1.

Understand the steps of a research project o Analytical calculations

o Numerical simulations
Research steps
o Questionnaires/focus groups
Generally, the research process can be divided into 6 steps:
(Or a combination)
1. The problem:
5. Analyze
2. The sample
 The data measured on its own does not provide research answers, it must be analyzed.
3. Research Design  There are typically three steps required:
o Data processing
4. Measure
o Data grouping
5. Analyze o Data analysis
6. Conclude 6. Conclude
The problem  Describe results of your research and the implications.
 Some general problem formulation that is typically very broad  How do the results compare to your expectations?
 Why are we interested in this problem?  What are the limitations of your research?
o Society relevance  What would be your next steps?
o Fundamental knowledge  What is the relevance of your study?
o Society relevance
2. The sample o Fundamental knowledge
▪ The problem is usually too broad to tackle all together, therefore it must be split.
▪ It is expected that each of these portions (or samples) will contribute to addressing the question(s)
posed by the problem. 2. Learn the language of research
▪ Choosing the right sample or portion of your problem is crucial for the success of your research.
▪ Knowing the limitations of your sample is extremely important as well. Hypotheses
▪ A hypothesis is a specific statement of prediction.
3. Research Design ▪ It describes in concrete terms what you expect to happen in your study.
▪ Research design provides the glue that holds the research project together. ▪ Depending on your hypotheses you will have a di erent type of study.
▪ A method is chosen to do the research, that clarifies how the researcher is going to proceed. This
method will indicate what I am going to measure and why I am measuring it? Types of studies
▪ In order to design properly your own research, one must be aware of the systems and methods 1. Descriptive studies: designed to document the values of some unknown variable(s).
used by peer researchers. Example: proportion of votes between two parties.
▪ Then, one can develop their own method or utilize an existing one that has been used for a
di erent sample or even a completely di erent problem. 2. Relational studies: investigating the relationships between two or more variables.
Example: relationship between age and voting proportions.
4. Measure
3. Causal studies: designed to determine whether one or more variables cause or a ect one or
 Measurement is the process of observing and recording the observations that are collected more outcome variables.
as part of a research e ort. Example: causal e ect of advertising campaign on voting patterns.
 This measurement can take di erent forms.
o Experiments

What is a variable? o Theory -> Hypothesis -> Observation -> Confirmation

▪ A variable is the smallest block in the language of research.
 Inductive reasoning moves from specific observations to broader generalizations and
▪ The values a variable can take are called attributes.
theories.
▪ The attributes of a variable can be: quantitative (numerical) or qualitative (non-numerical).
▪ Variables should be exhaustive (all possible answers) and mutually exclusive (not simultaneous). o Observation -> Pattern -> Tentative Hypothesis -> Theory
▪ Beware: qualitative values can be based on quantitative judgements and all qualitative data can be
Some colleagues' examples of wrong statements via induction
summarized and manipulated numerically!
 We observe that people who drank water will at some point die. Therefore, drinking water is
Types of relationship between variables
fatal.
• A relationship refers to the correspondence between two variables.
 There was no tra ic when I crossed the road for the past days, so I won't need to look around
• The pattern of the relationship can be:
in the future.
o None
o Positive  1/3 of the students in this university like to eat apples, therefore 1/3 of all university students
o Negative must like to eat apples.
o Curvilinear
 Dutch people are very tall. People from Nepal are not. The flatter the country, the taller the
• The nature of the relationship can be:
people.
o Correlational
o Causal  Every time I am waiting for a bus for too long, I light up a cigarette to kill the wait and, every
o Third-variable problem time I do so, the bus arrives. Therefore, next time I will light up the cigarette, so that the bus
will arrive sooner.
Danger: Correlation is not causation
 The Earth has a moon, Jupiter has moons, therefore all planets have moon(s).

