System and method for automatically capturing technology transfer knowledge
By automatically capturing and converting implicit knowledge and generating a complete production process model, the implicit knowledge communication problem in technology transfer between pharmaceutical R&D and production teams is solved, achieving faster time to market and lower costs, and improving Accuracy and efficiency of technology transfer.
Patent Information
- Application Number
- CN202010831314.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2019-08-19
- Filing Date
- 2020-08-18
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2040-08-18
AI Technical Summary
During the technology transfer between pharmaceutical R&D and production teams, implicit knowledge is difficult to communicate in formal terms, resulting in extended transfer projects, increased costs and delayed time to market.
Using methods and systems for automatically capturing technology to transfer knowledge, by obtaining production process models, automatically identify missing information and using chatbots and video recording technologies, audio and text information are collected from people participating in new product development, and the implicit knowledge is converted into explicit Knowledge to generate a complete production process model.
Accelerate the technology transfer process, reduce production preparation time and cost, improve the initial transfer accuracy, reduce development costs, and promote collaboration between R&D and production teams.
Smart Images

Figure CN112465288B_ABST
Abstract
Description
Technical Field
[0001] The present invention involves the transfer of technology from a first location where a product is developed to another location where the developed product is to be manufactured. Background Art
[0002] The purpose of pharmaceutical R&D is to develop new products or molecules that can then be used at an industrial level. Typically, at a certain point in the development of a new product, the new product will be authorized for market entry, which requires industrialization of the manufacturing of the new product. This entire process is called New Product Introduction (NPI). This means that the product and process specifications that have been developed in the laboratory will be transferred from the R&D laboratory to the production plant. Such a transfer is called technology transfer (or more commonly "tech transfer"). In practice, R&D and production teams used to work in silos, and the collaboration between them was limited: they tended to interact only during technology transfers.
[0003] Technology transfer is a process (or even a project) that includes:
[0004] - Physically transfer new products, their manufacturing processes and associated analytical methods (i.e., precise methods used to measure process / product parameters or quality attributes, usually in a quality control laboratory) from the issuing location (i.e., R&D laboratory) to the receiving location (i.e., industrial site);
[0005] - Potentially increase batch size while performing transfers;
[0006] - Ensure that the receiving site has the correct understanding of the manufacturing process within its environment.
[0007] The problem may arise from the fact that the issuing location is typically focused on the need to develop new products as quickly as possible (time-driven development), while the receiving location is typically focused on production needs that are closely related to costs (cost-driven production), which results in two different approaches regarding the required information related to technology transfer.
[0008] In practice, there are two types of knowledge that need to be shared between the people at the issuing location and the people at the receiving location:
[0009] -Explicit knowledge, which is formal and explicit and can be easily communicated (i.e., product specifications);
[0010] - Tacit knowledge, which is highly personal, difficult to formalize, and therefore difficult to communicate to others. Tacit knowledge has an important cognitive dimension.
[0011] Throughout the technology transfer project, the issuing site must provide the receiving site with documentation that explains how the relevant new product was developed. This is a very arduous task: for example, a process step that takes 3 days to operate in the laboratory usually means 3 months of documentation work including meetings and discussions. On top of this, it also happens that the resulting documentation may lack some relevant information for the receiving site.
[0012] For example, it is particularly problematic to capture tacit knowledge from R&D experts, which may be necessary for the industrial production of new products. This may then affect the duration of the transfer project and, therefore, the time to market and the costs associated with a particular project. Summary of the invention
[0013] The object of the present invention is to propose a system and method for improving the transfer of technology between an issuing location and a receiving location, the system and method being capable of effectively capturing tacit knowledge and converting tacit knowledge into explicit knowledge, so that information about new products or processes can be easily transferred to the receiving location, thereby eliminating the need to create a large amount of documentation for the new process or product, and saving time and cost for producing new products.
[0014] According to the present invention, this object is achieved by a method and system for automatically capturing technology transfer knowledge according to the object of the present invention.Preferred embodiments of the present invention present other advantages of the present invention.
