual mechanical failure.

[0082] In one embodiment, determining the sustained time-consuming characteristics of the evolution slope and the decay numerical increment maintaining the corresponding direction within the same time window includes: obtaining the absolute timestamp shared by the aligned conversion ratio sequence and the rate of change sequence; using the absolute timestamp as a matrix mapping index; performing planar pairing and assembly of the evolution slope and decay numerical increment corresponding to the same absolute timestamp to obtain a two-dimensional coupling matrix; using the two-dimensional coupling matrix to track the change trajectory of the evolution slope and the decay numerical increment on the time axis; and determining the sustained time-consuming characteristics of the evolution slope and the decay numerical increment maintaining the corresponding direction within the same time window based on the change trajectory.

[0083] The evolution slope of the acoustic work characteristics of the internal mechanical collision corresponding to the same absolute timestamp is paired and assembled with the attenuation increment of the value reflecting the discharge characteristics of the external moist biological material flow based on time, to obtain a two-dimensional coupling matrix containing the synergistic interaction of multiple influencing factors.

[0084] The two-dimensional coupling matrix is ​​constructed as follows: obtain all absolute timestamps shared by the aligned transformation ratio sequence and the rate of change sequence, and arrange them in chronological order to form a row index sequence; for each absolute timestamp, calculate the corresponding evolution slope value and decay value; take the evolution slope value and decay value as two parallel column elements, and combine them with the corresponding absolute timestamp row index to form each row record of the two-dimensional coupling matrix.

[0085] Each row in the two-dimensional coupling matrix contains three data items: the absolute timestamp, the evolution slope corresponding to the absolute timestamp, and the attenuation increment corresponding to the absolute timestamp. The timestamp interval between adjacent rows is determined by the sensor sampling period and is kept uniformly distributed.

[0086] The evolution slope and attenuation increment are tracked along the time axis using a two-dimensional coupling matrix. This includes: starting from the first row of the two-dimensional coupling matrix, reading the evolution slope and attenuation increment values ​​row by row, and observing their increasing or decreasing trends over time; when the evolution slope and attenuation increment values ​​show the same direction of change at multiple consecutive time points, i.e., they increase or decrease simultaneously, the start and end timestamps of the same direction of change are recorded, and the time span between the end and start timestamps is used as a continuous time consumption feature; the determination of the same direction of change allows for setting a small fluctuation tolerance to eliminate false reverse fluctuations caused by sensor sampling noise.

[0087] By using a two-dimensional coupling matrix to synchronously track the correlation trajectory of the evolution slope and the attenuation value increment on the time axis, it can be determined whether, for example, when the acoustic crushing efficiency begins to slowly decline, the concentration of the external high-viscosity discharge flow also experiences a corresponding contraction hysteresis due to incomplete crushing.

[0088] Based on the number of consecutive rows in the two-dimensional coupling matrix where the evolution slope and the attenuation increment change in the same direction and the corresponding time span, the continuous time consumption characteristics are determined, which facilitates the subsequent identification of hidden mechanical faults. Among them, the change in the same direction means that the evolution slope and the attenuation increment show an increasing trend or a decreasing trend at multiple consecutive time stamps.

[0089] In step S105, the results of the self-diagnosis of the abnormality of the crushing equipment are determined.

[0090] The evolution slope, attenuation increment, and duration characteristics are combined to construct a multidimensional spatiotemporal state quantity, which is then used to determine the self-diagnosis results of anomalies in the crushing equipment.

[0091] In one embodiment, the evolution slope, attenuation increment, and duration feature are combined to construct a multidimensional spatiotemporal state variable, including: marking the duration feature as a continuous duration value when the evolution slope is within a preset gentle slope range and the attenuation increment is within a preset slow increment range; determining the absolute value of the conversion ratio of the crushing equipment at the current monitoring node; and concatenating the continuous duration value, evolution slope, attenuation increment, and absolute value of the conversion ratio to construct a first state variable as a multidimensional spatiotemporal state variable.

