ECO management is an innovative solution for energy saving and reducing negative impact on the environment. The platform is using artificial intelligent hardware and software for monitoring and managing consumption using machine learning methods to stipulate optimal consumption, climate and technological processes.

We provide

Quick and free installation on-site
Active management of clients' engineering system equipment without needing to modernise them
Measuring energy resources on managed equipment
Energy monitoring (monitoring business processes)
Monitoring climate parameters
Environmental friendliness by reducing carbon footprint.

How EcoManagement system works

• Data sensor information is uploaded to a cloud database for further processing

• Artificial intelligence-based algorithms send control signals to controllers which adjust climate, equipment and lighting

• All data on power consumption, equipment operation and incidents are available in customer’s personal online account.

Monitoring energy consumption in real-time
Ecomanagement system components
  • Collecting and storing data on energy consumption, climatic parameters and facility activity on a per second basis
  • Recognition of typical power consumption profiles based on machine learning methods
Analysis of gathered data
  • Automatic detection of abnormal and atypical equipment behavior
  • Search for the most promising optimization options (based on data accumulated at previous facilities)
  • Building a prediction model to forecast the dynamics of consumption throughout the day/week/season
Construction of a two-level automatic load control system
  • The internal control loop is based on a PID regulator, provides an acceptable level of control and remains operational in the absence of an internet connection
  • The outer control loop is based on Deep reinforcement learning and adjusts the parameters of the inner control loop to maximize equipment performance
Data is collected at a facility to establish the state of devices, climate and activity. EcoManagement uses the Modbus data transfer protocol, implemented through physical twisted pair cabling or wireless connection based on Wi-Fi, ZigBee, Z-wave and other protocols.
Based on current indoor climate prediction models, a multi-layer LSTM architecture is used inside facilities
  1. A set of input signals: (xi (t-l), xi (t-l + 1) ... xi (t))
  2. LSTM layer (N=200, activation function = ReLU)
  3. Second layer (N=1)
  4. LSTM layer (N=200, activation function = ReLU)
  5. TimeDistibuted layer (N = 200, activation function = ReLU)
  6. Fully connected layer (N = 1, activation function = linear)
  7. Exit layer: (yi(t+2), yi(t+2) ... yi(t+n))
The following are used as a quality assessment metric:
Active control
Currently implemented climate and illuminance models show prediction efficiency at the level of 90-95%, and we are constantly adding new objects and expanding our library of solutions

Based on all the data, our models can predict the state of a facility several cycles ahead in a dynamics of its change and determine how various factors can affect it, as well as suggest whether any action is required to bring it closer to is optimal state.
Depending on the situation and the rules of impact the system can:
Results and plans:
As a result of implementation, our clients receive a comprehensive solution that allows them to:
1. Significantly reduce the cost of electricity (on average up to 15 - 25% on site and over 50% on some equipment)
2. Ensuring optimal performance of equipment, extending its service life
3. Possibility to reduce the initial costs of creating a power supply system (at the design or construction stage of the facility)
4. Transparency of business processes and online monitoring of activities (KPI)
At the moment, we are applying these solutions in catering to restaurants, warehouses, stores.
We are working on projects for gas stations, school, office premises; we expect to implement them in the near future. As we add object, the efficiency of our system increases significantly, both for new and for previously implemented solutions.