[0015] The present invention relates to a method for automatically capturing technology transfer knowledge for production of new products, the method comprising:
[0016] - obtaining a production process model, wherein the production process model is a predefined structured model enabling listing of the sequential process steps required to produce the new product, wherein, for each step, the predefined structured model comprises one or several predefined fields, each of which is configured to enable recording of specific information about the production process of the new product (i.e., relevant technology transfer knowledge related to the relevant step);
[0017] - automatically capture missing information in the production process model, wherein the processing unit is configured to determine whether one or several of the predefined fields do not contain specific information, and for each of the predefined fields that do not contain the specific information, the processing unit is configured to:
[0018] i) automatically triggering and executing a chatbot to collect audio and / or text information from personnel involved in the development of a new product, and storing the collected audio and / or text information as specific information for a relevant field; and / or
[0019] ii) automatically starting a video recording process for each field identified as requiring equipment / machine processing for a process step of a production process, wherein the video recording process comprises: automatically detecting interaction with the equipment / machine; starting video recording of the processing of the equipment / machine; automatically stopping recording, in particular automatically stopping recording once the processing unit determines that the interaction with the equipment / machine has stopped for a predefined time threshold; and storing the video recording as specific information for the relevant field;
[0020] Until all fields of the predefined structured model have been reviewed by the processing unit;
[0021] -Once all fields have been reviewed by the processing unit, the complete production process model is exported.
[0022] The present invention also proposes a system for automatically capturing technology transfer knowledge for the production of new products, the system comprising:
[0023] - an information acquisition system connected to the processing unit and comprising at least an input device for collecting information about production steps, such as a camera system and / or a keyboard device and / or a mouse device and / or a touch screen device, the information acquisition system being further configured to acquire a production process model, the production process model being a predefined structured model enabling the listing of the successive process steps required for the production of a new product, wherein for each step the predefined structured model comprises one or several predefined fields, the one or several predefined fields being each configured to enable the recording of specific information about the production process of the new product;
[0024] - a memory connected to the information acquisition system and used to store the production process model and the specific information;
[0025] a processing unit, the processing unit comprising one or several processors, the processing unit being connected to the memory and the information acquisition system, the processing unit being configured to automatically capture missing information in the production process model by controlling the information acquisition system, by determining whether one or several of the predefined fields do not contain the specific information, and for each of the predefined fields that do not contain the specific information, the processing unit is then configured to:
[0026] i) automatically triggering and executing a chatbot to collect audio and / or text information from personnel involved in the development of a new product with the aid of an information acquisition system, and storing the collected audio and / or text information as specific information for a relevant field; and / or
[0027] ii) automatically starting a video recording process for each field identified as requiring equipment / machine processing for a process step of a production process, with the aid of the information acquisition system, wherein the video recording process comprises: automatically detecting an interaction with the equipment / machine; starting a video recording of the processing of the equipment / machine; automatically stopping the recording, in particular as soon as the processing unit determines that the interaction with the equipment / machine has ceased for a predefined time threshold; and storing the video recording in a memory as specific information for the relevant field;
[0028] Until all fields of the predefined structured model have been reviewed by the processing unit;
[0029] The processing unit is further configured to store the complete production process model in the memory once all fields have been reviewed by the processing unit.
[0030] Advantageously, the present invention proposes a method and system capable of automatically capturing information required for the production of new products within a release location by following a production process model that serves as a guide for determining what information is needed / missing and triggering specific actions for collecting the required / missing information. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] A preferred but non-exclusive embodiment of the present invention will now be described with reference to the accompanying drawings, in which:
[0032] Figure 1 A flow chart representing the method according to the invention is schematically shown;
[0033] Figure 2 An example of a production process model for a new pharmaceutical method according to the present invention is shown;
[0034] Figure 3 A preferred embodiment of the system according to the invention is schematically shown. DETAILED DESCRIPTION
[0035] Figure 1 An example of a method for automatically capturing technology transfer knowledge for the production of new products is schematically provided. The method enables automatic acquisition of knowledge about the production of new products and then easy transfer of the acquired knowledge to a receiving location where industrial manufacturing of the new products will take place.