3. Understand the induction and deduction method

4. Understand how research is thought up or conceptualized
The “hourglass” notion of research
Where do research ideas come from?
 Begin with broad questions
▪ Practical problems experienced in the field needing to be addressed.
 Narrow down, focus in ▪ Lack of specific knowledge in the field.
▪ Conclusions of other research studies.
 Operationalize
Feasibility
 Observe
 Is the research idea feasible?
 Measure
 Trade-o s between rigour and practicalities.
 Analyze data  Practicalities include:
o How long the research will take?
 Reach conclusions
o Ethical considerations?
 Generalize back to questions o Is there cooperation required?
o Are the required costs manageable?
Deduction vs Induction
o Do I have access to the tools needed?
 Deductive reasoning moves from the more general to the more specific.
Literature Review ▪ Determines what, why and how much is going to be studied.
• A literature review is designed to identify related research and to set the current research project ▪ 4 phases:
within a conceptual and theoretical context.
1. Objective of the research
• A good literature review answers:
2. Research framework
1. What is already known about my topic and in which discipline is the topic discussed?
3. Research Question/Conceptual model
2. Which theories and concepts are used and discussed within the scope of my topic, and how
are they defined? 4. Definition & Operationalization

3. How is my topic researched and what di erent research methods are there? Technical Research Design
▪ Set of activities needed to realize the objectives set out in the conceptual design phase.
4. Which questions remain unanswered and what has yet to be researched?
▪ Concerns decisions on how, where and when.
Recommendations for your literature search ▪ 3 phases:

1. Concentrate your e orts on the most credible scientific literature. 1. Research Strategy (how)

2. Identify the keywords of your topic and search in academic search engines. 2. Research Material (how)

3. Create a mind map that connects authors/research institutions and areas of knowledge. 3. Research Planning (where/when)

4. If you find a good baseline article, follow their references and where it is referenced.

6. Establish the framework for your conceptual research design

5. Learn about the strategic areas in your research design Types of Research Objectives

Research Design is split into  Theory oriented

 Conceptual Design  Practice oriented

o Research Objective Theory oriented research objectives

o Research Questions 1. Theory development: This type of research attempts to develop a theory about something
not yet known.
o Research Framework
2. Theory validation: This type of research attempts to find out if a proposed variable
o Definitions and Operationalization
relationship exists by testing existing hypotheses.
 Technical Design
Practice oriented research objectives
o Research Strategy
1. Problem Analysis: Recommendations on what technological developments must be made
o Research Material to solve a particular problem.

o Research Planning 2. Diagnosis: Recommendations to improve/optimize a current situation.

Conceptual Research Design 3. Design: Produces a solution to a technical problem.

▪ The modeling of the content of the research.
4. Change: Alterations to a practical situation are made and their impact is analyzed.

The research objective is … (a) … by … (b). Research Questions

(a) the contribution of the research project to the solution of the problem.
1. How do load demand patterns and solar variability a ect battery storage e iciency?
(b) description of the way the contribution will be provided.
2. How does predictive control perform compared to rule-based control in terms of e iciency
Criteria for a good research objective
and cost savings?
When formulating a research objective ensure you can answer these questions!
▪ Informative: Can somebody with generic engineering knowledge understand what the project is 3. Which system parameters most strongly drive overall system performance?
about?
Research framework
▪ Useful: What is the benefit of your research to the problem/solution?
▪ Feasible: Is it possible within your capabilities & resources?  Draw a schematic representation of the steps needed to answer your research questions.
▪ Clear: Is it precise in what your project’s contribution is?  Think about the knowledge that you need.
 Think about the methods you want to use.
Example Research Objective:
o Have they been developed by another researcher?
The research objective is to increase the e iciency of residential solar energy systems by developing
o Do you plan to develop your own method?
and testing a predictive control algorithm for battery energy storage management.
 A thorough literature study will help you, even if you want to develop your own method, you
Research questions can draw from the pitfalls of methods used in your field.
Research Questions will yield information that is useful & necessary for accomplishing the research  Using a known method is acceptable, remember that science is a collective endeavor. Your
objective. They enable us to design the technical part of our research. contribution might not be in the method but in the application, or in the unique combination
How to formulate a Research Question? of existing methods.

