CN-121973288-A - Nut cooking pretreatment slice processing method

CN121973288ACN 121973288 ACN121973288 ACN 121973288ACN-121973288-A

Abstract

The invention provides a nut cooking pretreatment slicing processing method, which belongs to the technical field of nut cooking pretreatment slicing processing, and is characterized in that the method comprises the steps of grading the particle size of a badam raw material, setting the cooking temperature and the cooking time in groups to finish cooking peeling, detecting the moisture penetration depth of peeled badam batch by utilizing online near infrared moisture gradient detection, returning unqualified batches to supplement cooking, sending the qualified batches to a tunnel oven for baking, transferring the qualified batches to a heat preservation intermediate bin to maintain the surface temperature of nuts, calculating the single feeding amount and the feeding interval time based on a thermodynamic residual temperature decay dynamics optimal slicing batch distribution algorithm, sending the single batch of slices into a slicer in an optimal slicing window according to the parameters, synchronously inputting the slicing yield and all process data of each batch into an artificial intelligent network training data set for iterative updating, and solving the technical problem of low slicing yield caused by the dynamic regulation and deletion of a causal kernel temperature window in the nut slicing processing process.

Inventors

LIU CHENGJIAN
ZHAO HONGEN
WU XIAOXIA
JI CHAO

Assignees

青岛伊德利食品有限公司

Dates

Publication Date: 20260505
Application Date: 20260402

Claims (10)

1. A nut cooking pretreatment slicing processing method, which is characterized by comprising the following steps: Detecting the particle size of the almond raw material, dividing the raw material into a plurality of particle size interval groups according to the particle size grading boundary, and respectively setting corresponding cooking temperature and cooking time for each particle size interval group to finish grouping cooking and peeling; Performing online near infrared moisture gradient detection on peeled badam batch by batch to obtain a moisture penetration depth detection value, comparing the moisture penetration depth detection value with a moisture penetration depth target value, entering a next step from a batch with the moisture penetration depth detection value reaching the moisture penetration depth target value, and returning to the complementary cooking from a batch without reaching the moisture penetration depth target value; Delivering the badam with the moisture penetration depth detection value reaching the moisture penetration depth target value into a tunnel oven for baking at a baking temperature, transferring the badam into a heat preservation intermediate bin after the baking time is over, and preserving the heat of the badam by the heat preservation intermediate bin at the temperature of the heat preservation intermediate bin; The multi-mode time sequence fusion attention network acquires the surface temperature time sequence data of the kernels, the near infrared water gradient time sequence data and the vibration acceleration time sequence data of the cutters in real time, outputs the starting time of an optimal slicing window and the ending time of the optimal slicing window, and transmits the starting time and the ending time to the conveyor belt control system; Based on a thermodynamic residual temperature decay kinetics optimal slice batch distribution algorithm, taking the current environment temperature, the rated processing rate of a slicing machine and the critical temperature of the slice as inputs, calculating single feeding quantity and feeding interval time, and enabling a conveyor belt control system to adjust the feeding rate according to the optimal slice window starting time, the optimal slice window ending time, the single feeding quantity and the feeding interval time, so that the Almond is fed into the slicing machine in batches from the optimal slice window starting time to the optimal slice window ending time to finish slicing; And counting the slicing yield of the current batch after slicing, and synchronously inputting the slicing yield, the current batch cooking temperature, cooking time, baking temperature, baking time, insulation bin temperature, nut surface temperature time sequence data, near infrared water gradient time sequence data and cutter vibration acceleration time sequence data into a multi-mode time sequence fusion attention network training data set for iterative updating of the multi-mode time sequence fusion attention network.
2. The nut cooking pretreatment chip processing method according to claim 1, wherein the determination of the particle size classification boundary is specifically performed by performing a particle size distribution statistical experiment on the badam raw material not less than the threshold of the particle size classification experiment batch, and selecting the particle size node with a larger distribution density as the particle size classification boundary according to the particle size distribution density.
3. The nut cooking pretreatment chip processing method of claim 2, wherein the determination of the cooking temperature and the cooking time is performed by performing a gradient cooking experiment not less than a threshold of a cooking gradient experiment batch for each particle size interval group, establishing a moisture penetration depth prediction model, and iteratively optimizing a cooking temperature value interval and a cooking time value interval corresponding to each particle size interval group with the aim of converging a moisture penetration depth detection value to a moisture penetration depth target value.
4. The nut cooking pretreatment slicing method as set forth in claim 3, wherein the determination of the target value of the moisture penetration depth is performed by slicing the badam of different moisture penetration depth detection values in a slicing experiment not less than the threshold value of the moisture penetration experiment batch, and the minimum moisture penetration depth detection value corresponding to the first time the slicing yield reaches the yield qualification threshold value is taken as the target value of the moisture penetration depth.
5. The nut cooking pretreatment chip processing method of claim 4, wherein the on-line near infrared moisture gradient detection is performed, specifically, the moisture content gradient at different depths below the surface layer of the nut is inverted by utilizing the absorption difference of near infrared light with different wave bands between the surface layer and the subsurface layer of the nut, so as to indirectly represent the moisture penetration depth detection value, and the moisture penetration depth detection value is output after the almond is scanned particle by particle.
6. The method of claim 5, wherein the determination of the baking temperature and the baking time is performed by performing a slicing experiment at different combinations of baking temperature and baking time, wherein the combination of the baking temperature and the baking time is determined such that the surface temperature of the nut is higher than the critical temperature of the slice after the end of the baking, by using the time series data of the slice yield and the surface temperature of the nut as the evaluation index.
7. The method of claim 6 wherein determining the critical temperature of the slices, specifically, performing slicing experiments at different surface temperatures of the kernels, recording the slag breaking rate corresponding to the surface temperature of each kernel, and taking the surface temperature of the kernel corresponding to the first time when the slag breaking rate exceeds the upper limit of the qualified slag breaking rate as the critical temperature of the slices.
8. The nut cooking pretreatment chip processing method of claim 7, wherein the determination of the temperature of the thermal insulation bin is specifically to determine the decay rate of the surface temperature of the almond at different temperatures of the thermal insulation bin, and the minimum maintenance value of the temperature of the thermal insulation bin is determined by iterative experiments under the constraint that the surface temperature of the almond is not lower than the critical temperature of the chip within the time from the start time of the optimal chip window to the end time of the optimal chip window.
9. The nut cooking pretreatment chip processing method of claim 8, wherein an input layer of the multi-mode time sequence fusion attention network receives three types of heterogeneous sensing data streams of nut surface temperature time sequence data, near infrared water gradient time sequence data and cutter vibration acceleration time sequence data, the three types of heterogeneous sensing data streams are respectively sent to three independent one-dimensional time sequence convolution branches to carry out local feature coding, the coding result is sent to a cross-mode cross attention module to be dynamically fused, and then sent to the circulation jumping unit, and an output layer outputs an optimal chip window starting time, an optimal chip window ending time and an expected chip output rate interval.
10. The nut cooking pretreatment chip processing method of claim 9, wherein a physical constraint layer is embedded at the network end of the multi-mode time sequence fusion attention network, and a loss function is embedded in a penalty term form after discretizing an after-heat attenuation differential equation, so that predicted values of the starting moment of an optimal chip window and the ending moment of the optimal chip window do not violate thermodynamic physical rules.