[0092] When processing bio-fermented materials that may contain lactic acid and high moisture content, the metal parts at the edge of the hammer blades of the crushing equipment may be subject to slight chemical corrosion in addition to friction from the bio-feed material. If the evolution slope is within the preset gentle slope range and the attenuation value increment is also within the preset slow increment range without abrupt changes, it can be determined that the crushing equipment has not suffered serious fracture damage in a short period of time.

[0093] In this embodiment, the preset gentle slope range, preset slow increment range, preset efficiency critical point, preset abrupt slope threshold, preset step increment threshold, and the period duration and abrupt change duration in the degradation rule conditions can be predetermined through offline calibration experiments. For example, under the condition that the crushing equipment is in a known healthy state, at least one complete production batch is continuously run. Following the same acquisition frequency and processing procedure as online monitoring, the acoustic signal sequence and ultrasonic amplitude sequence are acquired, and the evolution slope benchmark value, attenuation numerical increment benchmark value, and conversion ratio benchmark value under the healthy benchmark state are calculated.

[0094] When the crushing equipment is in a known wear-over-limit state, that is, when the wear of the hammer blades reaches the replacement threshold specified by the equipment manufacturer, the data collection and calculation process is repeated to obtain the wear evolution slope, attenuation value, incremental wear value, and conversion ratio wear value under the wear state.

[0095] The upper limit of the preset gentle slope range is set as a value at a certain ratio between the baseline value of the evolution slope and the wear value of the evolution slope. The ratio is selected in the range of 30% to 50% according to the requirements of sensitivity to false alarms. The preset efficiency critical point is set as the conversion ratio wear value. The cycle time is set as the statistical average of the historical operating time required for the crushing equipment to degrade from a healthy state to a wear-exceeding state. The statistical average is obtained by summarizing and analyzing the historical operation and maintenance records of multiple equipment of the same model.

[0096] Under the condition that the crushing equipment is in a known screen breakage state, the data acquisition and calculation process is repeated to obtain the evolution slope jump value and the attenuation value jump value under the breakage state; the preset abrupt change slope threshold is set as the evolution slope jump value; the preset step increment threshold is set as the attenuation value jump value; the abrupt change duration is set as the statistical value of the measured time from the occurrence to the stabilization of the screen breakage event. The statistical value of the measured time is obtained by analyzing the sensor data of historical screen breakage failure cases.

[0097] The relatively weak trend of evolution and deterioration may be masked by the noise caused by uneven moisture content in the biomass flow. The continuous time characteristic of the steady deterioration trend can be marked as a continuous duration value representing the long-term cumulative time span. In order to accurately assess the severity of the final state caused by wear accumulation and determine the maintenance timing, the absolute value of the conversion ratio of the crushing equipment at the current monitoring node can be determined. The absolute value of the conversion ratio indicates the work performance of the crushing equipment in processing biomass at the current moment.

[0098] By combining and manipulating structured data frame parameters, including continuous duration values, evolution slope, decay increments, and absolute values ​​of conversion ratios, a first state variable can be constructed to describe incremental metal fatigue damage.

[0099] The first state variable can be a feature vector representing multiple dimensions obtained by normalizing these different data and combining their features. This allows for comparison of the current state information to be determined with the reference information determined at historical times, thus determining the diagnostic result of the abnormal operating condition of the feed crushing equipment at the current time.

[0100] In one embodiment, the abnormal self-diagnosis result of the crushing equipment is determined by using multidimensional spatiotemporal state variables, including: determining the continuous duration value and the absolute value of the conversion ratio recorded in the first state variable; when it is confirmed that the continuous duration value is greater than the period duration in the preset degradation rule conditions and the absolute value of the conversion ratio is greater than the preset efficiency critical point, it is determined that the degradation rule conditions are matched and the wear exceeding the limit result is output as the abnormal self-diagnosis result of the crushing equipment.