1. Identify the main issue that will be investigated by describing the key concept that will be the Definitions and operationalization
focus of the work. Get to the details:
▪ Define the content of key concepts with exact definitions.
2. Think of this first in qualitative terms and then try to design the research question that can ▪ Write down which assumptions you are making and why.
link to quantitative evidence. ▪ Describe what you are neglecting and why.
3. Ensure that you at least identify and incorporate what the main drivers (variables that a ect ▪ Indicate what are the indicators/variables that you will use.
research outcome) might be. ▪ Outline how you will measure those indicators/variables.

4. Build towards a main hypothesis that can be tested. The details will allow you to establish the technical research design.

Some tips and common pitfalls

▪ The Research Questions are questions YOU the researcher must answer. NOT the questions the 7. Establish your technical research design
client has!
▪ You can NEVER have one sub question; you must have at least two. Strategic Framework for Technical Design
▪ Research questions are best if they are posed in an open-ended way. If you pose them as yes/no  Research Strategy
questions, you will limit your research.
▪ Make sure that your question is Relevant (connected to research objective), Researchable (not  Research Material
subjective or vague), Anchored (clear context) and Precise (specific drivers).  Research Planning
▪ If you struggle with the technical design, it is most likely that your set of research questions is
inadequate – feel free to go back and edit your research questions throughout your research if Research strategy
necessary.  Here you take your theoretical framework and start considering the practical aspects.
  The research strategy should define exactly how you want to do your research.  Respect
 Consider:
 Responsibility
o Suitability (Your skills, academic level of problem)
o Availability (are resources such as test equipment available)  Engagement
o Duration (Can it be done within the set time)
 Courage
o Ethics (should it be done; do you want to do it)
Research involving human test subjects
Research materials
▪ TU Delft has an ethics board for human research: Human Research Ethics Commission (HREC).
 Research materials include every tool you need to perform research. ▪ Permission must be sought for research that involves human test subjects. This includes amongst
 Depending on your topic this could be just a computer with a specific software or complex others: participants in questionnaires, interviews, and testers (such as test pilots).
laboratory equipment.
 Some recommendations:
o Make sure you have access to the materials you need as early as possible. 9. Learn about sampling and the sampling method
o Learning how to use tools takes time. Do not assume you will produce results right
Research Material
away.
o Tools might not perform as anticipated. Expect the failure points and have 1. What is the population of your study?
contingencies for them. Population is defined as “entire collection of things” in your field. The population refers to the
entirety of the problem and it is too large to tackle in one study.
Research planning
2. What subset of my population am I going to study & why?
 What is the time frame for all your activities?
Your chosen subset is called a Sample.
 All plans change – make one that allows flexibility for those changes.
 An initial plan should include: Sampling
o Milestones ▪ We can rarely study an entire population (The Problem); hence we must take a representative piece
o Deliverables (The sample).
o Holidays ▪ Ideally, a Sample is a representative cross-section of your population that allows you to make
 Use system engineering tools to create your plan: Work Breakdown Structure, Work Flow generic statements over the entire population.
Diagram and Gantt chart can be useful if used properly. ▪ Choosing the right sample will improve the quality and speed of your research.
▪ In order to be representative, this needs to be proved statistically and there are required sample
sizes.
8. Learn about ethics during your research process ▪ Sometimes our sample is not representative of the whole problem. In this case, one cannot
produce conclusion regarding the whole population, just the studied sample.
General recommendations
▪ Always ensure your work is reproducible and can be traced back to original data. Sampling method
▪ Keep your raw data.
1. Make it truly generic -> The theoretical population: To what/whom do you want to
▪ If given data, verify its origin.
generalize?
▪ Always give credit where it is due.
▪ Follow the university’s code of conduct. 2. Group you want to generalize -> The study population: To what population can you get
access?
Core Values (of TU Delft)
3. List how you draw population -> The sampling frame: How can you get access to them?
 Integrity

4. The actual units used in your study -> The sample: What/who is in your study? ▪ the greater the SD the greater the error (more variability in sampling distribution).
▪ the greater the sample size the smaller the error (nearing full population).