[0036] At step 101, the system according to the invention acquires a production process model. The production process model is based on a predefined structured model that enables listing of the successive process steps required to produce a new product. Preferably, the database or memory of the system according to the invention may include several predefined structured models, each of which has been developed for a specific type of production process, such as for the production of pharmaceuticals, physical materials, etc., wherein the predefined structured models may be classified according to the technology in the database. Since knowledge models are organized into assets, the predefined structured models are configured to construct product / process knowledge to be transferred during technology transfer. The system according to the invention is indeed capable of determining whether a knowledge asset fails within a production process model.
[0037] Typically, for each step of the production process defined by the predefined structured model, the predefined structured model includes one or several predefined fields, each of which is configured to enable the recording of specific information or knowledge about the production process of the new product. Therefore, the operator of the issuing location can pick / select one of the predefined structured models in the database or memory according to the type of process or product to be developed, and one of the predefined structured models must be transferred to the receiving location. Optionally, the operator can fill in one or some of the fields included in the selected predefined structured model. Once the predefined structured model has been selected by the operator, the system according to the present invention is configured to obtain the predefined structured model, that is, the system according to the present invention uploads the predefined structured model to, for example, the memory of the system as a production process model that needs to be completed. Therefore, the production process model obtained by the system according to the present invention is one of the predefined structured models stored in the database, selected by the operator and automatically completed by additional knowledge to describe the production process of the new product.
[0038] Figure 2 A preferred example of a production process model 200 is shown, which is configured for structuring knowledge related to a production process for a product, such as a new pharmaceutical method, in a list of different process steps.
[0039] For each process step, one or several fields are defined. The fields are, for example, at least one of the following or preferably all of them:
[0040] - a description field, in which the step has to be described and / or shown in text, for example by means of some pictures, and in which the name or reference or identifier of the process step can be recorded;
[0041] - Bill of Materials field, which is a list of input materials required to perform the step. For each input material, a list of suppliers can also be recorded;
[0042] - A field for listing a set of machines or equipment that may be involved in the production process with respect to the step under consideration. For each equipment / machine, some information can be recorded, such as the category of equipment / machine, the function of the equipment / machine, the specifications of the equipment / machine, etc.;
[0043] - a material field optionally used to provide a description of the material produced by this step;
[0044] - an instruction field configured to specify how the step must be performed, the instruction field being configured to include one or all of the following features:
[0045] a) electronic work instructions, including in particular work instructions to be presented to the operator responsible for performing the step, and optionally configured to specify know-how information regarding the equipment / machine processing of the step;
[0046] b) a list of critical quality attributes (CQAs) required to certify the quality of the product related to the step under consideration, the list being configured to enable specification of at least one value (e.g. set point, limit) for each CQA and optionally describing an analytical method to be performed in order to check each CQA;
[0047] c) a list of critical control parameters (CPPs) to be monitored during this step, the list being configured to enable specification of at least one value (e.g., set point, limit) for each CPP and optionally describing the analytical method to be performed in order to check each CPP and describing how each CPP affects one or several CQAs;
[0048] A history field configured to collect information about the history of experiments performed during the development of the step under consideration. Preferably, the history field is configured to record within the production process model and for each step its objectives, the settings that have been made (e.g. on the equipment / machine), the specific objectives of the experiments, the measurement results, etc.
[0049] As previously explained, the present invention proposes to create / generate such predefined structured models for different fields of a technology / product / production process, wherein the selection of one of said predefined structured models triggers the acquisition / upload of one of the predefined structured models by the system according to the present invention, and then triggers the completion of one of the predefined structured models by automatically capturing the missing information. The present invention proposes to digitally record the complete production process model for each new product process. To this end and preferably, the information acquisition system according to the present invention may also include a collaborative graphical user interface for acquiring knowledge from operators at the issuing location.
[0050] At step 102, the system is configured to automatically capture missing information in the production process model, wherein a processing unit is configured to determine whether one or several predefined fields are missing specific information, and for each of the predefined fields, determine which certain specific information is missing, and then the processing unit is configured to automatically execute a chatbot and / or acquire a video.