[0101] From the determined first state quantity, the continuous duration value used to characterize the cumulative degradation time and the absolute value of the conversion ratio characterizing the current specific state of degradation can be determined. The moisture content of the biomass at the bottom of different fermentation tanks has a certain natural fluctuation. When it is confirmed that the continuous duration value is greater than the cycle length in the set degradation rule condition of 150 minutes, the short-term disturbance fluctuation caused by the uneven moisture content of a single batch of biomass material is eliminated by using an ultra-long time window.

[0102] When it is confirmed that the continuous duration is greater than the cycle length in the preset degradation rule conditions, and the absolute value of the current conversion ratio is greater than the preset efficiency critical point, it indicates that the combined effect of corrosion and impact may cause the hammer blade edges to be completely rounded, affecting the effective cutting force of the hammer blade edges on the high-toughness bio-fiber. A large amount of wet material stuck in the inner cavity may only be subjected to ineffective rigid metal friction and pushing.

[0103] Under the condition that the physical parameters of the crushing equipment match the degradation rule, the wear result of the hammer wear being seriously excessive and causing serious obstruction of biological material discharge can be output as the abnormal self-diagnosis result of the crushing equipment, so as to remind maintenance personnel to intervene in time to replace the hammer array.

[0104] In one embodiment, the evolution slope, decay increment, and duration feature are combined to construct a multidimensional spatiotemporal state variable, including: when the evolution slope exceeds a preset abrupt change slope threshold and the decay increment simultaneously exceeds a preset step increment threshold within a preset evaluation time window, determining the over-limit start point and over-limit end point where an over-limit jump occurs; calculating the jump duration value by subtracting the over-limit start point from the over-limit end point, and marking the duration feature as the jump duration value; and concatenating the jump duration value, the evolution slope exceeding the preset abrupt change slope threshold, and the decay increment exceeding the preset step increment threshold to construct a second state variable as a multidimensional spatiotemporal state variable.

[0105] Unlike the slower chemical corrosion or mechanical wear and degradation processes, when the arc-shaped screen at the bottom of the equipment is struck by a sharp foreign object mixed in with the biological material and instantly physically breaks and tears, the large particles of fermented grain that were originally confined and repeatedly kneaded inside the screen will lose the constraint of the metal mesh. The unconstrained coarse and hard particles may directly impact the external discharge shell, or a large amount of wet and heavy material that has not been fully cut may be poured out due to the screen hole, causing a surge in high-frequency hard impact noise in the acoustic spectrum. The evolution slope within a preset evaluation time window of, for example, 5 seconds exceeds the preset abrupt change slope threshold.

[0106] A large number of substandard, high-moisture coarse particles, mixed with a large amount of suspended dust, flowed directly into the discharge port without filtration, causing the concentration of sound-blocking particles in the local space to increase. This caused the attenuation value sensed by the external ultrasonic receiver to simultaneously experience a cliff-like increase and exceed the preset step increment threshold.

[0107] When both parameters representing internal acoustic anomalies and external material flow viscosity concentration anomalies are in extreme deterioration, the sampling time vernier can be used to backtrack and determine the starting point of the signal exceeding the limit waveform and the ending point of the signal waveform reaching the extremely deteriorated peak steady state.

[0108] The transition time is obtained by performing time-domain difference calculation between the precise timestamp of the limit termination point behind the time axis and the timestamp of the limit start point ahead of the time axis. The difference calculation formula for the transition time is expressed as follows: , Jump time value, Exceeding the limit termination point, The starting point of exceeding the limit.

[0109] Since the material spill caused by screen rupture is extremely violent and short-lived, the sustained time characteristic used to measure the development speed of sudden adverse events can be marked as the jump time value. By combining the jump time value, evolution slope and attenuation value increment as parameters, a second state quantity can be constructed. The second state quantity is a multi-dimensional spatiotemporal state quantity that can be used to capture sudden risk conditions of crushing equipment in feed processing.