Convenience sampling
10. Learn about the types of samples
▪ The sample is selected because it is available.
Types of Samples ▪ This needs to be explicitly stated in your research.
▪ It can never be assumed that this sample represents the total population.
 Probability Sampling
o Simple random sampling – every element in population is known and has the same Purposive sampling
chance of being selected as a subject.
 The sample is selected because we desire the specific information it holds.
o Complex probability sampling – improve e iciency by no longer having every element
 Reasons can be:
have the same chance of being selected.
o Expert advice
 Non-Probability Sampling
o Quota needed (I need exactly this subjects)
o Convenience sampling – who of the population are available to be sampled.
o Heterogeneity needed
o Purposive sampling – who in the population will hold the desired information.
o Snowball e ect: respondent driven sampling. (one subject leads to another)
Simple Random Sampling  This needs to be explicitly stated in your research.
Drawing a sample from a population so that every possible sample has an equal probability of being  It can never be assumed that this sample represents the total population.
selected.
Example: From a list of clients draw a random sample.
11. Learn about data sources, types of data and data quality
Complex random sampling
Improve e iciency by no longer having every element have the same chance of being selected. Data sources
Some examples are:
 What sources are you going to use?
1. Systematic Random Sampling: Determine randomly where you want to start selecting and o Someone else’s data
then select every nth element in a randomly ordered list. o Theoretical data
o Simulated data
2. Cluster Random Sampling: Dividing population into groups, randomly selecting these
o Experimental data (testing/surveying/observing).
clusters and then sampling each element.
Data rules and regulations
3. Multi-stage Random Sampling: Similar to single stage sampling but reiterating the sampling
method within the sample.  Personal Data: We speak of personal data if data can be traced directly or indirectly to a
Example: Select 5 countries within EU to represent the EU, within that select 5 towns, 5 natural person. All personal data are subject to the rules of the General Data Protection
suburbs and 5 villages and within those 2 random ATM machines in order to determine the Regulation (AVG).
bacteria level on EURO notes.
 Commercially sensitive data: When you are doing your thesis within a company. Do not
Standard error sign Non-Disclosure Agreements (NDAs) or contracts with companies until you or your
The standard error is the precision of your statistical sample, and it is calculated from the standard supervisor has received legal advice from the university that it is OK to sign.
deviation (SD).
 Protected Data: Both the UN and the EU require countries to ensure that knowledge related
Standard deviation: the spread of variability of the scores around their average in a single sample
to nuclear weapons and missile technology is protected.
(square root of the variance).
SD is expressed in the same units as original measure and easy to interpret. Types of Data
In general: Generally, two types can be distinguished:
▪ Qualitative Data – Data collected by means of observation of the environment in which you are ▪ Allow your results to be used in a professional environment.
interested such as text, photos, people. ▪ Allow for academic reproducibility and transparency.
▪ Quantitative Data – Numerical results of measured/simulated variables from wind tunnel data,
Verification and validation in industry
FEM results, surveys.
 Verification is a quality control process.
The type of data you generate follows from your research strategy, research material and your
▪ Used to evaluate whether a product, service, or system complies with regulations,
defined variables.
specifications, or conditions imposed at the start of a development phase.
Sometimes, you will want to convert qualitative data in quantitative or vice versa.
▪ Verification can be in development, scale-up, or production.
Data quality ▪ Often an internal process.
Before starting to analyze you must ensure data is of su icient quality:
 Validation is a quality assurance process.
▪ Check for typing errors.
▪ Establishing evidence that provides a high degree of assurance that a product, service, or
▪ Check for outliers.
system accomplishes its intended requirements.
▪ Check for errors in coding your data.
▪ Often involves acceptance of fitness for purpose with end users and other product
▪ Check if data meets expectations.
stakeholders.
Some recommendations: ▪ Often an external process.
▪ Keep a good research logbook of what you do.
Verification and validation in your research
▪ Have a data management plan.
1. Observation & Formation of the topic
Research Logbook
Document where you document: 2. Hypothesis
▪ Choices you made in your research design.
3. Conceptual Design
▪ Assumptions you made.
▪ How you carried out your experiments. 4. Technical Design
▪ Any incidents during your experiments.
5. Data Gathering
▪ Where you have stored pictures & footage & relevant files.
▪ File names and changes per file. 6. Data Analysis
▪ People you talked to (including your meetings with your supervisors).
7. Data Interpretation
The more detailed your logbook is, the easier your thesis writing will be.
8. Validation & Verification, Revision of hypothesis