[0051] Preferably, the chatbot is configured for interacting with the operator by voice. For example, when the system according to the present invention determines that specific information, i.e., certain knowledge assets that require certain verbal explanations, are missing and / or have not been clearly defined, then the system is configured for automatically asking the operator predefined corresponding specific questions and for automatically recording the operator's responses in natural language and / or by using any of the collaborative graphical user interfaces.
[0052] Therefore, the chatbot is configured to collect audio and / or text information from the operator, and then the system automatically stores the collected audio and / or text information as specific information for the relevant fields. Typically, the chatbot will only focus on questions that require some verbal explanation (e.g., experimental observations or the reasons behind key decisions). Preferably, the information acquisition system according to the present invention includes a headset / microphone combination as an input device for recording each answer to the chatbot's questions in natural language. Preferably, each answer is recorded by the system in 2 formats: as an audio file and as text information after having been parsed by a speech recognition algorithm. Optionally, the system according to the present invention is configured to enable the operator to directly modify the displayed text information in a collaborative graphical user interface.
[0053] In the following, some examples of questions that may be asked by a chatbot according to the present invention are presented:
[0054] - in case of a process step called "recipe", where the system detects that there are no recorded operating instructions, then the chatbot is configured to provide the following message: "Do you want to describe some work instructions for the process step called 'recipe'?". If the operator replies "yes" within an acceptable response time predefined in the system, then the system is configured to automatically record the operator's instructions and store the instructions as text information as an EWI of the instructions field;
[0055] - If there is no documented “Impact on CQAs” for a given CPP, the chatbot will ask: “Do you know which CQAs are impacted by the ‘Reactor Speed’ CPP?”
[0056] - If no input reason exists for a given experiment, the chatbot will ask: "What is the purpose of experiment '248'?"
[0057] - If no input results exist for a given experiment, the chatbot will ask: "What are your observations from experiment '248'?"
[0058] -etc.
[0059] Thus, the system according to the invention is able to convert implicit knowledge into textual information. Advantageously, the textual information generated can then be used to perform some advanced search actions on previously input information from past experiments or product development, and / or to automatically translate the original language of the voice recording into another language, and / or to make associations between past experimental observations.
[0060] Preferably, once the system identifies an operator, for example, based on the operator's login to the system according to the present invention or the network connected to the system according to the present invention, who is involved in the production process of the new product, the chatbot is automatically started by the system according to the present invention. Preferably, the system can be configured to send an email to the operator, wherein the email includes the chatbot's question if the question has not been answered by the operator for a predefined time period since the last time the question was asked or since the last time the operator was identified. In this case, the system is also configured to collect email responses from the operator about the sent questions and record the responses as text information in the corresponding fields of the production process model. In particular, if the operator does not reply to the email after the predefined time period, the system is configured to send the previously sent email to the peer of the operator who is also part of the same technology transfer project. To this end, the system can include in its database a list of people involved for each technology transfer project.
[0061] Figure 3A preferred embodiment of a system 300 according to the present invention is shown. The system 300 comprises an information acquisition system 301, which is connected to a processing unit 302 comprising one or several processors, and the information acquisition system 301 comprises at least one input device configured to obtain information about a production process. Such an input device is, for example, a video camera system 303. Such an input device may also be a microphone, a keyboard, a touch screen, or any device that enables information to be obtained by observing scenes representing steps of a production process and / or by interacting with an operator. The system 300 may also comprise a memory or database 304 for storing any records and predefined structured models. The database 304 is, for example, connected to the processing unit 302 and / or the information acquisition system 301. Optionally, the system 300 may also comprise one or several displays 305 and / or a collaborative graphical user interface. The information acquisition system 301 is typically configured to be installed in a laboratory to obtain production process knowledge from the development steps of a production process for a new product.
[0062] Instead of the chatbot or together with or in collaboration with the chatbot, the processing unit 302 is configured to control a camera system 303 to obtain missing information about one or several fields of the production process model. For example, the processing unit 302 is configured to detect the processing or use of a specific machine or equipment in the laboratory, and the detection triggers the start of a video recording process. Once the interaction of the operator with the machine or equipment is detected or identified, the video recording process can automatically start the recording of the process steps of the production process with the help of a camera, and / or the chatbot can be used to ask the operator whether the processing must be recorded, and the recording is only started if a positive answer from the operator is identified. Video recording enables the collection of additional expertise about the machine / equipment processing for the steps considered in the production process.