[0110] In one embodiment, the method of using multidimensional spatiotemporal state variables to determine the abnormal self-diagnosis result of the crushing equipment includes: determining the jump time value recorded inside the second state variable; if the jump time value is less than the mutation duration in the preset degradation rule conditions, determining the matching mutation rule conditions; and outputting the screen breakage result as the abnormal self-diagnosis result of the crushing equipment.

[0111] The transition time value used to determine the speed of the event can be extracted from the second state quantity. In order to distinguish normal material falling surge, the sudden change time can be set to 2 seconds for example. If the transition time value is less than the sudden change time in the preset degradation rule conditions, it indicates that the phenomenon of high-frequency impact accompanied by dense fog obstruction is not a change process caused by the natural transition of biological material formula switching, but rather originates from the damage to the load-bearing structure of the internal metal isolation net.

[0112] After aligning and verifying the acoustic internal spatial characteristics and the ultrasonic external spatial characteristics, the abrupt change rules and conditions pre-defined for sudden failure conditions can be determined. The result of the screen breaking due to damage and leakage of a large amount of wet material can be used as the abnormal self-diagnosis result of the crushing equipment to output the highest level red alarm and trigger the emergency stop brake of the equipment in conjunction with the servo drive.

[0113] In one embodiment, an amplitude demodulation algorithm can be applied to the acoustic signal sequence to determine the low-frequency envelope excited by the internal drive rotor of the crushing equipment in a single rotation cycle, and to count the envelope repetition frequency of the low-frequency envelope within a preset evaluation time. The envelope repetition frequency is compared with the current motor speed frequency of the crushing equipment. When it is confirmed that the envelope repetition frequency matches the motor speed frequency, the peak-valley amplitude difference of the low-frequency envelope is determined. When it is confirmed that the peak-valley amplitude difference shows an expanding trend within a consecutive preset number of cycles, the rotor imbalance result is output as the abnormal self-diagnosis result of the crushing equipment.

[0114] During long-term processing, biological materials are prone to adhesion and clumping on one side of the rotor. When the hammers suspended on the heavy rotor assembly experience local breakage and detachment or severe unilateral material adhesion due to gravity, resulting in uneven mass distribution, they can generate strong periodic eccentric oscillation waves.

[0115] For acoustic signal sequences containing multiple complex frequencies in the form of original high-frequency carrier waves acquired by attaching to the outer shell, a digital amplitude demodulation algorithm based on envelope extraction can be applied.

[0116] After removing the masking interference of the high-frequency rigid impact oscillation waveform by using a low-pass filter, the low-frequency envelope excited by the overall center of gravity shift and the swinging of the outer shell can be determined within a single rotation cycle around the main shaft of the high-speed driven rotor inside the crushing equipment. The low-frequency envelope usually exhibits a continuous and regular undulating low-frequency wave-shaped topological feature in the two-dimensional coordinate system.

[0117] The peaks of the undulation period intervals of the wave-shaped topology feature are identified and counted, and the envelope repetition frequency of repeated fluctuations of the low-frequency envelope within a preset evaluation time of, for example, 15 seconds is statistically analyzed. The envelope repetition frequency obtained based on acoustic features is compared and verified with the motor speed frequency electrically fed back by the Hall encoder built into the powertrain of the crushing equipment in a positive and negative closed loop.

[0118] When it is confirmed that the envelope repetition frequency derived from acoustic back-reasoning matches the motor speed frequency read by electrical reading, it indicates that the periodic ultra-low frequency acoustic anomaly is caused by the eccentric mechanical defect of the rotor itself rotating with the main shaft.

[0119] The peak-to-valley amplitude difference between the highest peak amplitude and the lowest trough amplitude within a single wave undulation cycle of the low-frequency envelope is determined. The peak-to-valley amplitude difference is proportional to the magnitude of the eccentric ejection force, and the peak-to-valley amplitude difference can characterize the severity of the centrifugal ejection torque caused by the eccentric mass deviating from the axis.