9. Conclusion, reiteration if necessary

12. Learn about verification and validation
Verification and validation in your thesis
Definitions Each project is di erent, but some generic guidelines to consider are:
▪ How are you going to verify your methods?
 Verification is proving that your method of research has been used correctly and that it is
▪ How can your research goal be validated and proven to be true?
suitable for your research topic.
▪ How high will be the quality of the validation in your thesis research?
 Validation is proving that your outcomes are true and based on strong (scientific) evidence.
In your planning, it is worthy to:
Both are necessary to: ▪ Design a validation exercise into your thesis work.
▪ Add credibility to your results and conclusions. ▪ Time to perform the required validation, not just exploration and implementation.

13. Learn about conducting sustainable research Digital resources

By being conscious of your digital habits, you can reduce your carbon footprint and make your
Engineering for One Planet
research more environmentally friendly.
▪ When planning your research, you should consider the environmental, social, cultural, and
Some recommendations:
economic implications.
▪ Simulating Locally vs. on a Server: Running a simulation in a server is often unavoidable given
▪ To help you consider these aspects, the Engineering for One Planet (EOP) framework will be used.
the size of simulations but consider that they are very energy intensive (~5kWh/GB). By limiting the
▪ This framework will be used to cover these aspects:
amount of data transferred, and the time the servers are in use, you can minimize the footprint of
1. Value of research your simulations.
▪ Storing Files in a Hard-Drive vs. Cloud: storing your files in datacentres uses ~10E5 kWh/GB
2. Reproducibility of research
more than storing your files locally. Consider reducing the file size before backing up your files and
3. Resource management only backing up critical files in the cloud.
▪ Reading Online vs. O line: Websites are stored in datacenters which consume approximately
1. Value of Research
5kWh/GB to be run and cooled. If you need to access a document often, consider downloading it.
The EOP framework expects you to be aware of the risks that may come with your research, and
▪ Printing vs. Digital. A good rule of thumb to make sure you don’t over-print is to only print if you
requires you to forecast not only the near-term, but also the long-term cost and values of your work.
will read the document back at least 3 times.
Therefore, once you have decided on a topic, consider:
▪ Limiting additional screens when possible.
▪ Is there potential value to the scientific community or the industry from the findings?
▪ Can the research be built upon? Methodologies
▪ Is there a direct potential harm to the environment, society or economic implications of the To minimize the footprint of your research, you should be critical about the methodology you choose
research when used as it is intended? and its social and environmental impact.
▪ Can you foresee the research being used for unintended purposes with potential environmental, Before finalizing your method, you should reflect on the following points:
societal, cultural, or economic consequences? ▪ Is the desired quality/accuracy realistic given the resources available?
▪ Are there any alternative experimental set-ups that are less resource intensive with a reasonable
2. Reproducibility of Research
accuracy/quality?
Not only does reproducibility indicate transparency and rigor in the research, but it also allows for
▪ Is your sample size justified?
additional research exploring similar topics to be conducted using less resources.
Main barriers to reproducible research:
▪ Complexity of the work
14. Learn about data management
▪ Technology changes/advancements
▪ Human Error Why do we need to think about data management?