[0063] Preferably, the processing unit is configured to identify within the image of the laboratory each equipment / machine that may potentially be used with respect to the production process. Based on this location knowledge and / or based on the connection to certain equipment / machines, the system according to the present invention is able to identify whether an operator is about to interact with the equipment machine or is interacting with the equipment machine. The identification can be based, for example, on a machine learning algorithm configured to detect the interaction of the operator with the machine / equipment based on the stream of images acquired by the video system. For example, if an interaction is detected in the image acquired by the camera system 303, the processing unit 302 can automatically start video recording of the processing of the equipment / machine until the end of the interaction with the machine / equipment is detected, at which point the processing system 302 automatically stops recording and stores the corresponding video file in the memory 304 for a specific field of the production process model. The field is, for example, a field that includes an indication that the machine / equipment must be used for the corresponding step. For example, the processing unit is configured to identify each field of the production process model including the processing of the machine / equipment, and to ask via a chatbot which of the fields is the correct field for the relevant step. Preferably, the information acquisition system may also include a motion sensor that may cooperate with the camera system 303 to trigger the start of recording. The motion sensor is typically configured to detect motion associated with a single machine / equipment in the laboratory. In this case, whenever the motion sensor detects the single machine / equipment, the processing unit 302 will start the video recording process by activating the camera system pointing to the single machine / equipment or including the single machine / equipment in its field of view. Preferably, at the end of the video recording session, a message is sent to the operator of the laboratory, for example, via a chatbot or via a speaker of the information acquisition system, asking the operator whether he wants to keep or delete the video recording. Preferably, if the operator chooses to keep the video recording, the chatbot may be configured to require the operator to enter some context information on the recording. The microphone or any other input device of the information acquisition system 301 may be used to record the context information provided by the operator. Preferably, the context information provided by the operator is automatically recorded in two formats: as an audio file and as text information after having been parsed by a speech recognition algorithm. Optionally, the system is configured to enable an operator to modify displayed textual information directly in the collaborative graphical user interface.
[0064] The context information may typically contain a process step name, a reference or identifier, and a description of the step or some observations made by an operator in the lab, etc. The video recording together with the recorded context information may then be automatically and collectively stored in the memory 304 (at the correct location with respect to the field due to the added context), thereby allowing the system to link it to the correct production process field. Ultimately, the recorded information with respect to the processing of the machine / equipment may be automatically integrated into some electronic work instruction for the production process step in question, and may then be automatically displayed by the system to a shop floor operator requiring the same type of equipment / machine processing on the production line.
[0065] Therefore, the processing unit 302 is configured to check whether each field of the production process model has been reviewed. It must be understood by review that for each field, the system will automatically record some information or withdraw the field from the production process step because the field is irrelevant or not needed with respect to the production process step in the information obtained through the system and from the operator.
[0066] Once all fields have been reviewed by processing unit 302, at step 103 the system automatically outputs a complete production process model which can be used by the receiving site for industrial production of the product according to the production process steps developed by the issuing site.
[0067] Compared with existing solutions, the present invention presents the following advantages:
[0068] - It accelerates pharmaceutical time to market, automating the human knowledge transfer process in a very practical and flexible way, making this knowledge reusable for future development projects in R&D and during technology transfer to production, reducing development costs and improving the right first time rate (i.e. the rate at which things are done right the first time to better manage risk assessment, which means more prevention of defects than defect detection and rework).
[0069] - It is able to extract implicit knowledge based on the interaction with the operator or by recording the actions / experiments implemented in the laboratory, making the corresponding knowledge automatically available within a predefined structured production process model.