[0120] When it is confirmed that the peak-valley amplitude difference shows an increasing trend within a consecutive preset cycle of, for example, 10 consecutive cycles, it indicates that the wear of the metal gap of the support bearing caused by the eccentric force is in a vicious cycle and is aggravated. The rotor imbalance caused by the displacement of the center of gravity or the complete failure of the dynamic balance of local heavy paste can be output as the abnormal self-diagnosis result of the crushing equipment, reminding maintenance personnel to adjust the counterweight balance, so as to avoid the safety accident of fatigue fracture of the main bearing seat due to long-term eccentric alternating stress.

[0121] Figure 2 The waveform diagram shows the acoustic signal sequence and low-frequency envelope characteristics of the crushing equipment, as shown below. Figure 2 As shown, Figure 2 The upper part shows the original high-frequency carrier wave signal sequence acquired by the acoustic receiver. Its vertical axis represents the relative amplitude of the acoustic wave signal sequence, which includes a variety of complex frequency aliasing waveforms generated when the crushing equipment is operating.

[0122] After applying an amplitude demodulation algorithm to the acoustic signal sequence, it is possible to extract, such as Figure 2 The lower half shows the low-frequency envelope, which exhibits a continuous and regularly undulating low-frequency wave-like topological feature in a two-dimensional coordinate system. Figure 2 As can be seen from the diagram, by identifying the undulation period interval between adjacent peaks of this wave-shaped topological feature, the corresponding envelope repetition frequency can be statistically determined.

[0123] Within a single wave cycle, the peak-to-trough amplitude difference can be intuitively determined by the vertical difference between the highest peak amplitude and the lowest trough amplitude. Figure 2 The extracted time-domain and amplitude features can directly map the intensity of low-frequency oscillations excited by rotor eccentricity, thus providing data support for the abnormal diagnosis of rotor imbalance.

[0124] In one embodiment, a thermal imaging image of the outside of the top feed channel of the crushing equipment can also be obtained, and the spatial geometric contour of the continuous high temperature zone induced by the friction between the falling material and the inner wall of the channel can be determined from the thermal imaging image. When it is confirmed that the spatial geometric contour shrinks in the preset lower half of the feed channel and generates high-heat agglomerates within a specific height of the feed channel, the agglomeration blockage abnormality is output as the abnormal self-diagnosis result of the crushing equipment.

[0125] In abnormal processing environments with high ambient temperatures or high sugar and oil content in bio-fermentation materials, the biomaterials may stubbornly adhere and clump together on the inner wall of the feed metal pipe as they fall inside. To facilitate analysis of whether there are clumps within a specific height of the feed channel, an infrared thermal imaging image of the outer metal wall of the top feed channel of the crushing equipment can be obtained.

[0126] Under the influence of gravity, the falling material undergoes viscous, decelerated frictional motion against the inner wall of the metal channel. The potential energy of frictional resistance is converted into heat energy and dissipated, and then conducted through the thin metal wall to the outer surface. From the color temperature field cloud array of the thermal imaging image, the spatial geometric contour of the continuous high-temperature zone caused by the continuous friction between the internal falling material stream and the inner wall of the channel can be determined by finding the extreme value line of the temperature gradient.

[0127] Under normal, unobstructed operation with a completely smooth inner wall, the spatial geometric profile of the continuous high-temperature zone exhibits a cylindrical distribution with width variations within a preset normal fluctuation range due to the uniform distribution of frictional forces across the entire cross-section.

[0128] When it is confirmed that the spatial geometric contour of the continuous high temperature zone suddenly shrinks in the lower half of the feed channel and is accompanied by irregular twisting and deformation of the edges, it indicates that the flow cross-sectional area of ​​the region is greatly reduced due to the adhesion of the bio-adhesive to the wall, which causes the subsequent material flow to be squeezed and the friction work to be greatly intensified.