▪ Intellectual Property Rights ▪ Storing and documenting correctly our data facilitates Open Science and makes the
reproducibility of research higher.
To overcome the barriers of technology changes and human errors, good data management is
▪ Good data management includes not only the storing of raw data but also the metadata and the
necessary.
workflow applied to the analysis.
3. Resource Management ▪ Producing good documentation and metadata ensures that data can be understood and used in
Being aware of the resources you are using during your research, which will help you make more the long term.
environmentally conscious decisions. ▪ To do this, a system must be designed beforehand. It is very hard to organize your data only at the
Resource management will be divided into two relevant categories: end of the process.
▪ Digital resources
Data management plan template
▪ Methodologies
I. DATA DESCRIPTION
  Type of data ▪ If you leave TU Delft (or are unavailable), who is going to be responsible for the data resulting from
this project?
 File format(s)
Other data management recommendations
o How will data be collected (for re-used data: source and terms of use)?
 Use descriptive, self-explanatory file and variable names.
 Purpose of data processing
 Be consistent with your filing system.
o How much data storage will you require during the project lifetime?
 Keep an uncorrected original data file.
 Storage location
 Store your data as much as possible in non-proprietary software based on open formats.
o Who will have access to the data?
 Document your workflow.
II. DOCUMENTATION AND DATA QUALITY
▪ What documentation will accompany the data?
▪ Methodology of data collection
15. Learn how to apply project management techniques to a research project
▪ Documentation in an Electronic Lab Notebook
▪ README file Why Project Management in a Research Project?
▪ Data dictionary explaining the variables used
1. Organized Progress: It ensures structured timelines and milestones, helping researchers
▪ Other - explain below
stay on track.
III. STORAGE AND BACKUP
2. Resource Optimization: It maximizes the e icient use of resources, such as time, and
▪ Where will the data (and code, if applicable) be stored and backed-up during the project lifetime?
research tools.
▪ Design a back-up strategy from the start.
▪ Use the 3-2-1 rule of data storage: 3 copies, 2 storage media, 1 copy in a di erent physical 3. Risk Mitigation: It identifies potential risks early on, and this way it reduces the likelihood of
location. unexpected setbacks.

IV. Legal and ethical requirements 4. E ective Communication: It facilitates transparent communication with the supervisory
▪ Will you work with any other types of confidential or classified data? team. This makes sure that the challenges of the project are clearly shared.
▪ Does your research involve human subjects or 3rd party datasets collected from human
Estimating your tasks
participants?
The first step in project management is to make estimates about the steps needed to complete the
V. DATA SHARING AND LONG-TERM PRESERVATION project.
▪ What data will be publicly shared? To accomplish that we first outline all the tasks needed and then estimate how long each task might
▪ How will you share your research data (and code)? take.
Manage & Share | [Link]
There are three main estimating methods that can be used:
▪ Under what license will be the data/code released?
▪ Comparative estimating
▪ CC0 (No Rights Reserved, Public Domain)
▪ Parametric estimating
▪ CC-BY (Attribution)
▪ Bottom-up estimating
▪ CC BY-SA (Attribution-ShareAlike)
▪ CC BY-NC (Attribution-NonCommercial) Work Breakdown Structure (WBS)
▪ A Work Breakdown Structure visually divides the master thesis project into tasks, breaking down
VI. DATA MANAGEMENT RESPONSIBILITIES
the research into smaller, more easily manageable pieces.
▪ Is TU Delft the lead institution for this project?
▪ The WBS usually is divided in work packages, tasks and (maybe) sub-tasks.

