MILQ-QC-TOOL 018176
018176
MILQ-QC-TOOL
Project full title:
The development of predictive models on the Internet for optimisation of heat treatment of raw milk in small and medium-sized dairy companies
sixth framework programme
horizontal research activities involving smes
co-operative research
FINAL Activity Report
Period covered: from 16/9/2005 to 15/12/2007 Date of preparation: 15/01/2008
Start date of project: 16/9/2005 Duration: 2 years, 3 months
Joost van Dijk Revision: 1
Van Dijk Kaasmakerij BV
NIZO food research BV
M.A.I Schutyser
Wirelessinfo
P. Gnip, P. Horak
January 2008
11
January 2008
MILQ-QC-TOOL 018176
1 Executive Summary 3
1.1 Aim 3
1.2 Tasks and objectives 3
1.3 Results 4
1.4 Milestones and deliverables 5
1.5 Corrective actions 7
2 Project objectives and major achievements 8
2.1 General Project Objectives 8
2.1.1 Objectives and targets of this project 8
2.1.2 International state-of-the-art 10
2.2 Summarise the objectives 11
2.3 Problems and corrective action 12
3 Work packages 13
3.1 Progress towards objectives WORK PACKAGE 1: Identification of products, processes, soft- and hardware requirements. 13
3.1.1 Task 1.1 – Collection of processing data and product data. 13
3.1.2 Task 1.2 – Collection of experimental data 15
3.1.3 Task 1.2 – Collection of data to validate the customised/calibrated predictive models 16
3.1.4 Deviations from the project work program and corrective actions 20
3.1.5 Deliverables 20
3.1.6 Milestones 21
3.2 Progress towards objectives WORK PACKAGE 2: Development of the first prototype 21
3.2.1 Task 2.1 – Customisation and validation of the predictive models 22
3.2.2 Task 2.2 – Evaluation of the first prototype 23
3.2.3 Deviations from the project work program and corrective actions 24
3.2.4 Deliverables 24
3.2.5 Milestones 25
3.3 Progress towards objectives WORK PACKAGE 3: Development and hosting of Internet application 25
3.3.1 Task 3.1 – Development of a secure Web page 26
3.3.2 Task 3.2 – Development of a Web database structure 26
3.3.3 Task 3.3 – Development of the basic simulation environment 28
3.3.4 Task 3.4 – Hosting of the Web application 29
3.3.5 Deviations from the project work program and corrective actions 30
3.3.6 Deliverables 30
3.3.7 Milestones 30
3.4 Progress towards objectives WORK PACKAGE 4: Training and implementation of simulation software 31
3.4.1 Task 4.1 Preparation of a user manual 31
3.4.2 Task 4.2 Arrangement of help-desk facilities 32
3.4.3 Task 4.3 Training of SMEs with the modelling tool 32
3.4.4 Deviations from the project work program and corrective actions 34
3.4.5 Deliverables 34
3.4.6 Milestones 34
3.5 Progress towards objectives WORKPACKAGE 5: Project management activities 35
3.5.1 Task 5.1 Overall project management 35
3.5.2 Task 5.2 – Exploitation and dissemination 37
3.5.3 Deviations from the project work program and corrective actions 37
3.5.4 Deliverables 37
3.5.5 Milestones 38
4 Consortium Management 38
5 Other issues 41
Annex I – Project deliverables 42
1 Executive Summary
1.1 Aim
The aim of the project has been to develop a web application, accessible on the internet, for optimization of heat treatment e.g. pasteurization processes in the dairy industry. This project has been carried out by a consortium of SME’s and two research providers. One of the fundamental ideas underlying the project was that SME’s should benefit from the applicable technology that has been developed. The application has potential to improve product quality, maximize food safety and minimize production costs. No integrated user friendly and web based model is available yet to achieve this task.