[0070] Other advantages of the present invention are:
[0071] -A deeper understanding of products and processes;
[0072] - Reduction in time and effort for technology transfer projects;
[0073] - Reduction of paper documentation;
[0074] -First correct improvement;
[0075] -Deeper collaboration between R&D and production stakeholders;
[0076] -Continuous improvement of closed-loop manufacturing methods;
[0077] - Correlation between laboratory batches / experiments and production batches;
[0078] -Reduction in time and effort for future product development projects.
Claims
1. A method for automatically capturing technology transfer knowledge for production of new products, the method comprising: a) acquiring (101) a production process model, wherein the production process model is a predefined structured model that enables listing of the successive production process steps required to produce the new product, wherein, for each step of the production process, the predefined structured model comprises one or several predefined fields, wherein the one or several predefined fields are each configured to enable automatic recording of specific information about the production process of the new product; b) automatically capturing (102) missing information in the production process model, wherein the processing unit (302) is configured to determine whether one or several of the predefined fields do not contain the specific information, and automatically start a chatbot for each of the predefined fields that do not contain the specific information and / or automatically start a video recording process for each field identified as requiring equipment / machine processing for a process step of the production process, until all fields of the predefined structured model have been reviewed by the processing unit, wherein the video recording process includes: automatically detecting an interaction with the equipment / machine; and starting a video recording of the processing of the equipment / machine in the event of a detected interaction; c) Once all fields have been reviewed by the processing unit, the complete production process model is output (103).
2. The method according to claim 1, wherein: The processing unit (302) is configured to collect audio and / or text information with the help of the chatbot and regarding process steps for producing the new product, and store the collected audio and / or text information as specific information for relevant fields.
3. The method according to claim 1 or 2, wherein: The video recording process also includes: automatically stopping said recording; and The video record is stored as specific information for the relevant fields.
4. A system for automatically capturing technology transfer knowledge for production of new products, the system comprising: an information acquisition system (301), connected to the processing unit (302), comprising at least an input device for automatically collecting information about production steps, and configured to acquire a production process model, the production process model being a predefined structured model enabling the listing of the successive process steps required for producing the new product, wherein, for each step, the predefined structured model comprises one or several predefined fields, each of the one or several predefined fields being configured to enable the recording of specific information about the production process of the new product; - a memory (304), the memory (304) being connected to the information acquisition system (301) and being used to store the production process model and the specific information; - the processing unit (302) being configured to automatically capture missing information in the production process model by controlling the information acquisition system (301) until all fields of the predefined structured model have been reviewed by the processing unit, and The processing unit (302) is also configured to automatically store the complete production process model in the memory once all fields have been reviewed by the processing unit (302), wherein the processing unit (302) is also configured to determine whether one or several of the predefined fields do not contain the specific information, and automatically start a chatbot for each of the predefined fields that do not contain the specific information and / or automatically start a video recording process for each field identified as requiring equipment / machine processing for a process step of the production process until all fields of the predefined structured model have been reviewed by the processing unit, wherein the video recording process includes: automatically detecting interaction with the equipment / machine; and starting video recording of the processing of the equipment / machine when an interaction is detected.
5. The system according to claim 4, wherein: For each of the predefined fields that does not contain the specific information, the processing unit (302) is configured to: i) With the aid of the information acquisition system (301), the chatbot is automatically triggered and executed to collect audio and / or text information from personnel involved in the development of the new product, and the collected audio and / or text information is stored as specific information for relevant fields.
6. The system according to claim 4, wherein: The video recording process further includes: automatically stopping the recording in case the detected interaction ends; and storing the video recording in the memory (304) as specific information for relevant fields.
7. System according to one of claims 4 to 6, wherein: The information acquisition system (301) includes a camera system (303).
8. System according to one of claims 4 to 6, wherein: The information acquisition system (301) includes a motion sensor.
9. The system according to claim 8, wherein: The information acquisition system (301) includes a motion sensor.
Citation Information
Patent Citations
A method and system for knowledge extraction
CN109086440A
Method for managing knowledge flow to create value
EP1418517A1
Rapid knowledge transfer among workers
US20050027696A1
System and method for managing the development and manufacturing of a pharmaceutical drug
US20050197786A1
Automated batch manufactuirng
US20070050070A1