[0129] Based on infrared thermal field distortion and high-temperature accumulation, the agglomeration and blockage abnormalities that severely hinder the feeding flow efficiency and seriously threaten the survival rate of biological flora can be output as the abnormal self-diagnosis results of the crushing equipment.

[0130] In this way, the output of diagnostic results that integrate the spatial temperature field distribution characteristics can prompt on-site operators to use manual unblocking or activate the high-pressure pneumatic vibration tapping auxiliary device to thoroughly dredge and clear the feed channel, ensuring that the feed channel of the crushing equipment can maintain a smooth feeding state.

[0131] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0132] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A self-diagnostic method for abnormal operating conditions in a feed processing process, characterized in that, include: Acquire the acoustic signal sequence of the crushing equipment in the feed processing line, convert the acoustic signal sequence to the frequency domain to generate an acoustic spectrum, and determine the high-frequency energy value of the corresponding mechanical collision and the low-frequency energy value of the corresponding material friction in the acoustic spectrum. The ratio of high-frequency energy value to low-frequency energy value is used as the energy conversion ratio. The energy conversion ratios are spliced ​​together according to the sequence acquisition order to obtain the conversion ratio sequence. The ultrasonic amplitude sequence of the crushing equipment is acquired, the ultrasonic amplitude sequence is denoised to determine the attenuation value, the attenuation change rate of the attenuation value is determined, and the attenuation change rate is spliced ​​according to the sequence acquisition order to obtain the change rate sequence. Align the conversion ratio sequence and the rate of change sequence along the timestamp dimension, determine the evolution slope of the aligned conversion ratio sequence and the decay increment of the aligned rate of change sequence, and determine the continuous time consumption characteristics of the evolution slope and the decay increment maintaining the corresponding trend within the same time window; the decay increment is the integral accumulation result of the rate of change sequence within the preset evaluation time window. The evolution slope, attenuation increment, and duration characteristics are combined to construct a multidimensional spatiotemporal state quantity, which is then used to determine the self-diagnosis results of anomalies in the crushing equipment.

2. The self-diagnosis method for abnormal operating conditions in the feed processing process according to claim 1, characterized in that, Acquiring the acoustic signal sequence of the grinding equipment in the feed processing line, and acquiring the ultrasonic amplitude sequence of the grinding equipment, including: The acoustic receiver is attached to the surface of the crushing equipment shell to continuously collect the vibration waves excited by the mechanical collision in the inner cavity to obtain the sound wave signal sequence. The ultrasonic probes installed on both sides of the discharge chute at the bottom of the crushing equipment continuously emit ultrasonic waves toward the powder falling channel and receive the residual ultrasonic waves that penetrate laterally into the powder falling channel. The residual ultrasound waves are processed by energy attenuation conversion to obtain the ultrasound amplitude sequence.

3. The self-diagnosis method for abnormal conditions in feed processing according to claim 1, characterized in that, Determine the high-frequency energy values ​​corresponding to mechanical collisions and the low-frequency energy values ​​corresponding to material friction in the acoustic spectrum, including: In the acoustic spectrum diagram, the hard impact segment is identified as the high-frequency range, and the high-frequency energy value is obtained by amplitude integration of the energy distribution within the high-frequency range. In the acoustic spectrum diagram, the flexible friction section is identified as the low-frequency range, and the low-frequency energy value is obtained by amplitude integration of the energy distribution within the low-frequency range.

4. The self-diagnosis method for abnormal conditions in feed processing according to claim 1, characterized in that, The ultrasonic amplitude sequence is denoised to determine the attenuation value, including: The ultrasonic amplitude sequence is segmented according to a predetermined step size to obtain multiple sets of time-domain data segments; For each set of time-domain data segments, the arithmetic mean calculation is performed to obtain the average attenuation amplitude corresponding to the time-domain data segment, and the average attenuation amplitude is used as the attenuation value of the corresponding time node.