1.2 Tasks and objectives
The project has been completed in December 2007. The following tasks have been carried out during the course of the project:
1. Inventory of field conditions for heating processes at sixteen European small and medium dairy enterprises; pasteurizing equipment and processes have been described for each of the participating companies;
2. Selection of process-product combinations and collection of technical processing data and product data. The data collection phase has been of prime importance for the development of the prototype;
3. Development of a first prototype of the predictive model, definition of parameters used to optimize processes in the dimensions: production efficiency; pasteurizer fouling; food safety;
4. Design and implementation of a web-database structure and interface with the predictive model;
5. Develop and host a uniform simulation environment as Web application on the Internet;
6. Use the model to predict optimal circumstances for the consortium members selected production processes and report on the recommendations;
7. Collection of experimental data required for customisation and calibration of the available predictive models; i.e. new process settings have been recommended to consortium partners, they have implemented these settings and taken additional samples to validate the model;
8. Use field-data for model validation; additional reports to the participants with supporting evidence;
9. Carry out additional kinetic parameter estimation experiments to validate the model;
10. Write a manual for the resulting application called WebSim-MILQ;
11. Develop learning tools for end users; The tool is user friendly but still requires specialist training, a case has been defined and an e-learning platform has been created;
12. Define a structure for exploitation after project has ended.
13. To provide a forum for knowledge exchange during workshops for all participants
14. To train the dairy firms how to use the knowledge and show the potential of the tool to speed up product development cycles.
1.3 Results
Optimisation of heat treatments in dairy industry: Development of a
new web-based tool with predictive models (WebSim-MILQ).
Abstract
The main reason for heat treatment of milk is inactivation of harmful bacteria (e.g. Listeria monocytogenes) and enzymes such as milk lipase and plasmin. Besides inactivation of microorganisms and enzymes, heat treatment also adversely affects taste and quality of products. Heat treatment procedures such as high temperature/short time (HTST) pasteurization and ultra-high temperature (UHT) treatment have therefore evolved. Throughout 15 years NIZO experts work with predictive models describing chemical, physical and biochemical reactions that occur during heating of milk. This work has lead to the development of NIZO Premia, software with various computer models for optimization of dairy processes (heat treatment, membrane filtration, spray drying, cheese production, and evaporation). Main application of NIZO Premia is improving food safety and quality, reducing of energy consumption, and accelerating new product development. Many of the available predictive models are however not suitable for SME’s, because they are designed for use by R&D experts and thus not always user-friendly. Therefore, 16 small and medium dairy enterprises (SME’s) joined forces to set-up an EU project to develop a web based version of the modeling tools for heat treatment (Websim-MILQ available at websim.milq.org). NIZO food research and Wirelessinfo developed the web application in such a way that it is user-friendly and designed to the specifications of the SME’s. In the project the models were applied for optimization of a selected heat treatment in each SME. A stepwise approach was followed that consisted of process data collection, product analysis, predictive model implementation, and finally optimization of the heating process. This approach led to concrete improvements with respect to process efficiency (e.g. increased cheese yield), food safety, and reduction of fouling. The web application itself was introduced to the SME’s through training sessions and via the development of ultilingual tutorials. Shortly, the web application will be available to any interested dairy company.
Keywords: predictive modeling, heat treatment, web application
Illustration of WebSim-MILQ’s interface
1.4 Milestones and deliverables
Table: Milestones
Milestone no. / Milestone name / WP no. / Date due / Actual delivery date / Lead ContractorMonth 9 / Month 9: After task 1.1 and 1.2. all data required for customising and calibrating the models will be available, which will be done in work package 2. / 1 / 01/09/06 / 01/10/06 / NIZO
Month 14 / Month 14: At the end of this work package all required data, including validation data, are collected. The validation data will be used in work package 2 to validate the customised/calibrated models. The data will be stored in the Web database structure which will be developed in work package 3 (task 3.2 / 1 / 01/02/07 / 01/10/07 / NIZO
Month 12 / Month 12: The first off-line prototype of the model based on the customised/calibrated models will be available and ready to be evaluated by the SMEs. This prototype will also be used as a basis for building the Web enables version of the model (work package 3, task 3.3). / 2 / 01/09/06 / 01/12/06 / NIZO