▪ The preparation of a WBS for the project must be done down to a level of detail that allows for an  Software guides
estimation of the time required per task.
ACADEMIA
▪ The individual estimates are then summed to give the total estimate for the project.
 Assignments
Resource Management
▪ In project management terms, resources are the people, equipment, facilities or any other entity  (Master / PhD/ ...) thesis
that is needed to complete an activity.
 Conference articles
▪ Some resources are re-usable (such as people), and others are consumable (such as a plate of
aluminum alloy).  Journal articles
▪ Most resources are not always available and require tasks to be performed to have access to
The goal is to communicate information needed to UNDERSTAND A SUBJECT or to CARRY OUT A
them.
TASK.
▪ To understand the complete picture, the resource or resources required to work on each task must
be described and their availability (or lack thereof) known. What to write in a Technical Report?
A technical report conveys SPECIFIC INFORMATION about a TECHNICAL SUBJECT to a SPECIFIC
Project Planning
AUDIENCE for a SPECIFIC and PRACTICAL PURPOSE.
▪ Project Planning is the time-based articulation of the overall project plan.
▪ A proper schedule can only be put together when all conceptual and technical research design has What to write in the Research Methodologies Report?
been outlined. A technical report conveys your RESEARCH PLAN about YOUR RESEARCH TOPIC to a TECHNICAL
▪ Putting together a plan allows you to predict the overall project duration and when activities and PERSON FAMILIAR WITH YOUR FIELD with the purpose of ORGANISING YOUR RESEARCH
events are planned to happen. ACTIVITIES.
▪ The most common way of showing a project’s plan is to use a Gantt chart.
Any Report Structure
▪ A Gantt chart allows you to detect the critical path or paths for the project. A critical path will
determine the shortest time in which a project can be completed.  HEAD = INTRODUCTION: It helps the reader understand the report but it has to be
proportional to the main body.

 MAIN BODY: Divided into sections but they have to be connected to each other in a nice
16. Learn why write a Technical Report & Learn what to write in a Technical Report
flow.
Why write Technical Reports?
 FEET = CONCLUSIONS: The whole report stands on them. No new information here.
“Everybody hates to write them,
Everybody hates to read them, Research Methodologies Report Structure
Almost nobody reads them,
 HEAD = INTRODUCTION: It helps the reader understand the importance of your research.
They take a huge amount of time to write,
Why is it important to study this?
They are absolutely crucial to your success in engineering!”
 MAIN BODY: Describes your research plan.
Because you have to: Communicating your work is not something to add to your job, it is part of
your job! o Research questions

INDUSTRY o Methods, tools and expected results

 Project results report o Planning

 Project handover documentation  FEET = CONCLUSIONS: Describes concisely all the aspects of your report. If you just read
the conclusions, you get a good idea of your plan.
 Instruction manuals
When to write? Some recommendations Numbered style e.g.: AIAA style

 Write little and often, avoid binge writing, embrace snack writing.  Citation in the document:

 Writing clarifies your thinking. o Parenthetical citation: E.g.: The most important parameter to scale a sandwich
composite shell is the ratio between the radius and the equivalent thickness of the
 If you usually: “think – think – think” and then “write – write – write”.
sandwich structure [1].
Try instead: “think – write – think – write – think – write”.
o Narrative citation: E.g.: Uriol Balbin et al. [1] demonstrated that the most important
 Set specific times to write.
parameter to scale a shell is the ratio between the Radius and the equivalent
 Set specific places to write. thickness of the sandwich structure.

 Reference list: References ordered numerically, in the order in which they appear in the text.
[1] Uriol Balbin, I., Bisagni, C., Schultz, M. R., and Hilburger, M. W. "Scaling Methodology
17. Check that your reference list is correct
Applied to Buckling of Sandwich Composite Cylindrical Shells," AIAA Journal Vol. 58, No. 8,
Some suggestions 2020, pp. 3680-3689.
▪ Use at least 3 di erent document types.
Advantage: A reader who wants to check all references while reading a paper, can easily move from
▪ Use at least one book (Usually this is used for theory background).
one reference to the next without having to search for the next author’s name in a (long) list.
▪ Use at most two websites (or no more than 10% of the literature list).
Disadvantage: In-text references do not contain any useful information themselves.
▪ Include patents, standards, technical documents etc.
▪ Most of your sources should be peer-reviewed journal articles. Final Reference Checklist
▪ Check the topicality of the information: Most sources must be less than 10 years old.
 All illustrations, that are not your own, are referenced.
Choose one reference style and stick to it!
 All statements, that are not your own, are substantiated with references. (Where did you find
Author-date style e.g.: APA style the information?)