5. The self-diagnosis method for abnormal operating conditions in the feed processing process according to claim 1, characterized in that, Determine the duration characteristics of the evolution slope and the decay increment maintaining the corresponding trend within the same time window, including: Obtain the absolute timestamp shared by the aligned transformation ratio sequence and the rate of change sequence. Use the absolute timestamp as the matrix mapping index. Pair the evolution slope and decay value increment corresponding to the same absolute timestamp in a plane to obtain a two-dimensional coupling matrix. By using a two-dimensional coupling matrix to track the changes in the evolution slope and the decay value increment on the time axis, the duration characteristics of the evolution slope and the decay value increment maintaining the corresponding trend within the same time window are determined based on the change trajectory.

6. The self-diagnosis method for abnormal operating conditions in the feed processing process according to claim 1, characterized in that, The evolution slope, decay increment, and duration characteristics are combined to construct a multidimensional spatiotemporal state quantity, including: If the evolution slope is within a preset gentle slope range and the decay value increment is within a preset slow increment range, the continuous time consumption feature will be marked as a continuous duration value. Determine the absolute value of the conversion ratio of the crushing equipment at the current monitoring node; By concatenating the continuous duration value, evolution slope, decay increment, and absolute value of the transformation ratio, the first state variable, which serves as a multidimensional spatiotemporal state variable, is constructed.

7. The self-diagnosis method for abnormal operating conditions in the feed processing process according to claim 6, characterized in that, The results of anomaly self-diagnosis for crushing equipment are determined using multidimensional spatiotemporal state variables, including: Determine the continuous duration value and the absolute value of the transition ratio recorded within the first state variable; When the continuous duration is confirmed to be greater than the cycle duration in the preset degradation rule conditions, and the absolute value of the conversion ratio is greater than the preset efficiency critical point, the degradation rule conditions are determined to be matched and the wear exceeding the limit result is output as the abnormal self-diagnosis result of the crushing equipment.

8. The self-diagnosis method for abnormal operating conditions in the feed processing process according to claim 1, characterized in that, The evolution slope, decay increment, and duration characteristics are combined to construct a multidimensional spatiotemporal state quantity, including: If the evolution slope within the preset evaluation time window exceeds the preset abrupt change slope threshold, and the decay value increment simultaneously exceeds the preset step increment threshold, determine the over-limit start point and over-limit end point where an over-limit jump occurs. The jump time value is obtained by subtracting the jump start point from the limit termination point, and the continuous time feature is marked as the jump time value. The jump time value, the evolution slope exceeding the preset abrupt change slope threshold, and the decay value increment exceeding the preset step increment threshold are concatenated as parameters to construct the second state variable as a multidimensional spatiotemporal state variable.

9. The self-diagnosis method for abnormal operating conditions in the feed processing process according to claim 8, characterized in that, The results of anomaly self-diagnosis for crushing equipment are determined using multidimensional spatiotemporal state variables, including: The transition time value recorded in the second state variable is determined. If the transition time value is less than the mutation time in the preset degradation rule condition, the mutation rule condition is matched and the screen breakage result is output as the abnormal self-diagnosis result of the crushing equipment.

10. The self-diagnosis method for abnormal operating conditions in the feed processing process according to claim 1, characterized in that, The method further includes: An amplitude demodulation algorithm is applied to the acoustic signal sequence to determine the low-frequency envelope excited by the internal drive rotor of the crushing equipment in a single rotation cycle, and the envelope repetition frequency of the low-frequency envelope within a preset evaluation time is statistically analyzed. The envelope repetition frequency is compared with the current motor speed frequency of the crushing equipment. When it is confirmed that the envelope repetition frequency matches the motor speed frequency, the peak-valley amplitude difference of the low-frequency envelope is determined. When it is confirmed that the peak-valley amplitude difference shows an expanding trend within a continuous preset cycle, the rotor imbalance result is output as the abnormal self-diagnosis result of the crushing equipment.

Images