Month 18 / Month 18: At the end of this workpackage the predictive models will also be validated for the SMEs and each SME will know how to use the models. For each SME the selected heating process has been optimised with respect to product quality and fouling. This means that from this stage all SMEs can run their production with lower costs, whereas product quality and safety are the same. Furthermore, the Web enables version of the model can now be completed (workpackage 3). / 2 / 01/06/07 / 1/9/07 / NIZO
Month 7 / Month 7: Hosting of the Web page will be available and facilitated during the rest of the project. / 3 / 15/04/06 / 01/05/06 / Wirelessinfo
Month 24 / Month 24: At the end of this work package the predictive models will be available on a secure and safe Internet site. / 3 / 01/10/07 / 01/10/07 / Wirelessinfo
Month 24 / Month 24: After this workpackage the implementation of the modelling tool is at each SME is established. All SMEs are capable of working with the Internet application and a manual will be available on how to use this application on the Web. / 4 / 01/10/07 / 15/12/07 / NIZO
Month 23 / Month 23: Licence contract and consortium agreement / 5 / 01/09/07 / 15/12/07 / Van Dijk
Month 27 / Month 27: Final review report. / 5 / 31/12/07 / 15/12/07 / Van Dijk
Month 27 / Month 27: Final Plan for using and disseminating knowledge / 5 / 31/12/07 / 15/12/07 / Van Dijk
Table: Deliverables
Del. no. / Deliverable name / WP no. / Lead participant / Estimated person-months / Nature[1] / Dissem. level[2] / Actual Delivery date (last version)D1 / Work sheet with process and product data / 1 / NIZO / 15 / R / CO / 15/12/07
D2 / Work sheet with experimental data / 1 / NIZO / 39 / R / CO / 15/12/07
D3 / Work sheet with validation data / 1 / NIZO / 23 / R / CO / 15/12/07
D4 / Customised models – first prototype / 2 / NIZO / 29 / P / RE / 15/12/07
D5 / Evaluation of comments SMEs on first prototype / 2 / NIZO / 6 / R / RE / 15/12/07
D6 / Report with results of optimised heat treatment / 2 / NIZO / 24 / R / PU / 15/12/07
D7 / Secure Internet page with public and private part / 3 / Wirelessinfo / 4 / P / PU / 15/12/07
D8 / Hosting, support and maintenance during the project / 3 / Wirelessinfo / 4 / O / RE / 15/12/07
D9 / Web database structure / 3 / Wirelessinfo / 3 / O / RE / 15/12/07
D10 / Brief report of test results simulation environment / 3 / Wirelessinfo / 4 / R / RE / 15/12/07
D11 / Brief report future maintenance protocol / 3 / Wirelessinfo / 1 / R / CO / 15/12/07
D12 / User Manual / 4 / NIZO / 5 / R / RE / 15/12/07
D13 / Report help-desk protocol / 4 / NIZO / 6 / R / CO / 15/12/07
D14 / Models available on the Internet / 4 / NIZO / 10 / P / CO / 15/12/07
D15 / Draft plan for using and dissemination knowledge / 5 / Van Dijk / 1 / R / RE / 15/12/07
D16 / Licence contract / 5 / Van Dijk / 0.5 / O / CO / 15/12/07
D17 / Consortium agreement / 5 / Van Dijk / 0.5 / O / CO / 15/12/07
D18 / Final plan for using and dissemination knowledge / 5 / Van Dijk / 1 / R / RE / 15/12/07
D19 / Annual reports / 5 / Van Dijk / 3 / R / RE / 15/3/07; 30/1/08
TOTAL / 179
1.5 Corrective actions
To validate the model additional micro-heater experiments have been conducted to estimate kinetic parameters and additional visit have been carried out to examine equipment in cases where the model yielded inexplicable results.
The investigation offered two type of explanations. Care should be given to the way sampling was conducted in the field to minimize variances and error margins. The other type of explanation related to the quality of data that were provided from the field. It turned out that actual flows, temperature readings that were assumed to be true in fact were not. After remedying both types of error sources the model predictions yielded acceptable results with a margin of about +/-1.5%.
The second type of corrective actions relates to consortium management. It was deemed necessary to maintain a consortium of sixteen members as originally was foreseen in the project description. This has lead to some replacements during the course of the project and a number of task running concurrently. The effect has been a three month delay in relation to the originally approved starting date of the project.
Due to changes in the consortium some tasks have been carried out later than originally was planned. Overall the work planning was workable. An extension was requested until the 15th of December to complete training of the participating companies and the training material.
2 Project objectives and major achievements
2.1 General Project Objectives
2.1.1 Objectives and targets of this project
Introduction
The aim of this project was to develop an easily accessible user-friendly application for heat treatment in dairy processes that is available on the internet and can safely accessed from terminals in the own environment. The project is set up around the bracket of small and medium sized companies (SMEs) to optimise heat treatment in the production of dairy products. The application is easy to use and the results should be interpretable for the quality/product design personnel of the participating dairy companies. Now the project phase has finished the application will also be made available for other users. Measures have been taken to manage maintenance, security and marketing of the application past the project horizon.