 Citation in the document:  Direct citations are put in quotation marks; paraphrases are written in your own words.

o Parenthetical citation: E.g.: The most important parameter to scale a sandwich  Each citation or paraphrase includes a correct in-text reference according to the rules of the
composite shell is the ratio between the radius and the equivalent thickness of the chosen reference style.
sandwich structure (Uriol Balbin et al., 2020).
 All references are correct and complete, according to the rules of the chosen reference style.
o Narrative citation: E.g.: Uriol Balbin et al. (2020) demonstrated that the most
 The literature list contains all references used in the text.
important parameter to scale a shell is the ratio between the radius and the
equivalent thickness of the sandwich structure.  The literature list contains only references used in the text.

 Reference list: References ordered alphabetically, by author’s name.

Uriol Balbin, I., Bisagni, C., Schultz, M. R., & Hilburger, M. W. (2020). Scaling methodology
18. Learn some best practices for writing a report
applied to buckling of sandwich composite cylindrical shells. AIAA Journal, 58(8), 3680–
3689. The reader is the most important person
Step 1: DEFINE YOUR TARGET READER
Advantage: A reader can recognize important original texts from the in-text reference, without
▪ Who is going to read this?
having to check the reference list.
▪ What are they interested in?
Disadvantage: Longer in-text references can make a text more di icult to read.
▪ What do they need this information for?”

Step 2: GUIDE THEM IN:  The nomenclature is consistent throughout the report.
The reader does not know your work. You want the report to attract and maintain readers’ attention.
 All acronyms are defined the first time they are mentioned.
Readers first look at the titles, then at the pictures, then at the conclusions, and then to the rest of
the report.  All figures are readable in pdf format.
▪ The layout must look nice (right placement, high resolution images).
 All figures and tables include a caption.
▪ Images and block diagrams help convey information much more than detailed explanations.
▪ Images, tables and graphs must be clear and self-explanatory.  Figures are clear (units, axes labels, legend …) without reading the text.

How to write: 4Cs style principles  The title captures the contents of the report.

 CLEAR: Easily understood by the intended audience.  The introduction captures the attention of the ideal reader to the report.

 COMPREHENSIVE: All the necessary information is present.  Conclusions are understood without reading the rest of the report.

 CONCISE: Clear without verbiage (=as short as possible).

 CORRECT: Accurate and grammatically correct.

Some graphs style tips

▪ Always use tick marks with major gridlines at the start (often 0,0) and end of the graph.
▪ If more than one line, use a legend. Make sure the legend does not overlap with your lines.
▪ Keep any dots the same color as the indicated line.
▪ Use either straight or short dashed lines. Avoid dotted lines (can’t see on large screens) or long
dashed lines (hard to capture nonlinearities).
▪ Be consistent with the units, either “Force, kN” or “Force (kN)”.
▪ Be consistent with the font type and size. Try to avoid scientific notation and make sure the font
size is large enough.

AI tools for technical and scientific writing

Be responsible in your use and pay attention to their limitations. If you use it to:

 Improve or generate text: sometimes it is too redundant (overwriting), you might need to
edit it out!

 Generate ideas or brainstorm: might provide good ideas but remember that it can only look
back in time!

 Provide content or references: prone to error, be aware!

Final writing checklist

 No spelling or grammar errors.

 Every paragraph conveys one or two ideas.

 The paragraph format is consistent throughout the report.