Description

Nut cooking pretreatment slice processing method Technical Field The invention belongs to the technical field of nut cooking pretreatment slice processing, and particularly relates to a nut cooking pretreatment slice processing method. Background Slicing processing of nuts such as badam is an important link in the nut deep processing industry chain. In traditional slicing processing, the cooking pretreatment usually depends on the experience of operators to set fixed cooking temperature and cooking time, static parameter control is adopted in the baking and heat preservation links, the feeding scheduling of the slicing machine depends on manual judgment or fixed beat, and the whole process lacks the capability of sensing the internal moisture state and surface temperature of the nuts in real time. The traditional method can not adaptively adjust the technological parameters according to the grain diameter difference of the badam in the stewing stage, so that the moisture permeation of partial batches is insufficient, and the toughness of the nuts does not reach the standard. In the stage of transferring from baking to slicing, the surface temperature of the kernels naturally decays along with the environment, the traditional method lacks quantitative prediction and active management and control capability of the temperature decay process, and the surface temperature of the kernels often deviates from a proper interval during slicing, so that the slag crushing rate of the slices is increased, and the yield is reduced. In the current nut slicing processing, as the multi-process links from cooking and baking to slicing are longer, and the surface temperature attenuation of nuts is influenced by multi-factor coupling such as particle size distribution, ambient temperature, batch size and the like, the static scheduling scheme in the prior art cannot respond to the temperature attenuation difference among batches in real time, and it is difficult to ensure that each batch of materials is sliced in an optimal temperature window. That is, the prior art has the technical problem that the dynamic regulation and control of the causal kernel temperature window in the nut slicing process results in low slicing yield. Disclosure of Invention In view of the above, the invention provides a nut cooking pretreatment slicing processing method, which can solve the technical problem of low slicing yield caused by the dynamic regulation and control loss of a causal kernel temperature window in the nut slicing processing process in the prior art. The invention provides a nut cooking pretreatment slice processing method, which comprises the following steps: Detecting the particle size of the almond raw material, dividing the raw material into a plurality of particle size interval groups according to the particle size grading boundary, and respectively setting corresponding cooking temperature and cooking time for each particle size interval group to finish grouping cooking and peeling; Performing online near infrared moisture gradient detection on peeled badam batch by batch to obtain a moisture penetration depth detection value, comparing the moisture penetration depth detection value with a moisture penetration depth target value, entering a next step from a batch with the moisture penetration depth detection value reaching the moisture penetration depth target value, and returning to the complementary cooking from a batch without reaching the moisture penetration depth target value; Delivering the badam with the moisture penetration depth detection value reaching the moisture penetration depth target value into a tunnel oven for baking at a baking temperature, transferring the badam into a heat preservation intermediate bin after the baking time is over, and preserving the heat of the badam by the heat preservation intermediate bin at the temperature of the heat preservation intermediate bin; The multi-mode time sequence fusion attention network acquires the surface temperature time sequence data of the kernels, the near infrared water gradient time sequence data and the vibration acceleration time sequence data of the cutters in real time, outputs the starting time of an optimal slicing window and the ending time of the optimal slicing window, and transmits the starting time and the ending time to the conveyor belt control system; Based on a thermodynamic residual temperature decay kinetics optimal slice batch distribution algorithm, taking the current environment temperature, the rated processing rate of a slicing machine and the critical temperature of the slice as inputs, calculating single feeding quantity and feeding interval time, and enabling a conveyor belt control system to adjust the feeding rate according to the optimal slice window starting time, the optimal slice window ending time, the single feeding quantity and the feeding interval time, so that the Almond is fed into the slicing machine in batches from the op